CONFIRMATION OF FAINT DWARF GALAXIES IN THE M81 GROUP

The initial CFHT/MegaCam r'-band imaging survey was carried out over three semesters: 2005B–2006B. A contiguous 65 deg² region centered on M81 was observed in a mosaic pattern which resulted in each region being imaged twice with combined total integration times of 1096 s. Full details of this survey are presented in Chiboucas et al. (2009). The M81 Group has a 230 kpc second turnaround radius of spherical collapse (Tully 2010) which, at the group distance of 3.7 Mpc, corresponds to 3.65 deg on the sky. The MegaCam survey therefore extends beyond this radius and provides nearly complete areal coverage of this nearby galaxy group. The survey limiting magnitude for resolved stars was r' ∼ 25.0–25.5, dependent on seeing variations throughout the survey. Given the high level of extinction in this part of the sky (Schlegel et al. 1998; Schlafly & Finkbeiner 2011; see Table 2), we found we could resolve stars down to $M_{r^{\prime }} \sim -2.9\pm 0.5$. The TRGB in r' is −3.1 ± 0.1. Stars at the TRGB are just at the edge of detection while young main-sequence (MS), red supergiants, and intermediate-aged asymptotic giant branch (AGB) stars are easily resolved over the low surface brightness diffuse late type MS population light.

Exploiting this, we searched the imaging for previously undetected faint dwarf galaxies. Our expectation was that any dwarf galaxy, no matter how faint or diffuse, should have been detectable, as long as it had built up a significant enough red giant branch. In retrospect from results discussed further along, this requirement of a well developed red giant branch is satisfied for galaxies brighter than $M_{r^{\prime }} \sim -9$. Using two methods (a two-point correlation routine and visual inspection by two of the authors) to search for concentrations of resolved stars which would indicate potential dwarf galaxies, we identified a total of 22 candidate dwarf group members. This conceivably doubles the number of known group members, and includes three blue compact dwarfs (BCDs), a tidal dwarf, and a number of dI and dSph candidates. Three of the candidates were previously known galaxies. However, d0958+66 (KUG 0945+670) and d1012+64 (UGC 5497) had no velocity or distance measures that would associate them with the M81 Group. These two galaxies are compact star-forming objects which were found to resolve into individual stars in our MegaCam imaging and thus considered strong candidates for group membership. A third object, d0959+68, was originally noted by Durrell et al. (2004). In a survey for resolved red giant stars near the core of the M81 Group, they discovered a concentration of blue stars which they suggested was a likely tidal dwarf within the group.

Artificial galaxies having a wide range of sizes, total magnitudes, and stellar surface densities were inserted in the imaging to determine survey completeness and recovered using the same methods described above. Results indicated that limitations did exist for this survey, due in part to gaps between the CCDs and regions with very high extinction, and due to what are effectively total magnitude and surface brightness limits. Out to a distance beyond the second turnaround radius (roughly equivalent to the Virial radius), Chiboucas et al. (2009) contend that the survey is ∼90% complete to $M_{r^{\prime }} = -9.8$.

In order to securely establish group memberships, we need to measure distances to these candidate galaxies. Previous work has shown that it is possible to measure TRGB distances for resolved stellar populations in dwarf galaxies out to ∼10 Mpc (see, e.g., Makarov et al. 2006; Rizzi et al. 2007; Jacobs et al. 2009). The M81 Group is nearby at ∼3.7 Mpc, a distance where the brighter members of galaxy stellar populations, including red giant stars, are resolved in single orbit space-based imaging. We therefore make use of the high resolution of HST to obtain two-color photometry of resolved stars in these candidates. From the resulting CMDs, we can measure the TRGB and use this to measure distances and confirm group membership. With the Advanced Camera for Surveys (ACS) on HST, we can reach V = 28.0 and I = 27.0 with integrations taken together in a single orbit. At the distance of the M81 Group, the TRGB is found at I ∼ 23.8, assuming M_{I, TRGB} = −4.04 ± 0.12 for metal-poor populations (Bellazzini et al. 2001). ACS imaging probes deeper than 3 mag below the TRGB assuring secure determinations of the TRGB and, with uncertainties in the tip magnitude typically 0.1 mag, distance determinations accurate to 200 kpc. In addition, with effective radii ⩽30 arcsec, the 202 × 202 arcsec field of view of ACS provides essentially full coverage for these candidates. With WFPC2, the data only reach ∼1.5 mag below the TRGB, uncertainties in the distance measurement are larger, and bigger galaxies extend beyond the field of view.

Imaging was obtained in both F814W and F606W bands for each of our 22 candidates. Although Johnson–Cousins V- and I-band filters are traditionally used to obtain TRGB distances, the F814W filter transforms well to Cousins I, while the F606W provides adequate color separation and is chosen over the F555W for its greater throughput. An initial set of 14 candidates was observed in Cycle 16 (GO-11126). Due to the failure of the ACS, the program was carried out using the lower sensitivity WFPC2 camera. A second set of 8 candidates was observed with the repaired ACS/WFC (GO-11584) in Cycle 17. One candidate was re-observed with ACS in order to probe the stellar population to fainter magnitudes. Regardless of instrument, each candidate was observed in both bands in a single orbit. This meant two exposures with combined total integration times of 900–1000 s in F814W and 1000 s in F606W for WFPC2 observations, and ∼1250s in F814W and F606W for ACS/WFC observations. See Table 1 for a summary of observations. Thumbnail images of each candidate are shown in Figures 1–3. Color images are shown for a subset of these in Figures 4 and 5. Color figures for all can be found in the Extragalactic Distance Database (EDD) at http://edd.ifa.hawaii.edu.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We started with bias corrected and flat-fielded data produced by the STScI pipeline. Photometry was then performed using the point-spread function (PSF) fitting stellar photometry packages from A. Dolphin. HSTPHOT (Dolphin 2000b) was originally run on the WFPC2 data, while the ACS module of DOLPHOT was used for the ACS data. Pre-processing with these routines included bad column and hot pixel masking, cosmic ray rejection, and sky determination, as described in the manuals. Sources are then simultaneously detected and photometered in both F814W and F606W images, using PSF fitting with model PSFs from TinyTim.⁴ Aperture corrections were made based on stars in uncrowded regions. Charge-transfer inefficiency and zeropoint corrections are also applied, and measured magnitudes are provided as both flight system instrumental VEGAMAG magnitudes and as corresponding Johnson–Cousins transformed apparent magnitudes. Charge-transfer-efficiency (CTE) corrections and zeropoints are taken from Dolphin (2000a) for WFPC2 data, while DOLPHOT makes use of the newest Chiaberge (2012) CTE corrections and revised magnitude zeropoints from the STScI Web pages.

