1 Introduction

Blue compact galaxies (BCGs) have luminosities in the range $M_B \simeq -12$ mag to $M_B \simeq -21$ mag (Kunth & Östlin 2000). Much work has been devoted to the study of BCGs, including image (such as Kunth et al. 1988; Doublier et al. 1999; Cairós et al. 2001), and spectra (such as Kunth & Sargent 1979; Terlevich et al. 1991; Kinney et al. 1993; Kunth et al. 1997) observation in all wavelengths. These galaxies have an average surface brightness higher than 20 mag arcsec^-2 (Gordon & Gottesman 1981), and are characterized by their compact morphology and blue rest-frame colors ( B-V <0.45, B-r < 0.6; Pisano et al. 2001). Most of BCGs have a saturated region surrounded by no, or only small amounts of, nebulosity, and show no regular structures, such as spiral arms, but may display irregular features, such as jets, filaments, bridges, etc.(Gordon & Gottesman 1981). BCGs were first observed spectroscopically by Sargent & Searle (1970). Their optical spectra show strong narrow emission lines superposed on an almost featureless continuum, similar to the spectrum of an H  II region. The least blue systems generally show both emission and absorption lines, while the bluest usually show only emission lines. The blue rest-frame colors and strong narrow emission line spectra indicate intense current star formation activity (Cairós et al. 2001). Population synthesis models yield typical star formation rates between 1 and 20 $\mbox{${\; {M_{\odot}}}$ }$ yr^-1 (Mas-Hesse & Kunth 1999; Östlin et al. 2001).

Neutral hydrogen observations, in the 21-cm line, of BCGs have shown that BCGs are typically rich in H  I gas, and there are indications that they have large H  I halos (Thuan et al. 1999; Pisano et al. 2001). Another interesting property found among BCGs is that some of them have significant under-abundance of elements heavier than helium. The abundances of heavy elements in these galaxies range between $\mbox{${\; {Z_{\odot}}}$ }/50$ and $\mbox{${\; {Z_{\odot}}}$ }/2$, and put them among the least chemically evolved galaxies in the universe. The two most metal-deficient galaxies known, I Zw 18 ( $Z_{\odot}/50$) and SBS 0335-052 ( $Z_{\odot}/41$) are BCGs (Doublier et al. 1999; Izotov & Thuan 1999). These properties make BCGs represent an extreme environment for star formation, that differs from that in the Milky Way and in many other nearby galaxies. Detailed studies of these galaxies are not only important for understanding their intrinsic properties, but also crucial for understanding star formation processes, galaxy evolution and cosmological parameters (Kunth & Östlin 2000; Izotov et al. 2001).

The Hubble Space Telescope and the new generation of 8 m class telescopes have extended our knowledge of the early Universe by identifying galaxies down to magnitudes $B\sim 28$ mag and redshifts $z\ge 3$. Studies of intermediate-redshift galaxies have revealed a population of compact, luminous galaxies with high star formation rates. Jangren et al. (2002) have compared several of the brighter objects of some intermediate-redshift samples and found that most of the galaxies have small sizes, high luminosities (hence high surface-brightness), and very blue colors, similar to the BCGs (Pisano et al. 2001). Therefore, the more accessible nearby BCGs should help us better understand these distant objects.

Despite the extensive work carried out during the past decade, age and star formation histories of BCGs are not very constrained. With the advent of modern detectors, many BCGs (>90%) have been found to contain an underlying old stellar population (Kunth & Östlin 2000, 2001). It suggests that most BCGs undergo a few or several short bursts of star formation followed by longer, more quiescent periods (Kong & Cheng 1999). An important question is whether these BCGs are unique objects or not? Only a large spectral sample can help us answer this question. The data published so far are not sufficient to allow a systematic analysis of the star formation history of BCGs, because detailed and exhaustive studies have concentrated on particular type of BCGs, such as very metal-deficient galaxies.

Motivated by these facts, we have undertaken an extensive study of a large sample of BCGs, including the collction of IRAS, H  I and optical spectroscopy data. Our ultimate goal is to perform a detailed analysis of these galaxies, focusing on the aspects of the various stellar components, the star formation rates, the metallicities (chemical abundances), the age of the underlying population, dust contents and star formation history.

In this paper, we present the blue compact galaxies sample, the optical spectroscopic observations and the method of reduction. The paper is organized as follows: in Sect. 2, we present the galaxy sample. In Sect. 3, we describe the spectral observations. The data reduction, and calibration are outlined in Sect. 4. The spectrophotometric atlas is presented in Sect. 5. In Sect. 6, we assess the data quality and spectrophotometric errors. Finally, some concluding remarks are given in Sect. 7. In the forthcoming papers of this series, we will study the constraints set by these observations on the physical properties of BCGs in our sample.