Astron. Astrophys. 346, 285-294 (1999)

3. The SUMER spectrometer

SUMER is described in detail by Wilhelm et al. (1995a). First results together with operational performance characteristics are discussed in Wilhelm et al. (1997) and Lemaire et al. (1997). The instrument has no inflight wavelength calibration system and, moreover, no stable optical bench other than is achievable with a light-weight aluminium structure adequate for space applications. It has a spectral pixel size of  mÅ in first order at 1540 Å and  mÅ in second order of diffraction at 770 Å. The Doppler formula then yields a velocity sensitivity at these wavelengths of 8.2 km s^-1 per pixel. With simulation studies (Wilhelm et al. 1995b), it could be shown that line shifts of strong lines can be determined relative to the unshifted line with a precision of 0.1 of a pixel even with short integration times of several seconds. Similar results have been obtained under operational conditions (e.g., Warren et al. 1997; Brekke et al. 1997; Judge et al. 1998). The mechanical deformations of the optical bench caused by residual periodic temperature variations of  50 mK of the thermal control system led to spectral shifts of  0.15 pixel within 2 hours, and also to a trend of  0.5 pixels within 8 hours. This can be seen from Fig. 1 for the observations discussed here. The drift is caused by a thermal adjustment to the conditions of the specific observational sequence. The thermal control situation is, however, dependent on the specific mode of observation, and has to be studied on a case by case basis (see, e.g., Curdt et al. 1997).

Fig. 1. Spectral line shifts caused by thermal deformations of the optical bench of SUMER. The periodic contribution is correlated with the cycle of the thermal control heater. The drift reflects the thermal adjustment of the instrument to the specific mode of operation, which is slightly different for each sequence of observations. The shift was determined by averaging the line positions of Ne VIII (770) along the slit and over  10 raster positions during the scan of September 21, 1996 (solid line) and of September 22, 1996 (dotted line). "Red" and "Blue" indicate the direction of the effect, if interpreted as Doppler shift.

Because of our accuracy requirements, we have to improve on the standard SUMER data analysis methods. We intend to do this by using information derived from the data directly as described in Sect. 5.1. It is important to mention that Ne VIII (770) can be observed in second order with the SUMER detector A (with a wavelength range in first order from 780 Å to 1610 Å). In this case the first order lines near 1540 Å are seen in the background of the intense neon line. This is essential for our arguments, as the grating formula, is strictly linear in m, where m is the order of diffraction, d is the grating spacing, is the angle of incidence on the grating, and is the angle of reflection off the grating. Up to  40 Å can be exposed simultaneously on the detector in first order; 20 Å thereof on the potassium bromide (KBr) photocathode. The selection of a wavelength band for observation is done by putting a certain wavelength on a specific reference pixel of the detector. Without re-adjusting the grating position, one therefore can look at the strong lines Ne VIII (770), Si II (1533) and C IV (1548, 1550) together with numerous weaker lines, mainly of the Si I and C I spectra. By rastering the slit position across the Sun, maps of certain solar areas have been built up.

© European Southern Observatory (ESO) 1999

Online publication: May 6, 1999
helpdesk.link@springer.de