2 Description of the observations

2.1 Optical imaging

We have imaged HCG 54 in the Johnson R band at the 2.5 m NOT telescope at el Roque de los Muchachos (La Palma) using the ALFOSC spectrograph during the nights of March 21st and 27th and May 16th 2001. The main characteristics of the observations are given in Table 1. The detector was a Loral/Lesser (CCD#7) $2048\times 2048$, with a spatial scale of 0.189 $^{\prime \prime }$/pixel, which gives a field of about $6\hbox{$.\mkern-4mu^\prime$ }5 \times 6\hbox{$.\mkern-4mu^\prime$ }5$. In order to cover a larger field (16 $.\!\!^{\prime\prime}$ 6 $\times$ 16 $.\!\!^{\prime\prime}$ 7) tracing the direction of the HI tail (see Sect. 3.3), a total of 10 shifted exposures were taken during the 1st run for 600 s each. The seeing was varying between 1 $.\!\!^{\prime\prime}$1 and 1 $.\!\!^{\prime\prime}$4 during the observations. Ten bias exposures were taken through this run and were used to construct a median bias which was subtracted from each image. Pixel-to-pixel variations were evaluated with a median normalized twilight flat-field. Then a bidimensional fourth order Legendre polynomial was fitted using IMSURFIT task in IRAF. Images were divided by this flat field. After correction of instrumental effects, the atmospheric extinction was determined and corrected from observations of selected fields from Landolt list of standard stars (PG0918 and PG1047). The rms errors of the standard stars in the final calibration are smaller than 0.07 mag. Finally all individual images were combined in a mosaic by using the IRAF SQIID package created by K. M. Merrit at NOAO, and using SQMOS, XYGET and NIRCOMBINE routines, together with a mask built for each frame that prevented the mean of the frame edges, as well as bad CCD lines. The resulting image is shown in Fig. 1, where three previously unclassified galaxies A1126+2051, A1127+2054 and A1127+2057 are also identified (see Sects. 3.1 and 3.2).

Figure 1: Full field of the R-Johnson band optical image of HCG 54 obtained as explained in the text (Sect. 2.1). The dotted lines delineate the HI integrated emission at a level of 0.5 $\times$ 1020 atoms cm-2. We have labeled the three dwarf galaxies discussed in the text (Sect. 3.2).

Figure 2: Combined frame of B, R and r' bands in order to better recover the faintest emission. Contours are shown for the central part that appears saturated in our greyscale. The objects identified by Hickson et al. (1989) as the members of HCG 54 are indicated with letters a to d. We also mark the tidal optical features described in Sect. 3.1 t1 to t4, and an example of the unresolved knots (k) that we detect close to HCG 54.

Figure 3: B-R map of HCG 54. Averaged colors of different regions are indicated by arrows, and the displayed range goes from 0 (white) to 0.8 (black). We have superimposed two R band isophotes as a reference.

Figure 4: R band image of HCG 54 showing the direction of the slits used for the spectroscopical observations in the main body of HCG 54 (see Sect. 2.2), and the sections indicated in Tables 3 and 4.

Figure 5: a) Up: spectrum in the direction joining HCG 54a and b (sp3), where the position is given in offsets with respect to the peak of emission of HCG 54b. The indicated regions are described in Sect. 3.2. Bottom: R band profile along the slit.

Figure 5: b) Up: spectrum in the direction joining HCG 54c and d (sp4), where the position is given in offsets with respect to the peak of emission of HCG 54c. Bottom: R band profile along the slit.

Figure 5: c) Spectrum in the direction joining HCG 54d and A1126+2051 (sp1, Fig. 4), for the sections where emission was detected. Positions are given in offsets with respect to the continuum peak of HCG 54d along the slit.

Figure 6: Full spectrum of HCG 54a.

Figure 7: a) Full spectrum of HCG 54b.

Figure 7: b) Selected range of the same spectrum shown in a).

In the second run, a deeper red band exposure was obtained for the central part in order to better define the faintest optical structures. The red image is a composition of six 1200 s images with the Sloan-Gunn r' filter. A blue image was also obtained in order to be able to derive colour indices of the different components of the group. The Johnson B filter image taken in the 3rd run is a composition of four 1200 s frames. Slight dithering was applied between successive frames for both filters, in order to avoid cosmic rays and bad pixels. All frames (R, r' and B band images) were finally combined in a search for the best definition of the weakest structures (Fig. 2). None of the nights of these two last runs were photometric, so no calibration was obtained for the r' and B band images. Instead we adopted the one proposed by Hickson et al. (1989) for the (B - R) color map (Fig. 3), using the integrated B-R color of HCG 54b as a reference.

