3 Results and discussion

3.1 Morphology and total fluxes

The B (Fig. 1), R, and J images show a mostly regular galaxy of elliptical shape. Only few brighter and extended knots in a region to the South-East of the center can be seen. The H${\alpha}$ (Fig. 2) image reveals that these knots are HII regions. To the limits of the images, one bright and 4 fainter, smaller HII regions were identified. The largest one has a diameter of about 9.3$^{\prime\prime}$(0.25 kpc). These few HII regions are clumped in an area of no more than 25$^{\prime\prime}$ (0.65 kpc) diameter. This blob of HII regions is off-set from the center by 11$^{\prime\prime}$(0.31 kpc) or about one scale length of the stellar light distribution. Very little diffuse H${\alpha}$ emission outside the HII regions is visible, but the image which was taken for selecting proper long slit settings might have an exposure time which was too short to detect very faint structures like those describes for example in the case of NGC 4449 by Bomans et al. (1997). The long slit spectra along the major axis of UGC 685 cross some of the identified HII regions and indeed reveal the typical emission line pattern of HII regions. Outside these HII regions, but still well inside the low surface brightness envelope, sources with the shape of the point-spread function are visible. These objects trace the resolved brightest stars - mostly supergiants - in UGC 685 even though it is easily possible that some of these "stars'' may be in reality composed objects like blends of stars or stellar clusters.

A mask was constructed at the 25.0 mag/$\Box$$^{\prime\prime}$ level in B and the flux in B and R inside this isophote was determined, yielding B = 14.55 and B-R = + 0.97 (Table 2). Errors are less than 0.1 and are mostly of systematic nature. The B magnitude is in good agreement with those in the literature (see Schmidt & Boller 1992a). The absolute values derived from these observed magnitudes were computed using the distance estimate of 5.5 Mpc from chapter 3.4 and a correction for galactic foreground extinction as given by Burstein & Heiles (1982). No attempt was done to correct for internal reddening. As already mentioned, the J frames exclude a small part of the faint eastern outskirt, less than 4% of the total area. These outskirts contribute little light, nevertheless the derived J magnitude of 13.00 is more like a good lower limit to the total flux.

I measured the flux in the continuum-free H${\alpha}$ image and found a flux of $1~10^{-13}~{\rm erg~s^{-1}~cm}^{-2}$ which converts to an absolute flux in the H${\alpha}$ line as given in Table 2 (using D = 5.5 Mpc). The Poisson error of this value is small, less than 1%, but almost negligible compared to the error of the zero point calibration of about 20%. Hunter et al. (1994) observed H${\alpha}$ line fluxes for a sample of late-type galaxies, mostly dwarf irregulars. In comparison to this sample, the H${\alpha}$ line flux of UGC 685 is normal for an irregular dwarf of its total magnitude and indicates a relatively low recent star formation rate (see below).

3.2 Surface photometry

Figure 3 shows the surface brightness profile, the radial color profil and the variation of the ellipticity with major axis for B and R. UGC 685 has a rather elliptical shape with a mean ellipticity value of 0.33 for both colors. The ellipticity varies only slightly with radius while the (not shown) position angle increases steadily by the small, but significant amount of 20 degree from 10 to 70$^{\prime\prime}$. The surface brightness profile can be traced out to about 70$^{\prime\prime}$ in B and 65$^{\prime\prime}$ in R. It is interesting to note that in neutral hydrogen, the profile of UGC 685 was traced to a 2.6 times larger radius than here in the optical (see Hoffman et al. 1996 for the HI data). In J, the galaxy was traced down to 23.1 mag/$\Box$$^{\prime\prime}$and to a distance of 35$^{\prime\prime}$. Within this more limited range, the ellipticity is essentially the same (0.32) as in the CCD color bands.

In all three colors, the profile can be well described by an exponential law outside a central region of about 10$^{\prime\prime}$. The central profiles are significantly flatter. One should remember that in the very center, several knots (HII regions, stars, clusters) were removed, therefore it is difficult to trace the profile there. Outside 10$^{\prime\prime}$, I fitted an exponential brightness profile to the data and got a scale length of 12.4$^{\prime\prime}$to 12.7$^{\prime\prime}$ (Table 2). Thus, the scale length of the stellar distribution is 10 times smaller than the one of the HI distribution (Hoffman et al. 1996). The obtained parameters, already converted to metric values with the distance estimated below, are given in Table 2. They are quite normal for a galaxy of this size and luminosity. The central surface brightnesses are relatively high for an irregular dwarf with only moderate signs of recent star formation. A color profile was calculated (Fig. 3). Most of the body has the same color while the central part is slightly bluer. B-J, albeit having a larger systematic error, shows no significant variation with radius.

