5 Morphological classification and profile fitting

We performed the morphological classification of our sample galaxies in the framework of three classification schemes for ELGs, namely those proposed by Salzer et al. ([1989], SMB89 hereafter), Loose & Thuan ([1985], LTh85 hereafter), and Telles et al. ([1997], TMT97 hereafter). The scheme of SMB89 is the most quantitative one. It is based on measurable photometric parameters, as absolute magnitude, linear size and morphology, without making detailed use of the spectral properties.

Following the prescriptions of SMB89 we classified our ELGs as follows:

- Seyfert galaxies (Sy) and Starburst Nucleus Galaxies (SBN) have, on average, similar magnitudes ( $M_{ B} \leq$ -20) and sizes ($\sim$20 kpc). They can be disentangled from each other by including spectral line characteristics. The classification of a few Seyfert galaxies in our sample was obtained from the literature. Other ELGs with corresponding luminosities and sizes are classified as SBN (or as Giant Irregulars - GI, in a few cases);

- Dwarf Amorphous Nuclear Starburst Galaxies (DANS) are smaller ($\sim$10 kpc) and less luminous ( $M_{ B} \sim$ -18.5) symmetric disk galaxies with bright (stellar) nuclear region, of amorphous appearance, and show no features in their disk;

- H II Hotspot galaxies (HIIH) have sizes and luminosities as DANS but are irregular in shape, with strong emission regions everywhere in the galaxy, more often found off-center;

- Dwarf H II Hotspot galaxies (DHIIH) are as HIIH but smaller in size ($\sim$3.8 kpc) and in absolute magnitude ( $M_{ B} \sim$ -16.5);

- Sargent-Searle (SS) objects are the smallest ($\leq$ 2 kpc) and the least luminous (<M _B> = -14) ELGs;

- Interacting Pairs (IP) consist of two galaxies of comparable size, which show evidence of recent interaction.

The classes of DANS, HIIH, DHIIH and SS make up the broad group of Blue Compact Galaxies (BCGs), and most of them are also dwarfs, namely Blue Compact Dwarf Galaxies (BCDs). We assigned the general type of BCD to some less clear cases, where it was not possible to distinguish between different morphological subtypes. A few candidates in our sample were classified as Magellanic Irregulars (MI) and Giant Irregulars (GI).

The assigned morphological classes are given in Tables  6, 7 (Col. 12).

LTh85 studied a sample of $\sim$ 50 BCDs on deep CCD frames and proposed a classification scheme based on the regularity of the isophotal shapes of both the high-surface-brightness (HSB) star-forming regions and the low-surface-brightness (LSB) outer region. Each BCD is described by two letters, one lower case letter which describes the location and shape of the star-forming regions (n - for a single star-forming region located at the nucleus of the galaxy with circular or elliptical isophotes; i - for several star-forming regions which are not at the center and possess irregular isophotes), and one upper case letter which describes the shape of the outer isophotes (E - for circular or elliptical outer isophotes; I - for irregular outer isophotes).

LTh85 distinguished four main types of BCDs: - iE, the most common type in their sample; - nE; - iI, with subtypes of cometary(iI,C), and merger(iI,M) evidences; - i0, for pure stellar images without underlying older stellar population. Finally, $\sim$10 galaxies (out of $\sim$50) in LTh85 sample, which show some spiral structure, were reclassified as starburst spiral galaxies.

We classify our sample ELGs by means of above notation in Tables  6, 7 (Col. 13). For clearly non-dwarf ELGs with some spiral features we use notation S...

