High Dispersion Systematics

High Dispersion Systematics

This page presents currently known systematic effects in NEWSIPS high dispersion combined order spectra (MXHI files).

The high orders in IUE echelle spectra overlap. This means that at the shortest wavelengths covered by each camera, the spectrum in each order is contaminated by a contribution from its neighboring orders, making extraction of the spectra extremely difficult. The project elected to correct for order overlap with an approach which incorporates a global background subtraction and correction algorithm. Because the processing of the spectra had to be done in "batch mode", this routine had to be both general and robust enough to handle all spectra contained in the archive. Unfortunately, this algorithm often creates systematic effects in well exposed spectra of certain kinds of objects. Some currently well documented problems are:

  1. Time dependent effects in large aperture, high dispersion, combined order spectra.

  2. Non-zero Ly alpha absorption: In many instances, the extracted fluxes do not reach zero in the region of the heavily saturated Ly alpha 1215Å line. This is incorrect, and indicates an error in the reduction software, most likely in the background subtraction. This problem can usually be seen in spectra which have intrinsically high flux levels in the region of Ly alpha. The following images provide examples of this problem.

  3. Negative fluxes: In certain instances, the extracted spectra have negative fluxes. These are often in the cores of strong wind lines, particularly the C IV 1550Å doublet, as can be seen in the following examples, or in the weak or zero continuum regions of emission line sources such as or over a wide wavelength range of a relatively red object such as

  4. For a variety of reasons, the ripple correction software sometimes fails to perform well. In these cases, a characteristic wavy pattern can be seen in the combined order spectra. This usually occurs at the longest wavelengths, and is often worse in small aperture spectra. The effect can be seen most easily in intrinsically featureless spectra, such as those of hot main sequence stars subdwarfs. The following images provide examples of these problems:

  5. Reprint of the paper "Science Verification of the IUE Final Archive Data Products" (1998, ESA SP-413, Ultraviolet Astrophysics Beyond the IUE Final Archive, eds. Wamsteker and Riestra, ESA Pubs. Div., Noordwijk, The Netherlands, p723) as a PDF file. This paper describes many of the preceding problems and some low resolution problems which were detailed in Massa and Fitzpatrick 1999 ApJS, 126, 517).

    Many of the effects outlines above may result from poor background corrections (see, Myron Smith 1999 PASP, 111, 722). Users should be aware that combined order spectra MAY contain background correction errors and time dependent errors as large as ± several tens of percent. To obtain the maximum possible signal-to-noise from IUE high dispersion spectra (~3-5%), one should begin with the more fundamental data products (the SIHI or LIHI files) and perform customized extractions.


    Send comments to: Derck Massa