The exponential law parameters were used to calculate asymptotic total magnitudes of 14.11, 13.14, and 12.60 in B, R, and J, respectively.

H${\alpha}$ emission was only detected in a small area near to the center of UGC 685, slightly off towards the south-east. Thus, the recent places of star formation activity show a strong asymmetry in the angular distribution and are well localized. I measured the H${\alpha}$ surface brightness as a function of radius and averaged over the azimuth (for a better comparison with similar profiles for other irregular galaxies presented by Hunter et al. 1998). Naturally, this radial H${\alpha}$ profile is only valid for the small sector (Fig.  2). Outside, the H${\alpha}$ surface brightness is below the detection limits (about $1.5~10^{-16}~{\rm erg~s^{-1}~cm}^{-2}$). Assume that the H${\alpha}$ flux is a good measure of the total number of ionizing photons emitted in a galaxy. This number can be compared to the photons expected from massive stars for a given initial mass function (IMF). Taking into account the lifetimes of the massive stars, a formation rate of hot stars can be derived and extended to a total star formation rate (SFR) by extrapolating to all stars with the IMF. Hunter & Gallagher (1986, see also Gallagher et al. 1984) used a Salpeter function to establish a conversion formula which I used to transform the observed H${\alpha}$ surface brightness of UGC 685 into a star formation rate per ${\rm pc}^{2}$. A distance of 5.5 Mpc was used. Figure 4 show the radial profile. Given the uncertainty in the H${\alpha}$ flux calibration, the only secure feature is the off-center peak in the recent specific star formation activity, and an overall trend that the recent star formation activity drops to larger radii. The shape of the radial distribution of the recent star formation activity in UGC 685 is similar to the one observed in Sex A and IC 1613 (Hunter et al. 1998), even though its overall value is lower than in those two dwarfs. As discussed in more detail by van Zee et al. (1998), an azimutal average of the local SFR (as measured by the H${\alpha}$ flux) can be misleading for irregular dwarf galaxies where the star formation often takes place at only a few places. UGC 685 is obviously a good example for this case. Using again the derivation of Hunter & Gallagher (1986), the total H${\alpha}$ flux from chapter 3.1 can be transformed into an average SFR over the last $\sim 10^7$ yr. I derived 0.003 $M_{\odot}$ yr^-1 with an error based on the measurement alone of about 20%. Here again, the result depends on the applied distance value of 5.5 Mpc.

  

Figure 4: The radial distribution of the recent star formation activity as derived from the H${\alpha}$ flux applying the transformation relation between flux and star formation rate as given by Hunter & Gallagher (1986). A distance of 5.5 Mpc was assumed. This curve is only valid for the sector where HII regions are visible, outside this sector the star formation rate is below the detection limit

The radial color gradient shows no indication that the average SFR varies strongly with location (Fig. 3). A B-R color map shows a quite homogeneous distribution with only small (5% level) deviations in the central area, where recent (and localized) star formation is already indicated by the HII regions.

3.3 Resolved stellar population

A total of 209 stars are detected in both (B and R) frames, down to a limiting magnitude of 24 in both filters. These stars are identified in Table 3. Fainter than 22.5 in R and 23.0 in B, the photometric errors as estimated by the DAOPHOT PSF fitting algorithm are larger than 0.1. 73 of these stars are projected on or belong to UGC 685. As the CCD frames are by far larger than the galaxy, I can determine the surface density of fore- and background objects with stellar PSF. A statistical correction of 5.1 stars was calculated which leaves about 68 stars belonging to UGC 685. The color distribution of the field stars outside UGC 685 (see Fig. 5) shows that most of the contamination contributes in the red while practically all detected blue objects (B-R < 0.5) belong to the young population in UGC 685.

Artificial star tests were applied inside DAOPHOT to estimate the completeness of the star detection especially in the magnitude range of the supergiants of UGC 685. 50 stars were added in 1 magnitude bins in several experiments for each of these bins. The recovery rate indicates that the completeness starts to drop from almost 100% at 21.9 (22.7) in R (B) and the 50% completeness magnitudes are about 22.9 (23.7) in R (B).

  

Table 3: Stars found in B and R on the UGC 685 frames. Given are a running number (from B frame), relative frame coordinates in arc sec where +x increases towards East and +y increases towards South, the B magnitudes and their errors according to DAOPHOT (which does not include systematic errors of the calibrations) followed by the same photometric informations for R, and finally B-R

 

Table 3: continued

 

Table 3: continued

Figure 5 shows the color-magnitude diagram of the UGC 685. The left part shows all 209 detected points sources and indicates the 50% completeness limit while the right panel presents only those 173 stars where the photometric errors are less than 0.2 in B and R. The open symbols indicate the 136 sources which belong to the field around UGC 685 while the filled dots indicate the 73 sources in UGC 685. 56 of the UGC 685 objects have photometric errors less than 0.2 in both bands.

