2. Observations and data analysis
NGC 5252 was observed on March 6 and 9, 1997 with the spectrograph CARELEC (Lemaître et al. 1989) attached to the Cassegrain focus of the Observatoire de Haute-Provence 1.93m telescope. The detector was a 512 512 pixel, 27 27 µm Tektronic CCD. We used a 600 l mm-1 grating giving a dispersion of 66 Å mm-1. A Schott GG 435 filter was used in the red spectral range, 6305-7215 Å; the wavelength range covered in the blue was 4825-5730 Å.
The slit width was 2:000, corresponding to a projected slit width on the detector of 50 µm or 1.9 pixel; the slit P.A. was for the blue spectrum and for the red one. In each case, the galaxy nucleus was centered on the slit and 3 columns of the CCD ( 3:002) were extracted. The seeing was on both nights; the resolution, as measured on the night sky emission lines, was 3.5 Å FWHM in the blue, and 3.0 Å FWHM in the red regions. The spectra were flux calibrated using the standard star Feige 66 (Massey et al. 1988), also used to correct the observations for the atmospheric absorption.
To supplement our own observations, we searched the HST archives for spectra of the central region of NGC 5252. Three FOS (description by Ford & Hartig 1990) spectra were retrieved, corresponding to the nucleus and the two knots located 0:0036 NE and 0:0031 SW from it (details are given in Table 1). All three HST spectra were obtained under the same setting conditions, with the G570H grating (4.37 Å diode-1) and the FOS/RED detector (a 512 diodes linear array), resulting in a spectral range of 4570-6820 Å; a single 0:0026-diameter circular aperture was used. The spectra were submitted to the usual processes of substepping and overscanning, resulting in a 2 064 pixel coverage. Since each diode corresponds to 0:0031 in the dispersion direction, the resolution was estimated at about 3.7 Å FWHM. The HST data were processed by the calibration pipeline Calfos, which includes flat-fielding, subtraction of the background and sky, and wavelength and flux calibrations. All (HST and OHP) spectra were deredshifted to rest wavelengths with z = 0.023.
Table 1. OHP and HST observing log
The observing log is given in Table 1.
2.2. Data analysis
The presence of an old star population with many strong absorption lines can make the line fitting analysis difficult, especially in the blue spectral region. Bica (1988) has shown this old star population to have similar spectra in all Morgan (1958, 1959) classes; therefore, a suitable fraction of the spectra of the elliptical galaxies NGC 5982 and NGC 4365 (used as templates for the blue and red spectra, respectively) was subtracted from the observations to remove the old stellar population contribution. NGC 5982 was observed on March 6, 1997 with the same instrumental setting used for NGC 5252, while NGC 4365 was observed on February 28, 1984 with the Boller & Chivens spectrograph and the Image Dissector Scanner attached to the Cassegrain focus of the ESO 3.6 m telescope at La Silla; the dispersion was 59 Å mm-1 and the resolution, 4.5 Å FWHM. The subtraction of these template spectra from our observations resulted in much smoother continua and corrected the flux of the Balmer emission lines for the underlying Balmer absorption.
Inspection of our two-dimensional spectra shows the lines to be double and spatially extended in the nuclear region, confirming earlier results. HST and OHP spectra were analysed in terms of Gaussian components, as described in Véron et al. (1997). The emission lines H , [N ii] 6548, 6584, [S ii] 6716, 6731 and [O i] 6300, 6363 (or H and [O iii] 4959, 5007) were fitted by one or several sets of seven (three) Gaussian components; the width and redshift of each component in a set were taken to be the same. Therefore, in addition to the line intensities, the free parameters for each set of lines are one width and one redshift. The intensity ratios of the [N ii] 6548, 6584, [O iii] 4959, 5007 and [O i] 6300, 6363 lines were taken to be equal to 3.00, 2.96 and 3.11, respectively (Osterbrock 1974).