In their study of structural and morphological properties of H II galaxies TMT97 proposed a further morphological classification scheme, based on two main criteria: on the multiplicity of H II regions and on the shape of the outer isophotes. Both morphological classification schemes of LTh85 and TMT97 relay on similar structural features, but they are complementary to each other: the scheme of TMT97 specifies the multiplicity of the H II regions; LTh85 classes describe the outer isophotes in more detail. For comparison of different morphological classification schemes we assigne the TMT97-classes to our sample of ELGs too, by means of describing the morphology of each ELG with two integers. The first integer shows the number of resolved H II regions: 1 - single dominant giant H II region; 2 - double H II region; 3 - multiple ($\geq$ 3) H II region. The second integer characterizes outer structure: 1 - disturbed morphologies and irregular outer isophotes, extensions, fans or tails beyond the H II regions; 2 - symmetric and regular objects, regardless of the multiplicity of the star-forming regions (i.e. their internal structure). The assigned TMT97-classes are listed in Tables  6, 7 (Col. 14).

The perception of the morphology depends on redshift (and on observing conditions). TMT97 note that morphological details may be smeared out at higher redshift (z > 0.02), rendering the galaxies with a smoother compact appearance. In any case, the visual classification is increasingly difficult for faint and small-sized ELGs. Their poorly resolved images are dominated by bright H II regions, which can hide the presence of an underlying stellar component. Morphological information should be extracted using an objective classification method.

Next, we attempt to quantify the galaxian morphology by means of analysing their surface brightness profiles. Upon inspection of our SB profiles it becomes evident that many of them have linear parts and a central light excess. The occurence of an outer linear part in the SB profile was taken as an evidence for the presence of an underlying stellar disk component of the parent galaxy. We fitted the outer linear part by means of a simple exponential intensity law

$$I(r)~=~ I_0~{\rm exp}(-\alpha r). \protect$$ (4)

The total light emitted by the exponential disk can be computed as

$$L_{\rm T} = 2 \pi \alpha^2 I_{\rm0}. \protect$$ (5)

The excess light which remains after the subtraction of the exponential disk is often emitted by a single or multiple bright hotspot, which is typically offset from the geometrical center of the outer isophotes of ELGs. We refer to this excess light as to a nucleus which should be distinguished from the classical bulge. The nucleus typically shows an unresolved stellar profile. The light profiles of a few luminous SBNs in the present sample show resolved central spherical components (which can usually be fitted by an r^1/4-profile), superimposed on the excess light of the nuclear starburst. We did not attempt to disentangle between the bulge light and the extra light emitted by the starburst. Some SB profiles show central light depression instead of typical light excess and are more flattened than could be expected from poor seeing conditions. Flat SB profiles, which show convex curvature over linear radius, are characteristic of non-nucleated dwarf galaxies. Among the dwarf ELGs the convex SB profiles show weak star-forming activity, distributed in a number of faint hotspots. A number of ELGs show nearly unresolved stellar images and their SB profiles match closely the stellar SB profiles, without any reliable underlying extended envelope. Quasi-stellar BCDs at low redshift, which appear unresolved even on deep CCD frames, are of particular interest, since they are good candidates for being truely young galaxies (see Izotov & Thuan [1999], and references for a discussion). In addition, a few galaxies show definitely non-exponential SB profiles, i.e. they always show concave curvature over the linear radius.

The results of the profile fitting are summarized in the middle part of Tables 6, 7, with the data organised as follows:

Column 1: Galaxy name. Isolated galaxies are designated with an asterisk.

Column 2: Distance D (in Mpc) calculated from the measured redshifts of Popescu et al. ([1996,1998]), which were corrected for the Galaxy's motion with respect to the velocity centroid of the Local Group $V_0 = v_{\odot}$+300sinlcosb, (Sandage [1975]) and transformed into distances assuming H₀ = 75 km s^-1 Mpc^-1, q₀=0.

Column 3: Absolute magnitude corrected for foreground absorption as determined by Burstein & Heiles ([1978]).

Columns 4, 5: Effective ( $r_{\rm eff}$) and isophotal (r₂₅) radii in kiloparsecs.

Column 6: Seeing corrected approximate linear size of the star-forming regions ( $r_{\rm {\textup{H\, {\mdseries\textsc{ii}}}}}$) in kiloparsecs, measured in the B band as described in TMT97. $r_{\rm {\textup{H\, {\mdseries\textsc{ii}}}}}^2$ = 0.25( $FWHM_{\rm {\textup{H\, {\mdseries\textsc{ii}}}}}^2-FWHM_{\rm st}^2$), where $FWHM_{\rm {\textup{H\, {\mdseries\textsc{ii}}}}}$ is FWHM of the circular area centered on the burst region of the galaxy, $FWHM_{\rm st}$ is seeing FWHM of the stellar images.