A comparison with Geneva evolutionary tracks (Fig. 6) as well as with the Padua isochrones (Meynet et al. 1994; Bertelli et al. 1994) indicates that several blue supergiants and fewer yellow supergiants have been resolved in the mass range 15 to 60 $M_{\odot}$ (ZAMS). Given the higher field contamination, is difficult to detect a red supergiant population. Again, one should remember that some of these objects, especially of the blue ones, might be unresolved clusters. Having only one color in most cases, clusters (or blends) can not be distinguished. Naturally, clusters of blue stars can be brighter than individual stars. Contrary, HII regions have been identified unambiguously with the H${\alpha}$ image.

  

Figure 5: Left: B-R, B color-magnitude diagram of the 209 stars in the field of UGC 685. The 136 stars outside the galaxy are marked as circles, those 73 of UGC 685 (or projected on UGC 685) are marked by filled dots. The dotted line indicates the 50% completeness limit (see text). Right: The panel shows only those 173 sources which have photometric errors in B and R less than 0.2. 56 of these sources belong to UGC 685. UGC 685 clearly shows a population of blue and yellow supergiants pointing to recent star formation activity. Some red supergiants seem be present too, but are harder to disentangle from the field confusion

  

Figure 6: B-R, B color-magnitude diagram of 56 stars of UGC 685. Stars of the surrounding field and those with errors larger than 0.2 are excluded. The observations are corrected for the foreground reddening of the Milky Way. Geneva tracks for 15, 20, 25, 40, and 60 $M_{\odot}$ stars of a metallicity of Z = 0.001 are shown. The tracks were taken from the http version of Schaller et al. (1992) and converted from mass, temperature, and luminosity to the observers frame with Kurucz (1992) stellar atmospheres (where the tables with log g, T, and colors are taken again from the web) and finally shifted to a distance of 5.5 Mpc

3.4 The distance of UGC 685

The brightest blue supergiants (BBSG's) have a long tradition as distance indicators (K&T, R&RR). It is usual to determine the mean of the three brightest blue supergiants, B₃, and apply a calibration as provided by K&T and R&RR to derive the distance modulus of the considered galaxy. The detailed analysis of R&RR clearly shows that one is limited to an accuracy of 0.8 mag at best. Thus, B₃ can give only a first estimate of the true distance. This still excludes uncertainties which are introduced by the selection of the candidates as interlopers from the fore- and background as well as unresolved blends or clusters. Furthermore, the luminosity of the BBSG's depends on the luminosity of the host galaxy. This dependance has been explained through numerical simulations as a statistical effect (Greggio 1986, and references therein). The simulation further showed that deviations from the mean relation between B₃ and $M_{\rm gal}$ up to $\pm$ 1 mag. can be expected in the B band for galaxies which are as faint as UGC 685. This explains the rather low accuracy of the method as derived from empirical data by R&RR. The errors given in the following are based on the derivation by R&RR.

I identified the four brightest candidates for supergiants (Nos. 146, 194, 217, and 250 in Table 3) where I used B-R < 0.4 as color selection criterion for blue supergiants, approximating the B-V color selection of K&T. The three brightest candidates yield a mean magnitude of B₃ = 20.9 $\pm$ 0.6. To get an estimate of the uncertainty introduced by clusters or blends, I also calculated B₃ from the second to fourth brightest blue objects and got 21.4 $\pm$ 0.3. The values corrected for the Milky Way foreground extinction are 20.6 and 21.1, respectively. Following R&RR, their relation and figure, and using the above values for B₃ and total magnitude after reddening correction, I obtain an absolute magnitude for the brightest blue supergiants of about -7.9 and a distance modules (distance) of 28.7 (5.5 Mpc). According to R&RR, the minimum error of a distance modulus derived from their relation (10f) is 0.88. To this value, I added quadratically the error of B₃ and the - almost negligible - error in total magnitude. This yields a total error of 1.0 in distance modulus or about 35% in distance. The derived distance value was used to convert observed to absolute values as given in Table 2 and to adjust the tracks (Fig. 6). Following the calibration of K&T yields almost identical values.

I was able to identify four individual sources in the J frame which could be cross-correlated with the objects from the B and R frames (Table 4). The colors were corrected for the Milky Way foreground reddening. Assuming that these four objects are supergiants and that UGC 685 itself does not contribute internal reddening, one can translate the position in the $B-R_{\rm c}$, B-J color-color diagram to a spectral type using the table of Johnson (1966). The Johnson table was transfer into Cousins R with Bessels (1987) relation. Using the blue absolute magnitudes of supergiants for these spectral types as listed by Schmidt-Kaler (1982), one gets distance moduli as also given in Table 4. Three objects combine to m-M = 28.8 (5.8 Mpc), including the fourth one yields 29.1 (6.7 Mpc). I take this as a consistency check for the B₃ method. The deviating result for the fourth object shows the limitation of this approach and may be caused by a blend or one of the above assumptions may not be valid in this case.