Fitting our large aperture red spectrum with two sets of Gaussians gives unsatisfactory results, with large residuals not only for the H +[N ii] complex, but also for the [O i] lines, which have an obvious blue wing. The best solution is not obtained by adding a broad H component, but rather a third set of Gaussians. This third set of components has a relatively broad width ( 1 460 km s-1), 6584 H = 0.90 and extremely strong [O i] lines ( 6300 H = 0.73). We obtained only an upper limit to the strength of the broad component of the [S ii] lines; while the total [S ii] flux relative to H for the two narrow components is 1.2 and 0.9 respectively, this ratio is 0.5 for the broad component. The same model, with three sets of components, also succeeds in matching the blue spectrum, one set having a large width ( 1 080 km s-1) and 5007 H = 3.59. The width found for the H component is much larger than that of the equivalent H line; this, however, may not be significant since the errors in the H width should be larger, the narrow H and [N ii] components having a considerable relative strength. This "broad" line component has the characteristics of a Liner, while the two "narrow" line components are Seyfert 2-like, with weak [O i] lines and strong [O iii] emission (see Table 2). Although the width of the "broad" component may seem large for a Seyfert 2 or a Liner, it is not exceptional as the line width of the prototype Seyfert 2 galaxy NGC 1068 is 1 500 km s-1 (Marconi et al. 1996).
Table 2. Line profile fitting results for the OHP and HST spectra. Fluxes are in units of 10-16 erg cm-2 s-1. FWHMs have been corrected for instrumental broadening
We found no evidence for the presence of broad Balmer components typical of Seyfert 1 nuclei. The broad H component observed by Osterbrock & Martel (1993) and Acosta-Pulido et al. (1996) does not really seem to exist; this feature is rather due to the unresolved blend of the H and [N ii] components. Whittle (1985) has mentioned the possibility of misidentifying weak relatively broad wings to the H and [N ii] lines with a broad ( 2 000 km s-1) H component; we have an illustration of such a possibility in IRAS 13197-1627: Aguero et al. (1994) have fitted H and the H +[N ii] complex with a set of narrow Gaussian components, adding a broad component to the Balmer lines; however, Young et al. (1996) showed that the H +[N ii] complex can be very satisfactorily fitted by two sets of Gaussians (with FWHMs of 400 and 1 350 km s-1 and a 6584 H ratio of 1.94 and 1.35, respectively) without adding a broad H component.
Fitting the HST spectra showed that both the NE and SW knots (Fig. 1) have Seyfert 2-like spectra with relatively narrow lines ( 325 and 220 km s-1 FWHM respectively, corrected for the instrumental broadening) and a velocity difference of 345 km s-1 (the SW knot being redshifted with respect to the NE knot). The observed line ratios are 5007 H = 11.11 (7.09), 6300 H = 0.26 (0.29) and 6584 H = 0.88 (1.03) for the NE (SW) knot, respectively. The nucleus spectrum (Fig. 1) is quite different: the lines are broad and have a complex profile; they have been fitted with two sets of components and have line ratios typical of Liners (see Table 2).
Our larger aperture ( 3:002) included both the nucleus and the two bright knots (Fig. 2, upper panels). Subtracting the HST nucleus spectrum from our own, without any scaling, resulted in a spectrum which is well fitted by a set of two narrow components (Fig. 2, lower panels) showing that the broad lines come exclusively from the small 0:0026 aperture centered on the nucleus. The velocity difference between the two narrow line systems in the resulting spectrum is 250 km s-1, the FWHM is approximately the same in both systems ( 225 km s-1) and the line ratios ( 5007 H = 8.39 (6.73), 6300 H = 0.19 (0.23) and 6584 H = 0.82 (1.01) for the blueshifted and redshifted components, respectively) are very similar to those measured on the bright knots. The sum of the H fluxes of the two knots is equal to 20% of the total H flux measured on the differential (OHP - HST ) spectrum, confirming the finding of Tsvetanov et al. (1996), that the knots are embedded in faint diffuse gas.
© European Southern Observatory (ESO) 1998
Online publication: April 28, 1998