 

Table 1: Photometric observations.

Image	Date	$T_{\rm ex}$	Filter
Mosaic	21/03/2001	6000	R-Johnson
Center	27/03/2001	7200	r' Gunn
Center	16/05/2001	4800	B-Johnson

Figure 8: a) Left: HI column density contours 0.5, 0.7, 1.2, 2.0, 2.7, 3.4, 4.4, 6.7, 8.4, 10.1, 13.5, 16.9, 20.2 to 22.6 $\times 10^{20}$ atoms cm-2 overlapped on the R image of HCG 54. The atomic emission has been integrated in the velocity range 1334-1490 km s-1. The synthesized beam is 20 $^{\prime \prime }$$\times$ 16 $^{\prime \prime }$. b) Right: map of the first-order moment of the HI radial velocity field. The scale goes as in the wedge where the numbers indicate heliocentric velocities in km s-1. The contours go from 1340 to 1490 km s-1 with a step of 10  km s-1. The beam size is 20 $^{\prime \prime }$$\times$ 16 $^{\prime \prime }$.

Figure 9: Channel maps of the 21 cm line radiation superimposed on the R image of the group. Heliocentric velocities are indicated in each panel. Contours correspond to 2.4, 4.0, 5.6, 7.3, 8.9, 10.5, 12.1, 13.7, 15.3 K, and the rms noise of the maps is 1.2 K. The synthesized beam (15 $.\!\!^{\prime\prime}$8 $\times$ 14 $.\!\!^{\prime\prime}$5) is plotted in the upper left panel.

Figure 9: continued.

   
2.2 Optical spectroscopy

The spectra were also obtained with ALFOSC at the NOT 2.5m telescope, with the same detector used for the images and using Grism#4, 7 and 8, during April 2000 and March 2001. Table 2 summarizes the long-slit spectra taken for this study. The format is as follows: Col. 1 spectrum identification, 2 direction through which the spectrum was taken, 3 date of observation, 4 Grism used, 5 spectral dispersion, 6 number of exposures, 7 duration of each exposure in seconds, 8 spectral range, 9 position angle of the slit in degrees measured from N to E, and 10 width of the slit. The slit widths were chosen to match approximately the seeing of the nights. The direction of the slits listed in Table 2 as well as the different zones from which the spectra have been extracted along the slits are marked in Fig. 4. Observations with Grism#4 were performed in order to derive general spectral characteristics and kinematics of the galaxies, and to determine redshifts for possible satellite galaxies in the field. Grisms#7 and #8 were used to characterize the kinematics with a higher spectral resolution than for Grism#4. Observations with Grism#7 covered with good spectral resolution the range including [OII]3727 and [SII]6717/6731. In the observations performed with Grism#8 the [NII], H$\alpha$ and [SII] lines were detected and used to characterize the kinematics.

The spectra were reduced according to the usual methods, including subtraction of a mean bias calculated for each night, division by a median flat-field obtained for each configuration, as well as wavelength calibration based on HeNe calibration lamp spectra obtained after and before each exposure. The rms of the bidimensional calibration was 0.4 Å for Grism#4 and 0.1 Å for Grism#7 and Grism#8. A mean sky was obtained from object-free sections of each spectrum. In all cases several exposures were taken in order to increase the signal to noise ratio and to remove cosmic rays. The derived physical parameters described in Sect. 3.2 are given in Tables 3 and 4.

