From wilhelm@linmpi.mpg.de Thu Nov 20 15:02:58 2003
Date: Thu, 26 Sep 2002 12:11:20 +0200 (MET_DST)
From: wilhelm@linmpi.mpg.de
Reply-To: sumer-soft@esa.nascom.nasa.gov
To: sumer-soft@esa.nascom.nasa.gov
Subject: SUMER-SOFT: SUMER Line Width Correction

Subject: SUMER Instrumental Width

Dear colleagues,

since 1997 the SUMER line width correction programme ``con_width_funct_2.pro''
has been available as standard correction method. It assumes, 
based on optical calculations, a wavelength-independent instrumental width 
for detector A of 77 mA (7.7 pm) with a 0.3" slit, and 
for detector B of 128 mA (12.8 pm) at a wavelength of 680 A (68 nm),
110 mA (11.0 nm) at 1080 A (108 nm), and 95 mA (9.5 nm) at 1488 A (148.8 nm). 
All widths are given as FWHM and the wavelengths are in first order. 

As far as I know, only three estimates based on measurements
during the operational phase have been published: (1) 86 mA, detector A, 
1" slit (Wilhelm et al., 1997, Sol. Phys. 170, 75);  (2) 99 mA, 
detector A, 1" slit and 129 mA, detector B, 1" slit (Chae et al.,
1998; ApJ 505, 957). If de-convoluted to a 0.3" slit, Chae's et al. 
values correspond to 95 mA and 127 mA, respectively. It should be noted that
the value for detector B was given with a large uncertainty and, moreover, did
not cover measurements of the wavelength dependence expected for this detector.

The situation obviously called for a detailed study of the performance under
operational conditions, but the manpower to execute such a study was not
available. The first suspicion was that the instrumental widths could change
with time (for instance, depending on the high voltage at the MCP), 
however, I found in a quick survey that this is fortunately not the case, 
at least not as a significant effect.

In the meantime, several investigators using ``con_width_funct_2'' found that
at small measured widths, the SUMER correction led to line widths that were
too broad if compared with other data. In this situation, after discussions
with Werner Curdt, I have implemented ``con_width_funct_3'' based on 
Chae's et al. data. It is available in SolarSoft and its application is 
similar to ``con_width_funct_2'', which will continue to be available. 
If you want to run the programme outside the EOF environment, 
you first have to build up the corresponding look-up tables by executing 
the procedure ``non_width_3''. You also have to edit the address of the 
restore procedure in ``con_width_funct_3''.

There is a new feature implemented in ```con_width_funct_3''. If you prefer
to work in ``pure'' SI units, you can enter the wavelength in nanometre and
the width in picometre, the output will then automatically switch to picometer, 
but the angstroem and milli-angstroem option is still available.

For detector A the new corection function appears to be the best choice we
have, and I recommend to use it. To give an idea what the differences
compared to ``con_width_funct_2'' are, it should be mentioned that for a
1" slit at measured widths of 200 mA and 120 mA, the relative line
width decreases are 5 % and 25 %, respectively.

For detector B the situation is far less satisfactory. At short wavelengths
the new correction is not significantly different from the old one, whereas
at the long-wavelength limit it probably leads to an over-correction.
The best advice seems to be to use the new function also for B, but to be
aware of its short-comings. If a comparison between emission line widths
measured with detector A with those obtained from B can be made, more
confidence should be put in A.

Finally, if somebody is willing to embark on a systematic study of the
instrumental width of SUMER we would gladly accept such help.

If you have any problems with the new function, please do not hesitate
to contact us. 
        
With best regards,
Klaus Wilhelm