A detection threshold of 3.0σ was applied. This produced F814W and F606W matched catalogs with output including χ² for the fit, S/N, sharpness, roundness, position angle, crowding, object type (distinguishing between star, faint, and elongated/extended objects), and magnitudes. We then culled this output to reject anything that was not considered a point source with sufficiently high S/N. Final object lists contained only those objects classified as stars with χ² < 2.5, (sharpness)² < 0.09, S/N >5 in both bands, and quality flag 0. Final lists consisted of ∼10, 000 and typically less than 1000 good stars in ACS and WFPC2 fields, respectively.

Using artificial stars, we tested the detection completeness and photometric uncertainties of HSTPHOT/DOLPHOT. Between 100,000–200,000 sources were added using the Dolphin routines to WFPC2 and ACS imaging, one at a time so as to not introduce additional crowding. HSTPHOT/DOLPHOT were then run in the same way to recover these artificial sources, including implementing the same rejection parameters that were applied to the real data. Detection completeness is shown in Figure 6. For WFPC2 data, we find an F814W 50% completeness limit of ∼25.5  ±  0.2. The crowding of each field appears to affect the recovery rate, with crowded fields reaching ∼0.2 mag shallower than this and sparser fields probing ∼0.2 mag deeper. F606W data go 1.2 mag deeper than the F814W data. For ACS data, the difference in depth due to crowding is as large as ∼0.8 mag between dense and uncrowded fields. We find a 50% completeness limit of ∼27.3 ± 0.4 in F814W. F606W data reach about 1 mag deeper. At the TRGB at F814W ∼ 23.8, we expect 100% completeness in all datasets. For WFPC2 data, the 50% completeness limit is less than 2 mag below the tip. However, even in crowded fields, we expect completeness to be >90% one magnitude below the tip.

Zoom In Zoom Out Reset image size

Uncertainties in the photometry are also investigated using the artificial star tests. Results are displayed in Figure 7. The scatter at faint magnitudes becomes quite large. At F814W = 23.5, close to the expected TRGB magnitude for M81 Group members, the 1σ uncertainty is ∼0.055 mag for ACS and 0.11 mag for WFPC2 data. The uncertainty is slightly higher, by ∼0.03 mag, for the most crowded fields in each case. For individual stars at the TRGB, this translates to an uncertainty in distance of 100 and 200 kpc for ACS and WFPC2 data, respectively. The second turnaround radius for the M81 Group is 230 kpc, within which we expect all galaxies to be group members. Since the TRGB measurement is based on the ensemble of stars at the TRGB, the uncertainty decreases with increasing number of stars populating the upper TRGB. Makarov et al. (2006) find that the maximum likelihood method for determining the TRGB which we implement should produce reasonable results for RGBs consisting of >50 stars in the top magnitude before the tip. Thus, the photometric uncertainties are within the range needed to unambiguously establish group membership, at least for galaxies with an RGB population size sufficient for determining the location of the RGB tip.

Zoom In Zoom Out Reset image size

CMDs for all candidates are shown in Figures 8–11. Error bars denote 1σ uncertainties in magnitude and color as determined from the artificial star tests. The most prominent feature in many of these CMDs is the RGB at F606W − F814W ∼ 0.8, the presence of which is confirmation of the galaxian nature of these objects. The RGB is visible as a long, narrow strip in the ACS data, seen in at least five of the top six panels of Figure 8. For d0959+68 the presence of an RGB is more questionable. Since all candidates were centered in the WFC chip 2, we show the CMDs for stars in the bottom detector for several cases in Figure 9. Hints of an RGB in d0959+68 may indicate that the target extends onto the other CCD. The RGB has a wedge-shaped appearance in candidates shown in Figure 10 observed with WFPC2. The broadness at the detection limit is primarily due to large photometric errors. No such feature is apparent in the remaining five candidates observed with WFPC2 (Figure 11). For comparison we also include in this plot the stellar detections for the d0926+70 observation taken from a different chip away from where the target was centered and which exhibits fewer stellar detections. We have no independent control fields, but as many of our galaxies are spatially quite small, we can use part of some of the images as an approximation of the expected foreground stellar contamination.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In a few of these CMDs (d0959+68, d1012+64, d0958+66, and d1028+70), an MS is visible as a vertical strip around F606W − F814W ∼ −0.2. Blue loop stars are also likely present at bright magnitudes. These stars indicate the existence of a young stellar population. AGB stars, representative of intermediate-age populations, are also apparent in some of these CMDs above and redward of the RGB. These populations will be discussed further in Section 3.5.

Red giant stars have ceased hydrogen burning in their cores while continuing to burn hydrogen in outer shells. This shell burning causes the outer envelope of the star to expand and become redder and brighter while dumping material on a degenerate core, thereby increasing the core mass and temperature. Because of the degeneracy, the star cannot compensate for the increased core temperature and, when the core reaches the required temperature, at a constant core mass and therefore a predictable luminosity, the onset of helium burning occurs as a runaway thermonuclear reaction. The degeneracy is lifted before the star explodes, and the star then spends the next stage of its life on the lower-luminosity horizontal branch. Thus, the TRGB, immediately before a star undergoes the helium flash, has a near constant bolometric luminosity and can therefore be used as a standard candle. Although this luminosity is strongly dependent on the underlying stellar population metallicity and, to a much lesser extent, the population age, it is least sensitive to these evolutionary variables in the I band (Rizzi et al. 2007).

Accurate TRGB distance measurements can therefore be made with good calibration for the I-band tip luminosity. For this, we use the work of Rizzi et al. (2007). They define a zeropoint calibration of the TRGB, accurate to 1% statistical uncertainty, as a function of the stellar population color to account for variation due to metallicity and age. In addition, the zeropoint is provided in the HST flight system for F606W and F814W for both ACS and WFPC2 cameras. They find

$$\begin{eqnarray*} &&M^{{\rm ACS}}_{F814W} = -4.06 + 0.20 [(F606W - F814W) - 1.23] \\ &&M^{{\rm WFPC2}}_{F814W} = -4.01 + 0.15 [(F606W - F814W) - 1.12]. \end{eqnarray*}$$

The measurement of the TRGB can therefore be made directly within the flight system magnitudes as measured by DOLPHOT/HSTPHOT, eliminating additional uncertainty from magnitude transformations into the Johnson–Cousins system.

In order to determine the precise location of the apparent magnitude of the tip we use the maximum likelihood method of Makarov et al. (2006), as described in Jacobs et al. (2009). This has the advantage over edge detection algorithms in that no binning or smoothing is required, although a functional form must be assumed for the stellar LF. We use the TRGBTOOL program written by D. Makarov which assumes a simple power law with a cut-off faintward of the TRGB and a second power law with a different slope brightward. Default values for the slopes are provided although these can be varied as necessary to fit the data. The discontinuity in the fit determines the TRGB location. The procedure incorporates the results of the artificial star tests discussed above to account for completeness, crowding, and photometric errors and provides realistic uncertainties in the tip measurements.