Columns 7, 8: Exponential model scale length $\alpha^{-1}$ in arcseconds and in kiloparsecs, respectively.

Columns 9, 10: Exponential model central surface brightness $\mu^{\rm exp}_0$ and its corrected value $\mu^{\rm exp}_{0,\rm c}$, where flattening and foreground absorption corrections are applied as follows: $\mu^{\rm exp}_{0,\rm c} = \mu^{\rm exp}_0~+~2.5\,{\rm log}(a/b)-\rm Abs$.

Column 11: Assigned profile type with following notation. Type 1: pure exponential disk - d. Type 2: composite profile with nucleus or bulge - n/b and disk - d and probably an outer halo - h:. Type 3: central light depression - c or always convex curvature over linear radius) - cnv. Type 4: quasi-stellar (non-resolved) profile - st. Type 5: definitely non-exponential profile with concave curvature over linear radius - cnc.

Columns 12-14: Morphological classes according to SMB89, LTh85, and TMT97, respectively, as described above.

Typical errors of the exponential disk model parameters are given in the last row of Table 6. These errors are medians of the residuals of all multiple measurements of individual ELGs.

 

 

Table 2: Comparison of mean photometric characteristics of isolated and non-isolated ELGs

Sample	Size1)	z	MB	$B_{\rm T}-R_{\rm T}$	$<\mu_{\rm eff,c}^{B}>^{2)}$	$r_{\rm eff}$	r25	b/a	$c_{\rm in}$	$\mu_{\rm0,c}^{\rm exp}$	$\alpha^{-1}$
			[mag]	[mag]	[ $\frac{\rm mag}{\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hf... ...rlap{$\sqcap$ }$\sqcup$ } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi''}$]	[kpc]	[kpc]			[ $\frac{\rm mag}{\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hf... ...rlap{$\sqcap$ }$\sqcup$ } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi''}$]	[kpc]
Non-isolated ELGs	124(108)	0.041	-17.9	0.88	21.6	2.1	5.0	0.71	0.46	21.7	1.8
(total sample)		$\pm$ .048	1.8	.29	1.1	2.3	4,5	.18	.11	1.0	1.9
Non-isolated ELGs	69(59)	0.018	-16.8	0.88	21.7	1.0	2.6	0.68	0.48	21.7	0.8
(local, $z \leq$0.033)		$\pm$ .008	1.5	.26	1.1	.6	1.5	.19	.11	1.0	.5
Isolated ELGs	15(15)	0.021	-16.5	0.93	22.3	1.1	2.5	0.68	0.47	22.8	1.0
		$\pm$ .009	1.5	.27	.7	.7	1.4	.21	.09	0.8	.5

¹⁾  Sample size: the number of ELGs which were observed in B and/or in R (observed in B).
²⁾  The mean (effective) SB within the half-light radius, corrected for the galactic absorption and for the inclination.

Figure 5: Distribution of non-isolated (circles) and isolated (filled circles) ELGs at various redshifts (z)

Figure 6: Relations between the surface brightness characteristics: a) mean (effective) surface brightness within the half-light radius, b) log effective radius, c) exponential disk central surface brightness, d) log exponential disk scale length versus absolute B-magnitude. Big circles ($\circ$) mark non-isolated resolved ELGs in local volume, filled circles ($\bullet$) mark isolated resolved ELGs in the same volume. Small circles mark distant ELGs (outside the local volume). Unresolved ELGs with quasi-stellar images are marked by $\times$ - for non-isolated ELGs and by + - for isolated ELGs. d) shows the linear fit of log disk-scale-length -  absolute magnitude relation for isolated (continuous line) and non-isolated (dashed line) ELGs within the local volume