Date: Fri, 26 Aug 2005 09:19:46 -0700 (PDT)
From: Spitzer Science Center HelpDesk 
Subject: SPITZER: Spitzer IRS Long-High Sky Subtraction


Dear Spitzer Observers,

As a result of the continual increase in the number of degraded pixels in
the IRS Long High module due to the effects of Cosmic Rays, the IRS
Instrument Support Team strongly urges that all Long High (LH)
observations of sources with equivalent point source flux densities of 2.0
Jy or less at 25 microns be accompanied by off-source background
measurements. Examination of the properties of Long High warm, or "rogue"
pixels (pixels with abnormally high dark current that vary unpredictably
with time), shows that the rogue dark current can be reduced substantially
by subtracting a nearby background spectrum taken within a few hours of
the science spectrum.  Improvements in the signal-to-noise in the
extracted spectrum can be substantial, - up to 80% when averaged over an
entire order of Long High and when the integration time is split evenly
between target and background, since for faint sources the dark current in
the rogue pixels is the dominant source of noise.

This recommendation applies chiefly to those observers using IRS Staring
mode.  Observers who are mapping their sources may only need one
additional background observation, or none, depending on the extent of
their map.  An example of a portion of the LH array, before and after
background subtraction, is shown below.

To reduce the (increasing) number of rogue pixels, we will be reducing the
bias voltage on the Long High array in an upcoming IRS campaign (likely to
occur in October 2005).  This should reduce the current number of rogue
pixels (approximately 1200) by a factor of about 2.5.  This reduction in
bias should improve the overall signal to noise ratio of LH spectra even
if no additional background observations are obtained however, background
subtraction will further improve the signal-to-noise in faint source
spectra by about 30%, averaged over an order of Long High.

In addition, we are developing software for the detection and removal, via
interpolation, of rogue pixels in IRS data.  This software will be made
available to all observers via the SSC contributed software web site.  The
target date for release of this software is October 1st, 2005.  For more
information on rogue pixels in the IRS arrays, please refer to the
website:

        http://ssc.spitzer.caltech.edu/irs/roguepixels/

If an observer has to split out the LH/background observation into another
AOR because the current AOR has other slits selected, the cost of
splitting the observation into another AOR (a.k.a. the 215 sec overhead
charged to each AOR) will not be charged to the observer. However, the
duration of the background measurement must come out of the observer's
allocation.  The LH background AOR should have no peakup, should replicate
the cycles and ramp durations of the LH science AOR, and should be chained
to the LH science AOR.

The SSC urges observers to review their AORs immediately and contact us
via the Helpdesk (help@spitzer.caltech.edu) to request specific AORs be
put on hold to add background observations. AORs that are already
tentatively scheduled cannot be pulled from the scheduling queue for
revision. Background observations should be added in a timely fashion
(i.e. within a week or so of the request).

If you have any questions, please contact the Help Desk.


SSC IRS Instrument Support and Observer Support Teams