We run TRGBTOOL on HSTPHOT/DOLPHOT photometry for all the candidates. In most cases where we believe an RGB is visible in the CMD, the automated tip finder locates a tip magnitude close to the expected value for M81 Group members. In several cases, e.g., d0934+70 and d0959+68, the RGB is so poorly populated that uncertainties in the tip measurement are very large. In addition, for d0959+68 the tip measurement obtained from the tip finding software may be affected by AGB stars. The few cases with well populated RGBs have small TRGB measurement errors and well constrained distances.

The spatial distribution of detected stars in WFPC2 and ACS images for 14 candidates are displayed in Figures 12–16. Symbols are coded by region of the CMD. Black points denote all good stellar detections. Red pentagons are stars that fall within the range of the RGB: F814W_TRGB < F814W < 25 and 0.4 < F606W − F814W < 1.4. Black asterisks represent possible intermediate-age AGB stars having F814W < F814W_TRGB and F606W − F814W > 0.6. Blue open triangles denote potential young MS and blue loop stars with F814W < 25 and F606W − F814W < 0.2.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Although the stellar detections have been culled for only "good" stars, some artifacts are still present in the ACS imaging. All 14 objects shown here exhibit obvious concentrations of stars at the targeted location of the candidate, further proof that these are real galaxies. WFPC2 images not only have shallower magnitude limits than ACS, but cover smaller regions on the sky, with a field of view of only ∼150 × 150 arcsec as compared to the larger 202 × 202 field of ACS. Although concentrations are obvious in WFPC2 chip 3, the stellar populations spill out onto the other chips and outside the field of view. The smaller dwarfs observed with ACS are largely contained within the imaging; larger galaxies extend beyond the ACS images.

Many of these objects consist predominantly of RGB stars with perhaps a handful of stars in the AGB range. These two stellar types have similar distributions. A few objects exhibit evidence for a younger stellar population with upper MS and blue loop stars. These are found as concentrations in the core for some cases and scattered throughout the object in others. A wide range of galaxy sizes and stellar population densities is noted.

In Figures 17–25 we show the extinction-corrected CMDs of all expected members. Extinction in the ACS and WFPC2 bandpasses was determined using the reddening values from the dust maps of Schlegel et al. (1998) with the recalibration for extinction in different bandpasses by Schlafly & Finkbeiner (2011), as provided in their Table 6 assuming a Fitzpatrick (1999) reddening law with R_V = 3.1. The measured location of the TRGB is indicated with a broken solid line while dashed lines denote the uncertainty in the tip location. Table 2 provides TRGB F814W magnitude measurements, along with the extinction-corrected values, associated tip absolute magnitudes calculated as a function of TRGB color, and the resultant distance moduli and calculated distances. In addition, we correct our previous estimates of absolute r' magnitude (see Paper 1) for these new distance measurements. Uncertainties in extinction are not taken into account, so total uncertainty in distance measurements is expected to be larger.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 2. Distances and Photometric Properties of M81 Group Members

Galaxy	Cam	$r^{\prime }_{{\rm cor}_{s}}$a	μ○a	〈μe〉a	EB − Vb	TRGBmeas	TRGBo	(V − I)tip	Mtipc	DM	D	$M_{r^{\prime }_{{\rm cor}_{s}}}$d	MIe	Type
			r'	r'		F814W			F814W		(Mpc)
d0926+70	W	17.9	26.0	26.2	0.19	$23.92^{0.12}_{0.15}$	$23.63^{0.12}_{0.15}$	1.00 ± 0.07	−4.03	$27.66^{0.12}_{0.15}$	$3.4^{0.2}_{0.2}$	$-9.8^{0.5}_{0.5}$	−11.4	dI/dSph
d0934+70	W	18.0	26.2	26.4	0.25	$23.77^{0.70}_{0.08}$	$23.38^{0.70}_{0.08}$	1.00 ± 0.07	−4.02	$27.40^{0.70}_{0.08}$	$3.0^{1.1}_{0.2}$	$-9.4^{0.5}_{0.9}$	−10.5	dSph
d0939+71	A	18.5	25.5	26.1	0.04	$23.83^{0.32}_{0.22}$	$23.77^{0.32}_{0.22}$	1.02 ± 0.03	−4.10	$27.87^{0.32}_{0.22}$	$3.7^{0.5}_{0.4}$	$-9.4^{0.6}_{0.4}$	−9.6	dSph
d0944+69	A	21.1	25.9	26.7	0.09	$23.93^{0.37}_{0.19}$	$23.79^{0.37}_{0.19}$	1.05 ± 0.05	−4.10	$27.89^{0.37}_{0.19}$	$3.8^{0.6}_{0.3}$	$-6.8^{0.5}_{0.6}$	−9.1	dSph
d0944+71	A	15.2	23.4	23.9	0.03	$23.60^{0.06}_{0.06}$	$23.55^{0.06}_{0.06}$	1.13 ± 0.07	−4.08	$27.63^{0.06}_{0.06}$	$3.4^{0.1}_{0.1}$	$-12.4^{0.8}_{0.8}$	−13.2	dSph
d0955+70	W	17.9	26.6	27.1	0.16	$23.92^{0.35}_{0.31}$	$23.67^{0.35}_{0.31}$	1.05 ± 0.07	−4.02	$27.69^{0.35}_{0.31}$	$3.4^{0.6}_{0.4}$	$-9.8^{0.6}_{0.6}$	−10.6	dSph
d0958+66	W	14.7	21.6	22.3	0.07	$23.98^{0.05}_{0.05}$	$23.88^{0.05}_{0.05}$	1.00 ± 0.07	−4.03	$27.91^{0.06}_{0.06}$	$3.8^{0.1}_{0.1}$	$-13.2^{0.5}_{0.5}$	−13.7	BCD
d0959+68f	A	16.2	26.0	25.5	0.08	$24.13^{0.15}_{0.15}$	$24.01^{0.15}_{0.15}$	1.10 ± 0.10	−4.09	$28.10^{0.15}_{0.15}$	$4.2^{0.3}_{0.3}$	$-11.9^{0.8}_{0.8}$	−11.8	dI/tdl
d1006+67	W	18.0	26.2	26.5	0.06	$23.89^{0.14}_{0.12}$	$23.79^{0.14}_{0.12}$	1.12 ± 0.04	−4.01	$27.80^{0.14}_{0.12}$	$3.6^{0.2}_{0.2}$	$-9.8^{0.5}_{0.5}$	−10.3	dSph
d1012+64	A	14.5	21.5	22.3	0.02	$23.77^{0.05}_{0.05}$	$23.74^{0.05}_{0.05}$	1.10 ± 0.04	−4.09	$27.83^{0.06}_{0.06}$	$3.7^{0.1}_{0.1}$	$-13.3^{0.5}_{0.5}$	−13.9	BCD
d1014+68	W	18.5	27.2	27.3	0.05	$23.97^{0.18}_{0.20}$	$23.89^{0.18}_{0.20}$	1.00 ± 0.07	−4.03	$27.92^{0.18}_{0.20}$	$3.8^{0.3}_{0.3}$	$-9.4^{0.5}_{0.5}$	−10.5	dSph
d1015+69	A	19.1	25.5	26.0	0.05	$23.91^{0.14}_{0.12}$	$23.83^{0.14}_{0.12}$	1.00 ± 0.04	−4.11	$27.94^{0.14}_{0.12}$	$3.9^{0.3}_{0.2}$	$-8.8^{0.6}_{0.6}$	−9.5	dSph
d1028+70	W	15.5	22.4	23.3	0.04	$23.96^{0.05}_{0.05}$	$23.90^{0.05}_{0.05}$	1.03 ± 0.05	−4.02	$27.92^{0.06}_{0.06}$	$3.8^{0.1}_{0.1}$	$-12.4^{0.5}_{0.5}$	−13.1	BCD
d1041+70	W	18.5	26.1	26.2	0.06	$23.91^{0.14}_{0.16}$	$23.82^{0.14}_{0.16}$	1.03 ± 0.08	−4.02	$27.84^{0.14}_{0.16}$	$3.7^{0.2}_{0.3}$	$-9.3^{0.5}_{0.5}$	−10.7	dSph