In order to obtain redshifts and velocity dispersions we used the cross-correlation technique derived by Tonry & Davis (1979). The templates used for the spectra showing emission lines were: the brightest spatial section of HCG 54a and HCG 54c for the Grism#8 and Grism#7 spectra respectively, together with a synthetic spectrum built from the rest frame wavelengths of the emission lines. In the case of the absorption line spectra, where our goal was to determine the redshift of A1126+2051 and A1127+205, we used as template the spectrum of the radial velocity standard giant star HD186176 obtained with the same configuration. In order to improve the signal to noise ratio of the derived curves a binning of 2 pixels (0 $.\!\!^{\prime\prime}$38) has been applied to the spectra sp5 and sp6 in the spatial direction and of 3 pixels (0 $.\!\!^{\prime\prime}$57) to sp1 (Fig. 4). The radial velocity data in the direction joining HCG 54a and b are shown in Fig. 5a, in the HCG 54c - d direction in Fig. 5b together with the section of sp1 corresponding to HCG 54d, and in the HCG 54d - A1126+2051 in Fig. 5c. The spectra for HCG 54a and b taken with Grism#4 are shown in Fig. 6 and 7 respectively. Spectrophotometric standard stars were observed in order to get a proper calibration of the line fluxes in sp3 and sp4 (Fig. 4). However, since the nights of April 27 and 28 were non-photometric, only the relative fluxes of the emission lines are reliable. The rest of spectra were taken under photometric conditions.

 

Table 2: Spectroscopic observations.

Name	Direction	Date	Grism	R(Å/px)	N	$T_{\rm exp}$(s)	Wav.-range (Å)	PA (deg)	Slit width ( $^{\prime \prime }$)
sp1	HCG 54d-A1126+2051	21/03/01	GR4	3.0	3	1800	3018-9006	86	1.2
sp2	A1127+2057	21/03/01	GR4	3.0	2	1200	3018-9006	115	1.2
sp3	HCG 54a-b	27/03/01	GR4	3.0	3	1200	3018-9006	65	1.0
sp4	HCG 54c-d	28/03/01	GR4	3.0	3	1200	3018-9006	20	1.0
sp5	HCG 54a-b	27/04/00	GR8	1.2	3	1800	5816-8326	65	1.2
sp6	HCG 54c-d	21/03/01	GR7	1.5	2	1200	3815-6815	15	1.2

2.3 VLA HI observations

We have mapped HCG 54 with the VLA in the C array in August 1997. The synthesized beam is 20 $^{\prime \prime }$ $\times$ 16 $^{\prime \prime }$ (1.9 $\times$ 1.5 kpc at a distance of 19.6 Mpc) after natural weighting of the data. A velocity range between 1085  km s^-1 and 1730  km s^-1 was covered with a velocity resolution of 10.4  km s^-1 . The rms noise is 0.33 mJy/beam corresponding to a column density of 1.2 $\times$ 10¹⁹ at cm^-2. Assuming an HI linewidth of 30 km s^-1, the achieved HI mass detection limit is about 10⁶ $M_{\odot}$.

We have detected emission above 3$\sigma$ in the velocity range 1334.2 to 1490.2 km s^-1 . The integrated emission is presented in Fig. 8a superposed on the R-band image, and the velocity field is presented in Fig. 8b. The velocity channel maps are shown in Fig. 9. The total spectrum has been obtained by integrating the emission in the individual channel maps (Fig. 10, solid line). The total HI line flux detected, 6.23 Jy  km s^-1, is in good agreement with the single dish measurement by Huchtmeier (1997; Fig. 10, dotted line). The total HI mass derived is 5.6 $\times$ 10⁸ $M_{\odot}$.

 

Table 3: Line fluxes relative to H$\beta$and abundances¹ across slit position through HCG 54a-b.

