1 Introduction

Many investigations of solar structures require multi-dimensional spectroscopy, so it is a subject of intense instrumental developments. Usually, multi-dimensionality implies the spectroscopy in two spatial coordinates. Additional information such as the state of polarization and the temporal evolution are often measured too.

Data obtained by such an instrument can be described as a cube with two spatial dimensions and one dimension of wavelength (Fig. 1). Investigations of the state of polarization add for each measurement taken such a cube. A Stokes-V polarimeter for example splits the original cube into two sub-cubes. Time series are giving such a set of data cubes for each time a measurement is made. Hence, the amount of data obtained can thus become very large.

Classical spectroscopy yields single slices of this cube, for example: slit spectra, spectro-heliograms, narrow band filtergrams, etc... Several methods are known to obtain complete data cubes. Taking two-dimensional images while scanning through the wavelength is done with a Fabry-Perot interferometer (FPI, Bonaccini et al. 1989; Bendlin et al. 1992). An instrument with similar capabilities as a FPI is the Solar Optical Universal Polarimeter (SOUP, Title 1984; Title et al. 1986), based on a fast tunable birefringent filter.

The Multi-channel Subtractive Double-Pass Spectrograph (MSDP, Mein 1977, 1991) allows to obtain simultaneous spatial and wavelength information. Scanning in one spatial direction and obtaining slit spectra, with the slit perpendicular to the scanning direction, is another method (e.g. Johannesson et al. 1987; Ballesteros et al. 1996).

All these methods have advantages and disadvantages. Some instruments are difficult to operate, others mix information from different coordinates which have to be disentangled during data reduction. If high spectral resolution is needed, spatial scanning will be the choice. In this paper we present such a device: the Micro-Image-Scanner (MISC).

  

Figure 1: The hyper cube structure of multidimensional spectra. The front cube was derived from observational data