Notes. ^aMagnitude and surface brightness measurements from CFHT/MegaCam r' imaging (Chiboucas et al. 2009). ^bFrom Schlegel et al. (1998) dust maps; this was converted to extinction in each bandpass using the Schlafly & Finkbeiner (2011) recalibration of the dust maps. ^cAbsolute magnitude of the TRGB using the calibration from Rizzi et al. (2007). ^dIntegrated magnitude from the curve of growth fitting with cumulative Sersic function based on MegaCam imaging; extinction-corrected (Chiboucas et al. 2009). ^eEstimated magnitudes assuming model stellar populations for each galaxy; see the text. Uncertainties are typically 0.4 mag for ACS data and 0.6 mag for WFPC2. ^fTRGB measurements may apply only to tidal stream stars and not to the star-forming concentration proposed to be a tidal dwarf.

Download table as:  ASCIITypeset image

The largest uncertainty in TRGB measurements is for object d0934+70. This galaxy appeared to be strongly affected by galactic cirrus in MegaCam imaging, and is in a region with high extinction according to the dust maps of Schlegel et al. (1998). The RGB is poorly populated, leading to the large uncertainty in TRGB measurement. The RGB is also strangely shaped, displaying what appears to be color/metallicity shifts along the extension of the RGB. This could perhaps be caused by the low number statistics and large photometric errors, or by high and variable extinction spatially across this object. The maximum likelihood best-fit distance is 3.0 Mpc, but an extinction over-correction would produce an underestimate in the measured distance. We believe this may be the case. The large 1σ uncertainties place the distance between 2.8–4.1 Mpc. In comparison to the CMDs generated from "field" stars in regions on the detectors away from the targeted objects, there are significantly more stars associated with this object and these are arranged largely along the expected location for an RGB. Because of this, and coupled with the object's spatial location in the group, we consider it an M81 Group member, albeit with an uncertain distance measurement.

One object, d0959+68, was observed with both WFPC2 and ACS. We do obtain a TRGB measurement for this object from ACS data, with corresponding distance 4.2 Mpc. However, we consider this measurement uncertain due to the likely contamination of the RGB region with AGB stars. The majority of stars associated with this object are blue, young, MS, and blue loop stars. Since this object lies right along the tidal stream between M81 and NGC 3077, and given its young population and lack of a built-up RGB, we consider this object to be a tidal dwarf formed along this bridge of material likely having a distance similar to that of these two giants, ∼3.7–3.9 Mpc.

Other objects, d0926+70, d0955+70, d1006+67, d1014+68, d1041+70, d0939+70, d0944+69, and d1015+69, have uncertainties in distance ranging from 200 kpc to 1 Mpc. The TRGB measurements themselves place all of these objects near M81 at 3.69 Mpc. The large uncertainties are due to the nature of these objects. They all appear to be small dwarfs with very small stellar populations. With such poorly populated RGBs, the uncertainties in the distance measurements will be large. We expect these dwarf galaxies are all associated with the M81 Group.

Finally, four objects have small uncertainties in the tip magnitude due to the much more populated RGBs. Three of these were considered to be BCD galaxies in our first paper and are now established to be associated with the M81 Group. These are small objects with reasonably large and concentrated stellar populations consisting of both old stars in a well developed RGB and younger MS, blue loop, and AGB stars.

Thus, based on the distance measurements from the CMDs along with supporting evidence from concentrations of resolved stars at the targeted locations of the candidates, we confirm a total of 14 new M81 companion dwarf satellites. Of the original 22 candidates, 12 were expected to be members based on the degree of resolution of the stellar population in r' MegaCam imaging. Of these, we find that 11 are bona fide members of the M81 Group. Of 6 candidates that we thought were galaxian in nature but were uncertain members due to a lower degree of resolution, we find that 3 are indeed M81 Group members while 3 are non-members and likely more distant dwarf galaxies. Of the final 4 candidates considered to be possible artifacts or distant galaxy clusters, none prove to be group members. Nine of these confirmed M81 Group members exhibit predominantly RGB sequence stars in their CMDs displaying little or no evidence for any AGB, blue loop, or MS stars. These are the dSph galaxies. A tenth object may also contain a small number of younger stars. The remaining 4 dwarfs all exhibit more significant younger populations in their CMDs. We display the projected and three-dimensional (3D) distribution of 22 previously known and 14 new M81 Group members within the region of our survey in Figure 26. Table 3 lists the projected and physical separations of each group galaxy from M81. CMDs, spatial distributions, and color figures for each of these galaxies can also be found in the Extragalactic Distance Database (EDD), at http://edd.ifa.hawaii.edu, by selecting the catalog CMDs/TRGB.