ION	$\lambda$	$\char93 1$ (HCG 54b)	$\char93 2$	$\char93 3$ (HCG 54a)	$\char93 4$	$\char93 5$	$\char93 6$
[OII]	3727	$1.65\pm0.05$	$4.72\pm0.60$	$4.54\pm0.48$	$4.53\pm0.43$	$3.86\pm0.63$	$5.04\pm0.52$
[NeIII]	3869	$0.35\pm0.01$	-	-	-	-	-
H8+HeI	3889	$0.17\pm0.01$	-	-	-	-	-
H$\epsilon$+[NeIII]	3970	$0.26\pm0.01$	-	-	-	-	-
H$\delta$	4100	$0.24\pm0.02$	-	-	-	-	-
H$\gamma$	4340	$0.45\pm0.01$	$0.47\pm0.04$	-	-	-	0.43$\pm$0.04
[OIII]	4363	$0.04\pm0.01$	$0.14\pm0.01$	-	-	-	-
HeI	4472	$0.04\pm0.01$	-	-	-	-	-
H$\beta$	4861	$1.00\pm0.01$	$1.00\pm0.05$	$1.00\pm0.02$	$1.00\pm0.02$	$1.00\pm0.10$	$1.00\pm0.03$
[OIII]	4959	$1.60\pm0.02$	$0.68\pm0.06$	$0.79\pm0.04$	$1.56\pm0.07$	$0.53\pm0.09$	$0.63\pm0.04$
[OIII]	5007	$4.83\pm0.05$	$2.07\pm0.15$	$1.91\pm0.08$	$3.79\pm0.12$	$1.45\pm0.24$	$2.02\pm0.09$
HeI	5876	$0.15\pm0.01$	-	$0.25\pm0.03$	-	-	-
[OI]	6300	$0.05\pm0.01$	-	-	-	-	-
H$\alpha$	6563	$2.68\pm0.03$	$3.01\pm0.22$	$2.97\pm0.14$	$2.97\pm0.17$	$2.62\pm0.36$	2.81 0.15
[NII]	6584	-	-	$0.36\pm0.02$	-	-	-
HeI	6678	$0.03\pm0.01$	$0.26\pm0.03$	-	-	-	-
[SII]	6717,31	$0.23\pm0.01$	$0.65\pm0.06$	$1.03\pm0.08$	$1.16\pm0.06$	$0.52\pm0.17$	$0.93\pm0.09$
HeI	7065	$0.02\pm0.01$	-	-	-	-	-
[AIII]	7136	$0.08\pm0.01$	-	-	-	-	-
1.3*I(6584)/I(3727)		-	0.07	0.10	-	-	-
log $R_{\rm 23}$		0.90	0.87	0.85	0.98	0.76	0.89
P = [(1.3*I(5007)/I(H$\beta$))/ $R_{\rm 23}$]		0.79	0.36	0.35	0.52	0.33	0.34
log[1.3*I(5007)/I(3727)]		0.58	-0.24	-0.26	0.04	-0.31	-0.28
t[OIII] (104 K)		1.05	-	-	-	-	-
t[OII] (104 K)		1.14	-	-	-	-	-
104 O2+/H+		1.46	-	-	-	-	-
104 O+/H+		0.36	-	-	-	-	-
12 + Log O/H		8.26	8.30	8.30	8.20	8.10:, 8.60:	8.25:
LogNe2+/O2+		-0.67	-	-	-	-	-

¹ Uncertain values are indicated with ":''.

 

Table 4: Line fluxes and abundances across slit position through HCG 54c-d.

	$\lambda$	$\char93 1$	$\char93 2$	$\char93 3$
[OII]	3727	$3.26\pm0.37$	$4.53\pm0.58$	$3.35\pm0.50$
H$\gamma$	4340	$0.46\pm0.04$	-	-
H$\beta$	4861	$1.00\pm0.02$	$1.00\pm0.08$	$1.00\pm0.05$
[OIII]	4959	$0.76\pm0.05$	$1.64\pm0.16$	$0.80\pm0.08$
[OIII]	5007	$2.57\pm0.12$	$4.04\pm0.38$	$1.98\pm0.12$
H$\alpha$	6563	$3.12\pm0.18$	$2.97\pm0.36$	$2.69\pm0.21$
[SII]	6717,31	$0.78\pm0.17$	$0.91\pm0.17$	$0.79\pm0.10$
log $R_{\rm 23}$		0.82	0.99	0.77
P = [(1.3*I(5007)/I(H$\beta$))/ $R_{\rm 23}$]		0.51	0.54	0.43
log[1.3*I(5007)/I(3727)]		0.01	0.07	-0.12
12 + Log O/H		8.10:, 8.40:	8.20	8.10:, 8.45:

$^{\rm 1}$ Uncertain values are indicated with ":''.

Figure 10: Integrated HI emission as a function of the heliocentric velocity for the range where we detect emission. The solid line corresponds to our VLA data, while the dashed ones are single dish data from Huchtmeier (1997).

Figure 11: a) Upper-left: R band emission of A1127+2054 in greyscale with overlapped isophotes 22.6, 22.9, 23.3, 24.1 mag arcsec-2. b) Upper right: light profile of this galaxy in the R band as a function of the equivalent radius. c) Lower left: integrated HI emission of this galaxy where the contours correspond to 3.3, 6.6, 10.0 and 11.6 $\times$ 1020 atoms cm-2, and overlapped on the R image. d) Lower-right: map of the first-order moment of the HI radial velocity field. The scale goes as in the wedge where the numbers indicate heliocentric velocities in km s-1.