  

Table 4: Individual objects identified in B, R, and J within UGC 685. Listed are the numbers from Table 3, magnitudes and colors, the spectral types as estimated from the color-color values, and the individual dereddened distance moduli assuming that the objects are supergiants in UGC 685. See text for details

The two distance estimates presented here are in rather good agreement with an estimate based on the observed HI velocity and a Virgo infall model which yields about 6.0 Mpc (Schmidt & Boller 1992a). Thus, UGC 685 really belongs to the 10 Mpc sample and is quite isolated in space.

3.5 Star formation history

According to the results described so far, UGC 685 is in the right distance for an HST multicolor study which would provide the data for a detailed reconstruction of the star formation history (Schulte-Ladbeck et al. 1998, see also the review by Grebel 1997). Even though the ground-based data presented in this paper have a much more limited resolution, they already indicate that the recent star formation rate (SFR) is rather low and that it varied only slightly during the past.

According to the results in chapter 3.2 and 3.3, the observed total H${\alpha}$ flux corresponds to a recent (within the last 10⁷ yr) SFR of 0.003 $M_{\odot}$ yr^-1. As shown, this activity is spatially highly concentrated. The resolved supergiants and the underlying light distribution are related to slightly older star formation activities. The resolved supergiants are almost all inside 26 arcsec (major axis) which corresponds to 0.70 kpc or 2 scale lengths of the overall light distribution. This indicates that averaged over a longer time interval of $\sim$10⁸ yr, the SF took place all over the central part of UGC 685. This is supported by the absence of significant color gradients in B-R and B-J and the very limited deviation from the average in the color map.

A further estimator of the SFR of an actively star forming galaxy is its total blue luminosity. As described by Gallagher et al. (1984), this estimator $\alpha_{\rm L}$ is strictly valid only for a constant SFR and depends on the assumed IMF and its upper and lower mass cut-offs. It also slightly depends on the chemical composition of the system (see Gallagher et al. 1984, for complete references). The blue light is dominated by stars of $2.5~\mbox{\,$M_{\odot}$}\gt {M}_\ast \gt 1$ $M_{\odot}$ ($4\ 10^8\ {\rm yr} \gt {\rm age} \gt 6\ 10^9$ yr). Applying the Gallagher et al. formula (7) yields the same SFR as derived from the H${\alpha}$ flux. Thus, I have little evidence for strong variations of the SFR with time up to $\sim 10^9$ yr.

To strengthen this point, I used the total optical colors B-R and B-I (Table 2). I compared the observed, dereddened colors with Bruzual & Charlot (1994) models with exponential declining star formation rate of e-folding time $\tau$. Models with $\tau \le 5$ Gyr do not fit the colors at all and are far too red. A $\tau = 10$ Gyr model can reproduce the colors rather well with deviations of less than 0.1. Models with an exponential decay and a recent, even small, burst are in conflict with the data. Therefore, also the colors support the scenario of a simple star formation history.

In other words, the isolated dwarf irregular galaxy UGC 685 seems to have an approximately constant and low SFR, roughly stable for the past 10⁹ yr. The SFR is also quite low when compared to other galaxies of the same morphological type - actually, UGC 685 ranges at the lower end of measured SFR values of irregular galaxies. As I applied the formulae of Gallagher et al. (1984), Hunter & Gallagher (1985, 1986), it is straightforward to compare the obtained SFR's of UGC 685 to the values for their sample galaxies. For dwarf irregular galaxies, they found SFR values between 0.5 and 0.0002 $M_{\odot}$ yr^-1 (giant irregulars: 0.3 to 1.1). The SFR correlate with the absolute blue magnitude. Galaxies which have the similar $M_B \pm 0.5$as UGC 685 show similar SFR values.

As indicated by the amount of detected HI gas, this galaxy can continue with star formation at its today rate for quite a long time. To do so, the reservoir of far outlying HI gas has to be moved into those central regions where conditions obviously support star formation. As no external trigger seems to be present, this might be a very slow process, depending on the dynamical evolution of the extended HI disk. It might even be possible that the stellar body of todays UGC 685 will already be totally dimmed to an object resembling a dwarf spheroidal before this outlying gas is able to move in. In this context, it is interesting to note that Carignan et al. (1998) found some outlying HI gas in the Sculptor dwarf spheroidal galaxy.