Zoom In Zoom Out Reset image size

Table 3. Projected and Physical Separation of Group Galaxies from M81

Galaxy	Projected	Physical	$M_{r^{\prime }}$	Re	Type
	(deg)	(Mpc)		(kpc)
M81	0.00	0.00	−20.6	1.45	Sb
KDG61	0.49	0.043	−13.3	0.68	dSph
BK5N	1.16	0.071	−11.7	0.40	dSph
N2976	1.38	0.094	−18.0	0.77	Sc/pec
IKN	1.32	0.095	(−12.4)		dSph
M82	0.62	0.099	−19.9	1.05	Irr
FM1	0.98	0.101	−11.7	0.39	dSph
KDG64	1.63	0.119	−13.2	0.40	dSph
F8D1	1.87	0.127	−13.2	1.22	dSph
d0944+69	1.01	0.127	−6.8	0.09	dSph
Garlnd	0.82	(0.135)			pec/tdl
d1014+68	1.77	0.156	−9.4	0.37	dSph
HolmIX	0.18	(0.156)	−13.9	0.87	dI/tdl
KK77	1.64	0.158	−13.0	0.68	dSph
d1006+67	2.14	0.163	−9.8	0.32	dSph
d0939+71	2.63	0.173	−9.4	0.27	dSph
KDG63	2.67	0.177	−13.0	0.44	dSph
d0958+66	2.24	0.179	−13.2	0.22	BCD
N3077	0.78	0.181	−17.8	0.70	Irr
DDO78	3.18	0.227	(−12.8)	0.43	dSph
d1028+70	3.10	0.228	−12.4	0.25	BCD
d1015+69	1.75	0.237	−8.8	0.16	dSph
A0952	0.27	(0.238)			pec/tdl
d0955+70	1.34	0.260	−9.8	0.31	dSph
d1041+70	4.12	0.265	−9.3	0.19	dSph
HS117	3.01	0.308	−12.1	0.38	dI
I2574	3.04	0.322	−17.7	2.71	SABm
DDO82	3.39	0.329	−15.1	0.65	Im/BCD
d0944+71	2.59	0.335	−12.4	0.30	dSph
d1012+64	5.24	0.337	−13.5	0.25	BCD
d0926+70	2.89	0.345	−9.8	0.24	dI/dSph
HolmI	2.48	0.363	−14.6	1.30	dI
BK3N	0.18	(0.475)	−10.3	0.14	dI/tdl
BK6N	4.81	0.488	−11.7	0.27	dSph
d0959+68	0.54	(0.508)	−11.9	0.54	dI/tdl
d0934+70	2.19	(0.705)	−9.4	0.22	dSph

Notes. Values listed in parentheses are highly uncertain, especially for suspected tidal dwarfs where distances measured from RGB stars may apply to nearby galaxies/tidal streams rather than the tidal dwarf itself.

Download table as:  ASCIITypeset image

Figure 27 shows a projection of the 3D distribution of all known galaxies within a cube centered on M81 and extending 600 kpc on the cardinal axes. The plot is in supergalactic Cartesian coordinates and the orientation is chosen so measured distances (and their errors) are in the line of sight. This orientation allows a projection of the Schlegel et al. (1998) dust map to be superimposed as a reminder that a band of cirrus crosses the region. Revised (Schlafly & Finkbeiner 2011) selective to total extinction E(B − V) values reach 0.22 at the location of one of our galaxies, although the values are below 0.1 for 85% of cases, with a median of 0.06.

Zoom In Zoom Out Reset image size

The gas-deficient systems are given emphasis in this plot, with the new dSph/non-star-forming galaxies colored red and previously known gas-deficient systems colored orange. Normal giant and dwarf galaxies with both young and old stars and detected H i (Begum et al. 2008; Roychowdhury et al. 2012) are identified by open symbols, with the three newly discovered cases given emphasis. Five tidal dwarfs (young stars but no convincing evidence of old stars) are identified by green triangles. The measured TRGB distances for the tidal dwarfs are considered quite unreliable since the RGB stars in these fields may not be associated with the targets. All the tidal dwarfs are in close proximity to M81 in projection and we assume the M81 distance of 3.69 Mpc for them.

The projection chosen with Figure 27 is minimally influenced by errors in distance. We note an apparent flattening of the distribution of gas-deficient systems to the supergalactic equatorial plane. M81 itself lies just 40 kpc off this plane. The significance of this configuration can be evaluated from the histograms in Figure 28. The galaxy KKH57 lies outside the CFHT survey region so it will not be considered further. Of the 20 gas-deficient galaxies in the survey region only d1041+70 deviates substantially from the supergalactic mid-plane (or more precisely, a plane parallel to the supergalactic equator displaced 40 kpc to pass through M81). The rms deviation in supergalactic latitude (SGB) with respect to M81 is ±61 kpc for 19 galaxies excluding d1041+70. By contrast, the rms deviation in supergalactic longitude (SGL) with respect to M81 is ±123 kpc for the same galaxies. The difference between SGB and SGL scatter has a 95% probability of significance; that is, there is suggestive but not conclusive evidence for a physical flattening. The tentative postulation is that this sample on a scale of a few hundred kpc shares flattening about the same pole as the 50 times larger structure that it is embedded in, the Local Sheet (Tully et al. 2008).

Zoom In Zoom Out Reset image size

For the same 19 galaxies the scatter in distance is ±231 kpc. This value is the convolution of true scatter and measurement errors. If the true scatter in the line of sight is the same as the scatter in the SGL then measurement errors at the level of 5% are implied. This estimate of errors is in accordance with our expectation. This level of error in distances limits our ability to define the properties of a putative plane in the distribution of the M81 dwarfs. It can be supposed that such a plane is tilted to our line of sight. However, it would be necessary to have distances accurate to 1% to evaluate such a scenario, well beyond our present capability.

It is evident from Figure 27 that the distribution of galaxies with gas and young populations does not manifest the flattening seen by the gas-poor systems. Discounting M82 and NGC 3077, which are directly engaged in an encounter with M81, the later-type systems are at greater projected radii from M81 than the early types. This difference is illustrated in Figure 28. It is a familiar pattern (Grebel 1997; Tully & Trentham 2008). The median and mean projected separations from M81 for early types within the CFHT survey region are 122 and 133 kpc (149 and 162 kpc statistically deprojected), while the median and mean for late types are 158 and 169 kpc (193 and 206 kpc deprojected). Beyond the CFHT survey region, all known galaxies associated with the M81 Group except KKH57 are late types. In the common picture, minor galaxies that have made multiple crossings within a group halo have lost their gas through tidal disruption, stripping, or starvation. The outer extent of their distribution roughly coincides with the radius of "second turnaround" (Tully 2010), which is a direct function of the mass of the parent halo. The gas-bearing systems, by contrast, are likely more recent arrivals. They may be on first infall or they may have high orbital angular momentum that has kept them from harm's way.

The apparent flattening of the distribution of early types, although not overwhelmingly convincing, is given credence by similar phenomena seen more locally. The most remarkable case to date has been discussed by Ibata et al. (2013). Half of the small companions to M31 lie in a very thin (±13 kpc) disk with coherent kinematic properties. This disk is askew from the supergalactic plane but oriented such that the Milky Way lies in the same plane. Meanwhile most of the early-type companions to the Milky Way appear to lie in a plane (Lynden-Bell 1976; Metz et al. 2008; Pawlowski et al. 2012). This plane is essentially orthogonal to the supergalactic plane although the long axis is aligned with the long axis of the Local Sheet (Tully et al. 2008), i.e., toward the Virgo Cluster. The past histories of these planes and two other apparent planes of Local Group satellites are discussed by Shaya & Tully (2013). In review, in the few instances around nearby major galaxies where we have information, in every case there is evidence that gas-poor companions lie in flattened distributions.

To determine centroids of the new objects, we used the geometric mean of the distribution of the resolved RGB stars (or MS stars in the case of d0959+68) in the galaxy field using the F814W ACS/WFPC2 data. Ellipticity and position angle around these centroids were then derived from second moments (Secker & Harris 1997).

