Astron. Astrophys. 336, 743-752 (1998)

2. Observations

2.1. General overview

On May 25, 1995, a time series of a quiet region at the centre of the solar disk was taken with the Vacuum Tower Telescope (VTT) at the Observatorio del Teide on Tenerife. For this observation, the two-dimensional spectrometer mounted in an optical laboratory of the VTT (see e.g. Bendlin et al. (1992)) was used to obtain so-called white-light images and narrow-band filtergrams in the Na D₂ line strictly simultaneously with two CCDs. Its line formation height makes Na D₂ a suitable tool for studying the lower to middle chromosphere.

2.2. Short introduction to the two-dimensional spectrometer

As the spectrometer in the VTT is a non-standard and also a versatile instrument meeting different obervational needs (Volkmer 1995), a mere description of the observations gained with it will not provide sufficient information on the peculiarities of the data and their proper treatment in the reduction. A schematic representation as well as a detailed description of the spectrometer is given in Bendlin et al. (1992), while some components which were introduced more recently are treated in Bendlin & Volkmer (1995).

For two-dimensional spectroscopic observations, like here, the instrumental set-up mainly consists of a series of filters placed in the light path leading to "CCD 2" which is used to take narrow-band filtergrams. First, an interference filter (IF) with a FWHM of about 10 Å was chosen to pre-select the wavelength range around the Na D₂ line. A beam splitter cube (BSC) behind the IF reflects some light onto "CCD 1" which thus receives so-called white light, but about 90% of the transmitted light passes the BSC in the direction of CCD 2. The IF also serves to suppress neighbouring maxima in the transmission curve of the next filter. The latter is a universal birefringent filter (UBF) with a FWHM of 0.46 Å at the wavelength of Na D₂ (5890 Å). A considerably narrower pass band is produced by a Fabry-Perot interfermeter (FPI), the last filter in line. Specific information on the FPI used here and on the relevant formulae describing its performance can be found (e.g.) in Bendlin (1993) or Bendlin & Volkmer (1995), therefore only such characteristics are mentioned here which are essential in this context. The FPI's spectral resolution (and wavelength position of the selected transmission maximum) is variable, depending on the separation of its mirror plates. For the rather broad Na D₂ line, a mirror gap d of 1.25 mm yielding a spectral resolution of about was considered an acceptable compromise, as the free spectral range FSR thus becomes wide enough (1.38 Å) for scanning through a sufficiently large part of the line. The performance of the FPI is determined by a computer-controlled ramp generator feeding the piezoelectric stacks of one mirror with the required voltage to alter the mirror separation in a well-defined way. The images within a scan were taken in quick succession. After each individual exposure, the pass band of the FPI was shifted by 30 mÅ . This procedure was repeated from the beginning of the selected wavelength range for every scan of the time series.

The UBF is placed at a telecentric position, thus the transmission conditions for the different light bundles belonging to different points in the field of view are uniform. The same cannot be said of the FPI which leaves all parallel rays of light, belonging to any one point, with a blueshift due to its position at the pupil's image of the telescope. The blueshift is largest for the largest angles of incidence and vanishes for rays parallel to the optical axis. This effect is taken into account (and corrected for in the data reduction), as the FPI's position at the pupil's image guarantees the highest spectral resolution possible.

Two personal computers (PCs) execute the specific observational programme defined by suitably chosen observing parameters which are entered via a special software control programme. The Peltier-cooled CCDs connected with these PCs take images with a resolution of 12 bits. The maximum image acquisition rate is better than 3 full frames (corresponding to 384 286 pixels) per second. While a scan is taken, the CCD frames are stored in the PC-RAM, and in between two successive scans the data are transferred onto a hard disk.

When needed, several additional light paths branching off or leading into the main light path may be utilized via movable mirrors. Of foremost importance is the light path containing a continuum source to take flat-field images spanning the same wavelength range as the solar images in order to obtain "pure" transmission curves under the same conditions as the actual time series. It was therefore designed to imitate the solar light path as closely as possible to ensure that the same angles of incidence occur on the various optical surfaces of the spectrometer's components and that the same blueshifts of the transmission curves are produced by the FPI.

Some other light paths are equipped with helpful components to facilitate any necessary adjustments of the spectrometer.

2.3. Outline of the observations

From the wide range of observing parameters offered by the instrumental set-up and the software programme controlling the spectrometer, the following combination was chosen for the observed set of data:

The number of white-light and narrow-band scans in the time series is 128, each made up of 30 individual CCD frames. Every 56 s, a new scan was started, so the whole time series covers nearly two hours. With the selected mirror separation of the FPI, the spectral resolution reached about , corresponding to a smallest resolvable wavelength difference of  30 mÅ . The wavelength step between successive narrow-band images was 30 mÅ as well. For solar images (and, of course, the corresponding dark and flat-field images), an exposure time of 60 ms proved sufficient. In order to obtain solar flat-field scans, the telescope was defocused. The flat-field images taken with a continuum source, necessary for the narrow-band scans only, reached a fair signal-to-noise ratio after an integration time of 600 ms. They required another set of dark images with the same exposure time. With the CCDs' maximum field of view of 384 286 pixels and with the imaging achromat used here, a spatial resolution of 0:002 per pixel on the CCD-chips was achieved, yielding an (original) image size of 76:008 57:002.

The whole set of data necessary for the data reduction was supplemented with so-called grid images. They were taken after the final adjustment of the spectrometer with a grid placed at the prime focus of the VTT. These exposures are needed for the correct alignment of white-light images with the narrow-band filtergrams (see Sect. 3.1).

© European Southern Observatory (ESO) 1998

Online publication: July 20, 1998
helpdesk.link@springer.de