A quick look tool for EFOSC2 long-slit spectroscopy
The tool is based on the context/long of Midas. A basic knowledge of the context would thus be helpful in case of trouble. In any case, it's good to know that the current setup can be inspected with show/long, and parameters values (both variables and the software behaviour) can be changed with set/long parameter=xxx
The software should be as much non-interactive as possible, so it assumes that you have a guess session for the wavelength calibration. The guess sessions can be downloaded from:
efosc2_guess.tgz (5.7 MB)
Currently, the package contains sessions for grisms 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, and 16.
The tool is described here below, and in more detail in a short Manual in PDF format.
- 0 Prerequisites
- 1 Start MIDAS
- 2 Getting some help
- 3 Prepare the images
- 4 Apply the wavelength calibration<
- 5 Extract the spectrum
- 6 Create the response function
- 7 Apply the flux calibration
Download efosc2_qlook.tgz (7KB), and put all the *.prg in ~/midwork. Even better, put the files in a subdirectory (e.g. call it EFOSC2) and create symlinks in ~/midwork, so things are cleaner this way.
Put the tables of the guess sessions in ~/midwork and convert them into MIDAS tables, with a command like indisk/fits initial_xxx.tbl.fits initial_xxx.tbl.
- Arc line lists are stored in: /scisoft/midas/calib/data/spec/line/
- The atmospheric extinction curve is stored in: /scisoft/midas/calib/data/spec/exti/
- The standard stars flux tables are stored in /scisoft/midas/calib/data/spec/flux/ and /scisoft/midas/calib/data/spec/ctio/
The links above will give you the file list if you are browsing your local copy of this cookbook, but not on the web page!
- You have a guess session for your grism, so tables like the following exist (for example these are for grism 11):
- The names of the frames have the following form
BIASes are called : *Dark*
FFs are called : *FlatSpec*
Arcs are called : *HeAr*
Science spectra are called : *EFOSC_Spectrum*obj*
@@ inmidas !! of course
The following command will give you the list of available prg's
Midas xxx> @@ readme
Midas xxx> @@ prepare
The program does:
- Initialize default parameters for EFOSC2
- Convert FITS into BDF
- Rotate to have wavelength running "from left to right"
- Set the O_AIRM descriptor (it is not set automatically)
- Create the median BIAS frame and subtract from all images
- Create the master FF frame (the interpolation along the dispersion is done with a polynomial of degree 'fdeg', and 'fdeg' is set at the beginning
- divide each frame by the master FF
- create master HeAr frame
- Calibrate in wavelength, based on a guess session
The program calls two more macros: the initialization is done by:
And the wavelength calibration is done by:
In turn, waveguess displays the solution using
Now during the night, each time you get a new spectrum, you can go ahead with the following steps.
@@ calib xxx.fits:
Of course, this will first subtract the BIAS, divide by FF, trim the image, etc. (all the stuff that was done for the calib images).
The wavelength calibration also corrects for the distortions in the cross-dispersion direction, so you can now extract the spectrum.
The name of the new 2-d spectrum will be xxx.bdf
@@ exsingle xxx.bdf
The spectrum is extracted taking the extraction windows for the object and the sky from the file obdef.dat (a template is created by @@ defaultpars). The file looks like this:
--- obdef.dat ---
change these values if you wish
From top to bottom, the numbers are the lower sky window, the object window, and the upper sky window. The possibility of extracting more than one object from the 2D spectrum is forseen, so an object nr. can be specified with the last parameter, and it will be used to name the extracted spectrum, which will be called like 2D_spectrum_name_objnnn.bdf, where nnn is the object number defined by 'object_nr' in obdef.dat.
The spectrum is extracted by
which uses the plain average method. If you wish to use the optimal extraction, please see EXTRACT/LONG.
Now you could be satisfied with the extracted spectrum, but in case you observed a standard, you can flux-calibrate the extracted spectrum.
If you have a standard star spectrum, then first apply the wavelength calibration and then extract its spectrum, as you did for all the other spectra (so do @@ calib and @@ exsingle in sequence). After this execute
Which will find the response curve. By default, the response function is fitted with a polynomial of degree 12.
@@ fluxlong xxx_obj1.bdf