To obtain structural parameter measurements for each galaxy observed with ACS, we perform stellar surface density profile fitting. Stellar number counts were extracted in successive elliptical apertures around each galaxy out to a radius of 40 arcsec. This was truncated if counts exhibited a significant increase in slope, chosen empirically to be >30%, attributed to contamination by an excess of field stars. The central measurement was also discarded. Aperture sizes for each galaxy depended on galaxy stellar density and were optimized to contain statistically significant counts over the background while also maintaining enough apertures to perform the fitting. Foreground contamination was estimated in these same sized apertures using stellar detections in the other ACS chip, and a mean foreground stellar count density was determined for each galaxy. This was then subtracted from the galaxy counts. For larger galaxies, the stellar component may extend over the entire ACS field of view, and thus this background correction may be a slight over-correction. The effect of this would be to produce size measurements that are too small, and indeed for larger galaxies, these size measurements are almost certainly underestimates. Smaller galaxies are less likely to extend very far onto the other chip. However, these galaxies are more likely to suffer from statistical variance affecting the low number counts.

We then fit both the differential count density with a Sersic function and the curve of growth with a cumulative Sersic function. The generalized Sersic function is defined as

$$\begin{equation} I(r) = I_{\circ } \ e^{-(r/r_{\circ })^{1/n}}, \end{equation} \tag{ 1 }$$

where n is the Sersic parameter. For n = 1, this reduces to an exponential profile which usually provides a good fit for disk and dwarf galaxies, and for n = 4.0 becomes a de Vaucouleurs profile. The cumulative Sersic function is then

$$\begin{equation} I(r) = 2\pi \sigma _{\circ }r_{\circ }^{2}n \cdot \gamma [2n,(r/r_{\circ })^{1/n}], \end{equation} \tag{ 2 }$$

where γ[a, x] is the Incomplete Gamma function (Jerjen 1995),

$$\begin{equation} \int _{0}^{x}\exp (-t)t^{a-1}dt. \end{equation} \tag{ 3 }$$

The fitting of this nonlinear function to the data was done using a Levenberg–Marquard algorithm (Press et al. 1992) which performs a χ² minimization that implements an inverse-Hessian method far from the minimum and switches to a steepest decent method as the minimum is approached. In Figures 29–34, we display both curve of growth and count density profile fits. Parameters were extracted from the best of these two fits for the six galaxies imaged with ACS. WFPC2 fields are so small that even the smaller galaxies fill the field and it is not possible to obtain a good estimate of the foreground contamination. For these galaxies, we therefore defer to the MegaCam size measurements from Paper 1.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

From these fits we directly obtain scale length and profile type (Sersic index n), where scale length in the generalized Sersic function is directly related to the half-light or effective radius through

$$\begin{equation} R_{\circ }^{1/n} = R_{e}^{1/n} / (2.3026 b_{n}) \end{equation} \tag{ 4 }$$

with the approximation b_n = 0.868n − 0.142 (Caon et al. 1993). This assumes that the mean brightness of the stars does not vary as a function of radius. It further assumes that we can use counts in the opposing chip as an approximation for the foreground correction. For the larger galaxies, where we may be over-correcting for this affect, we find that if no background correction is applied, d1012+64 and d0944+71 would be 1 and 2% larger, respectively. Results are listed in Table 4. In Figures 29–34, we also display contour plots of these galaxies on which we overlay ellipses with shapes and sizes corresponding to the measured PA, ellipticity, and effective radius for each galaxy.

Table 4. Structural Properties of M81 Group Members

Galaxy	Cam	CFHTa	HSTb	PACFHTa	PAHSTb	α	δc	Re, CFHTd	Re, CFHT	nCFHT	Re, HSTe	Re, HST	nHST
						(J2000.0)	(J2000.0)	(arcsec)	(kpc)		(arcsec)	(kpc)
d0939+71	A	0.05	0.15	−80.9	3.1	9:39:16.01	71:18:41.00	15.3	0.27	0.66	8.9 ± 0.9	0.16 ± 0.02	0.75
d0944+69	A	0.09	0.04	−9.7	34.3	9:44:22.49	69:12:37.94	4.9	0.09	0.78	7.2 ± 1.3	0.13 ± 0.05	1.28
d0944+71	A	0.20	0.11	−27.4	3.6	9:44:34.37	71:28:55.60	18.3	0.30	0.61	21.4 ± 0.4	0.35 ± 0.02	0.67
d0959+68	A	0.10	0.22	35.1	73.5	9:59:34.90	68:39:25.78	26.6	0.48f	0.21	20.8 ± 2.8	0.37 ± 0.08f	0.34
d1012+64	A	0.15	0.14	17.0	4.4	10:12:48.41	64:06:26.21	13.9	0.25	0.88	14.9 ± 0.8	0.31 ± 0.02	0.67
d1015+69	A	0.07	0.17	−49.7	−41.0	10:15:06.89	69:02:13.81	8.5	0.16	0.68	9.6 ± 1.6	0.18 ± 0.05	0.99
d0926+70	W	0.18	0.26	−23.9	0.3	9:26:27.94	70:30:18.79	14.3	0.24	0.42
d0934+70	W	0.07	0.15	63.5	−82.9	9:34:03.22	70:12:58.32	15.3	0.22	0.42
d0955+70	W	0.04	0.08	86.8	84.8	9:55:14.14	70:24:25.96	18.6	0.31	0.61
d0958+66	W	0.36	0.23	−6.9	−7.5	9:58:48.74	66:50:57.34	11.6	0.22	0.71
d1006+67	W	0.05	0.17	86.1	50.2	10:06:46.80	67:11:59.78	18.1	0.32	0.44
d1014+69	W	0.01	0.08	−8.9	−34.6	10:14:55.80	68:45:29.71	19.7	0.37	0.29
d1028+70	W	0.20	0.20	−54.8	−67.0	10:28:39.98	70:13:59.84	13.4	0.25	0.88
d1041+70	W	0.18	0.28	26.0	26.0	10:41:18.14	70:09:13.75	10.9	0.19	0.30

Notes. ^aMeasurements based on the second moments of the light distribution in MegaCam r' imaging (Chiboucas et al. 2009), including both resolved and unresolved components. ^bMeasurements based on the second moments of the distribution of resolved RGB stars only. ^cCentroid based on the resolved RGB population in HST imaging. ^dFrom integrated Sersic profile fits to curves of growth in MegaCam r' imaging (Chiboucas et al. 2009). Values listed are the geometric mean of the semimajor and semiminor axes. ^eBased on Sersic profile fits to number counts in HST F814W-band imaging of resolved stars likely to be galaxy members. This assumes the mean brightness of stars is independent of radius, as expected for RGB stars. Measurements are for data taken with ACS only. Radii listed are the geometric mean of the major and minor axes. ^fFor D = 3.69 Mpc. Actual TRGB measured distance is 4.2 Mpc.

Download table as:  ASCIITypeset image

As we found with the CFHT imaging, the majority of the galaxies are best fit with Sersic index n < 1.0. It was determined that the low n fits were primarily driven by large cores, typical of dwarf galaxies, whereas in the outer radii, the surface brightness profile tends to drop exponentially.

For d0959+68, we fit only the blue stellar component which we take to be the tidal dwarf (see Section 3.5.5). However, these young stars are found in star-forming clumps, and the stellar surface density is not well fit with a Sersic profile. This is true in the case of the measurement made with CFHT imaging as well.

To estimate uncertainties in the size measurement, we repeat the process, using different regions of the images for the foreground correction. The 1σ uncertainties quoted in the table come from the variance in size measurements from a total of 20 profile fits. As can be seen in Table 4, the uncertainties are large for the fainter dwarfs, reaching as high as 12.5% in apparent size for the smallest dwarf d0944+69 (or 33% for physical size when including distance uncertainties). The largest and best populated objects observed with ACS, d0944+71 and d1012+64, have much smaller 6% uncertainties in physical size.

Differences from CFHT structural measurements can be attributed to large uncertainties in measurements for faint, low surface brightness objects due to the shallow depth of CFHT/MegaCam imaging, contamination by non-member stars in the CFHT imaging data, uncertainties in the foreground correction for ACS data and small size of the HST fields, and the different methods used. In the case of CFHT imaging data, we used the low surface brightness light profile, just visible above the sky in most cases, while for HST imaging, we use the distribution of individual stars with colors and brightnesses corresponding to the RGB. Shape measurements are also expected to differ since measurements from CFHT imaging were based on the light distribution and will be strongly affected by brighter sources including foreground stars and luminous blue MS stars, while the HST measurements are based on the RGB star counts (with some amount of foreground contamination), each weighted equally.

In Figure 35, we plot $M_{r^{\prime }}$ versus log R_e for the newly confirmed members with late and early types distinguished by color, along with previously known M81 Group members. Points are based on the MegaCam imaging for which we have size and magnitude measurements for all M81 members within the survey region. The new BCDs are seen as the high surface brightness objects at magnitudes similar to previously known galaxies while the majority of the new detections lie at magnitudes fainter than $M_{r^{\prime }} > -10$. A surface brightness limit <27 mag arcsec⁻² for all new detections is apparent. Also shown in the plot are size measurements based on the ACS imaging for five galaxies. These are shown as crosses at the corresponding r' magnitude. The points are seen to shift slightly although insignificantly in most cases. Only for our smallest dwarf d0944+69 is the shift noticeably large. The two dashed vertical lines indicate the region of parameter space where few star clusters and galaxies tend to be found. The CFHT/MegaCam imaging measurement places d0944+69 within this gap while the ACS measurement inflates it to well within the typical size range for galaxies.

Zoom In Zoom Out Reset image size

Measuring the brightness of these galaxies from the HST data based on the diffuse resolved stellar populations is not straightforward. For WFPC2 data, only ∼2 mag are visible below the TRGB, with incompleteness and increased photometric errors setting in before the detection limit. ACS data are not much improved for this purpose, reaching only ∼3 mag below the TRGB. In addition, the WFPC2 fields are too small to fully contain the full spatial extent of the galaxies, and foreground contamination is not, in many cases, easily determined. This is true for the larger galaxies imaged with ACS as well. Nevertheless, we attempt to determine absolute F814W-band (I) magnitudes for these galaxies using the following method. For dSphs with pure old stellar populations, we use Padova model stellar LFs with age 12.5 Gyr and metallicity determined from the color of the RGB based on isochrones. Details on the models used for the LFs and isochrones are provided in Section 3.5. We assume these objects have single stellar populations and that the models provide a good match to the data. We then fit the observed RGB LF with a power law. The RGB slope is typically ∼0.3 but we allow it to range between 0.25 and 0.35 although the models permit a slightly smaller range, 0.3 ± 0.025. Integrating over this fitted LF from the TRGB down to the base of the RGB, we obtain an estimate of the total magnitude of the RGB population. Using the model LF, we then determine the fraction of total flux that is in the MS and horizontal branch versus RGB stars and apply this correction to the RGB magnitude. Finally, we make a correction for the fraction of the galaxy that falls off the detector. This method relies on being able to determine the slope of the RGB LF, but sparse RGBs extending only 1.5–2 magnitudes below the tip in the case of WFPC2 data and contamination from foreground stars, cause poor fits. For WFPC2 data, we therefore simply assume a slope of 0.3, as suggested by the models. The largest uncertainties in the total magnitude come from the uncertainties in power-law slope and area correction and together are typically 0.4 mag for ACS data and 0.6 mag for observations with WFPC2. The effect of using different metallicity and age models for an assumed old (>4 Gyr) single stellar population is less than ±0.1 mag.

For objects with multiple or younger populations (see Section 3.5) we additionally include these estimated contributions to the total magnitude. For d0959+68, we only include the contribution for young MS stars under the assumption that this component constitutes the tidal dwarf and scale a model LF with an age of 30 Myr with Z = 0.008 to the observed flux brightward of M_F814W = −2 to estimate the total flux. Objects with obvious multiple components are more challenging and we only provide very rough estimates for these by assuming the composite of 2–3 discrete populations. For d0926+70 we use a model with a single old 12.5 Gyr stellar population with Z = 0.001, under the assumption that any younger population present will contribute negligibly to the total magnitude. For the BCDs d1012+64 and d1028+70, we assume the populations are a composite of an old 12.5 Gyr stellar population with Z = 0.0015 and 0.001, respectively, and a young 40 Myr population with Z = 0.008 and 0.003, respectively. In these cases, the contribution from the young component is only 0.1 mag. The intermediate-aged population of d0958+66 cannot be ignored. We therefore factor in missing flux below the detection limit for an assumed three discrete populations: 12.5 Gyr with Z = 0.0008, 1.5 Gyr with Z = 0.002, and 60 Myr with Z = 0.004. The intermediate-age model is scaled in a similar way as is done for the young population using the total flux of visible AGB stars brightward of the RGB. The young and intermediate stars here contribute 0.3 mag to the total based on these models with a model-dependent uncertainty of less than 0.1 mag. These estimated total F814W magnitudes are listed in Table 2.

We investigate the stellar populations and evolutionary histories of these confirmed dwarf galaxies. Attempts were made to perform stellar evolution synthesis on the most populated objects observed with ACS. However, with the bright limiting depth of the observations providing access to only the brightest RGB and AGB and handful of upper blue MS stars, along with the degeneracy between age and metallicity at the level of the RGB, uncertainties proved to be very large. Because of this we prefer to keep the discussion more general until deeper data are acquired. We therefore instead use isochrones, overlaying model isochrones on the CMDs for each galaxy in order to determine the plausible range in ages and metallicities of the stellar populations. Padova isochrones (Marigo et al. 2008; Girardi et al. 2008) are used, which are transformed directly on the ACS/WFC and WFPC2 filter systems. From the web interface form, we chose bolometric corrections from Loidl et al. (2001), circumstellar dust calculations as in Groenewegen (2006), and a Chabrier lognormal IMF (Chabrier 2001). We also chose detailed AGB tracks from Marigo & Girardi (2007). Padova isochrones were chosen over other models largely due to the fact that they include stellar isochrones down to very young ages, and several of our galaxies show evidence for young MS populations. Comparisons using Dartmouth (Dotter et al. 2008) and Basti (Pietrinferni et al. 2004) isochrones found these to be systematically more metal poor than the Padova isochrones. For internal consistency, we therefore only include Padova isochrones in this paper.

For old stellar populations, the color of the RGB is more greatly affected by metallicity than age. In Figures 17–25 we investigate the old stellar population by overlaying isochrones extending up to the RGB phase for a constant 12.5 Gyr age, with metallicities Z = 0.0001, 0.001, 0.002, 0.004, 0.01. For many of the galaxies, the CMDs are well described by a pure old stellar population without any indication of intermediate-age AGB stars or younger MS and blue supergiant stars. The large spread in the RGB sequence due to photometric uncertainties makes it difficult to constrain the metallicities. For WFPC2 observed galaxies, because of photometric uncertainties the old RGB stars can be described by a wide range of metallicities typically between 0.0001 < Z < 0.002. Since the photometric errors for the ACS data are much smaller, it is possible to put tighter constraints on the metallicity of the old stellar population. We find for these galaxies that the majority of RGB stars tend to fall between 0.0001 < Z < 0.001 or between 0.001 < Z < 0.002.

We display the RGB metallicity Z versus total r' magnitude for our sample of new M81 dwarfs in Figure 36. The bars along the Z axis are not standard uncertainties, but rather the metallicity of the isochrones having assumed age 12.5 Gyr and colors within ±1σ of the mean RGB color at F814W = 24.5. The photometric uncertainties are expected to produce most of the observed spread in these RGBs, although in a couple cases a width slightly larger than can be accounted for by photometric uncertainties (d1012+64 and d0959+68) indicates possible intrinsic broadening and a real metallicity spread. The metallicity ranges are provided in Table 5. For M81 galaxies not part of our HST follow-up, we use ACS data from the ANGST survey (Dalcanton et al. 2009) taken in similar passbands and reaching about 1 mag deeper. Using the same method we estimate the allowed metallicity range of the RGB for these galaxies. Early- and late-type galaxies are denoted by color, and as has been noted previously (e.g., Mateo 1998), separate relations are found for galaxies that are still forming stars and for those that are not, offset in the sense that late-type galaxies are more metal poor at a given luminosity. Given the large uncertainties, the differences are not significant in our data.

Zoom In Zoom Out Reset image size

Table 5. Summary of Stellar Populations

Name	Instr.	$M_{r^{\prime }}$	Re	Type	Age Young	Age Int.	AGB/RGBa	ZRGBb	Notesc
			(kpc)		(Myr)	AGB (Gyr)	(Gyr)
d1012+64	A	−13.3	0.25	BCD	<60	0.5–4	5–8	0.0004–0.0018	MH i = 1.04 × 106 M☉, vr = 190 km s−1
d0958+66	W	−13.2	0.22	BCD	⩾60	0.5–4	∼3	0.0001–0.003	MH i = 0.73 × 106 M☉, vr = 90 km s−1
d1028+70	W	−12.4	0.25	BCD	<60	1–4	5–9	0.0001–0.0023	MH i = 1.3 × 106 M☉, vr = −69 km s−1
d0944+71	A	−12.4	0.30	dSph	(⩾150)	(∼2)	old	0.0005–0.002	MH i < 0.31 × 106 M☉
d0959+68	A	−11.9	0.48	dI/tdl	<80			0.0001–0.0035
d0926+70	W	−9.8	0.24	dI/dSph	(⩾100)	⩾0.5	4–7	0.0001–0.0023	MH i < 0.31 × 106 M☉
d0955+70	W	−9.8	0.31	dSph				0.0001–0.003
d1006+67	W	−9.8	0.32	dSph				0.0004–0.004
d0934+70	W	−9.4	0.22	dSph		(∼1)		0.0001–0.004
d0939+71	A	−9.4	0.27	dSph				0.0002–0.001
d1014+68	W	−9.4	0.37	dSph		(∼1)		0.0001–0.003
d1041+70	W	−9.3	0.19	dSph				0.0001–0.003
d1015+69	A	−8.8	0.16	dSph				0.0001–0.001
d0944+69	A	−6.8	0.09	dSph				0.0003–0.0018

Notes. Values listed for ages and metallicities are approximate ranges determined from Padova model isochrones; see the text. Prominent young or intermediate populations are denoted in boldface, while more tentative identifications are listed in parentheses. ^aComposite age estimate for the older populations; see the text. These estimates are best used as relative age indicators. ^bThe range of metallicities which span the ±1σ width of the RGB, assuming a 12.5 Gyr age for the old stellar population. ^cH i measurements taken from Roychowdhury et al. (2012).

Download table as:  ASCIITypeset image

To assess ages, we can again use isochrones to investigate the existence of young, intermediate, and old populations, and can also compare the relative fractions of different populations. The older stellar populations (>1 Gyr) visible in these CMDs are formed by RGB and AGB stars. Since AGB stars evolve faster than RGB stars, the ratio of the two populations provides information about the ages of the older stars. Following Jacobs et al. (2011), we take the ratio of RGB and AGB stars as stars 2 mag below and 2 mag above the TRGB, respectively, which gives a crude estimate of the ratio of intermediate to old stars. We use the same color cuts for the ACS data that they use to extract counts from simulated CMDs for a range of metallicities and ages and compare our results to their simulations for single stellar populations as shown in Figure 37. For our WFPC2 data, due to the larger uncertainties, we broaden the color constraints defining the RGB by a factor of 1.5, consistent with the increased magnitude errors. Because the model values were extracted using older Bertelli et al. (1994) and Girardi et al. (2000) evolutionary tracks while we use more recent models to estimate metallicities, and as these are strictly model estimates, the reader is cautioned against directly reading off ages from the plot based on the measured population ratio and metallicity. However, we expect the trends to be valid and thus the ratio to be useful for estimating the relative ages of our galaxies. Of course, these results also assume a single age for the old stellar population, whereas some of these objects may have hosted continual or at least multiple bursts of star formation.

Zoom In Zoom Out Reset image size

3.5.1. BCDs: d1012+64, d0958+66, and d1028+70

3.5.2. dI/dSph Transition: d0926+70

3.5.3. dSph: d0944+71

3.5.4. Classical and Ultra-faint dSphs

3.5.5. Tidal Dwarf: d0959+68