Durham University
Astronomical Instrumentation Group
SMIRFS-IFU
Overview
The SMIRFS-IFU is a newly commissioned system for two-dimensional
( integral
field ) spectroscopy in the near infrared (1-2µm) (see paper on SMIRFS ;
PASP in press, 1999).
It has two
co-equal functions: firstly as a technology demonstrator for the more
ambitious integral field units which the AIG is producing for the 4m
Herschel Telescope and 8m GMOS). Secondly to
do new science! Examples include:
- Studies of radio galaxies which are have co-aligned radio and
optical structures. What causes this? jet-induced star-formation,
scattering of light from an obscured AGNs escaping along the ject axis,
or what? The IFU allows us to map the velocity field and study the
distribution of ionising radiation. Because of their high redshift (z >
0.6), the interesting lines ([OII], [OIII], H-alpha) appear in the near
infrared.
- Studies of optically-obscured AGNs (Seyfert galaxies). For example, the
strong [FeII] line emission has various possible origins:
X-ray illumination of clouds in the narrow line region,
shocks from supernova remnants in a circumnuclear starburst or shocks from
interactions between the radio jet and the interstellar medium.
- Studies of shocked regions in star-forming regions.
- Studies of gravitationally-lensed objects (arcs). Since these
are generally distant objects, the key lines for identification will
appear in the near infrared. With spatially resolved spectra, it should be
possible to recreate the original velocity field by undistorting the datacube.
Even for those that are too faint to recover spatial information, the
IFU is an efficient way to obtain information from the integrated light,
acting as an image slicer.
Description
The IFU works with
CGS4 on the UK
Infrared Telescope (UKIRT) to provide a basic integral field
capability for the near infrared (1 - 2µm in the J and H
bands).
The field of view is shown in the picture below.
Because of the use of hexagaonal microlenses, the filling factor of the
system is essentially 100%. The microlens array is located in the field
plate unit of the SMIRFS infrared multifibre system (Haynes
et al. , Spectrum, 7, 4).
This two dimensional field is reformatted into a single long slit,
pictured below, which is projected onto the slit of CGS4 (inside its
cryostat) by means of the SMIRFS slit projection unit which replaces the
CGS4 calibration unit.
The system is a retrofit and a prototype. This means that it has not
been possible to design the system for optimal sensitivity and that the
number of elements is relatively small - limited by the slit length of
CGS4 which is only 72 pixels with the short camera.
Sensitivity
With the CGS4 short camera (150mm) and the 75/mm grating, the limiting
sensitivity, estimated from the measured throughput of the IFU and
the posted CGS4 sensitivity, is:
- J-band 3-sigma 1 hour = 0.9 x 10^-19 W/m^2/arcsec^2
- H-band 3-sigma 1 hour = 2.3 x 10^-19 W/m^2/arcsec^2
With the CGS4 long camera and the 150/mm grating, it will be possible to
improve on these limits by large factors if the line falls in between
major OH sky lines. Based on information on the CGS4 web page where
the estimated sensitivity for the long camera is posted and assuming that
the 150/mm grating is used to yield resolving powers of 5900 in J and
5000 in H, so that OH lines can be rejected, we expect the following
sensitivities:
- J-band 3-sigma 1 hour = 0.2 x 10^-19 W/m^2/arcsec^2
- H-band 3-sigma 1 hour = 0.3 x 10^-19 W/m^2/arcsec^2
These figures will be revised when new information becomes available.
If beam switching is used (recommended) these sensitivities limits will
be increased by 1.4.
Despite the non-optimal design, this system gives significant advantages
over a series of stepped longslit exposures. The advantage is roughly
similar to the diference between CGS4 on UKIRT and a slit spectrograph
on GEMINI (e.g. GNIRS).
Status
The SMIRFS-IFU successfully went through technical commissioning in June
1997. Only the short CGS4 camera was available so that each IFU element
projected onto 1 pixel. (With the long camera, each IFU element will be
critically sampled. Note that the system sampling is set at the IFU
input where critical sampling is achieved for images with 1.2 arcsec
FWHM).
Despite unseasonal snow and almost continual fog, we managed to set it
up and take some science exposures. Here is an example of raw
and reformatted
data for a Seyfert-2 galaxy (X marks the one and only broken fibre).
The throughput of the IFU was close to that predicted (about
50%). The element to element variation in throughput was good for a
fibre system and we have demonstrated that this flatfields out. Because
you cannot nod along the slit with the IFU, we recommend beam-switching
for background subtraction.
The following information on the commissioning is available.
(These documents are gzipped postscript. They have been tested
on a HP laserjet 4M+ printing from both a PC - using a proper ADOBE
driver - and from a Sun server. They look better when printed in colour.)
Do you want to use the SMIRFS-IFU?
It is available for use in collaboration with the SMIRFS-IFU
team. Please contact Jeremy Allington-Smith.
Reports
Commissioning report (gzip/postscript 229kB)
UKIRT Newsletter article
(gzip/postscript 144kB) and
html version copied from the newsletter.
Stop press
The IFU was tested again in Feb 1998 with the short camera of CGS4 and
the 150/mm grating and used to carry out a PATT-supported study of
spectral line imaging of Seyfert galaxies (PI: Simon Morris). This run
was very successful with good weather and good performance from the
instrument. It is planned that the IFU will be available on a
collaborative basis in semester 98B.
More information will follow as it becomes available.
The sensitivity figures quoted above for this configuration will be revised.
Just for fun
Here's a simulation of the data you might expect from a totally fictitious
rapidly rotating emission line galaxy. For reasons of simplicity the
spectrum is actually an optical one since I couldn't face putting in an
IR sky spectrum. Like I said, its only for fun. Can you work out why the
wiggles look like they do? (Hint: the slit traces out a snake-like
pattern on the sky.) Next time I will make the input datacube
bigger. Note that this is applicable to the CSG4 short camera. With the
long camera, the slit projects onto twice as many pixels giving 2x
oversampling of each IFU element.
j.r.allington-smith@durham.ac.uk
24 Aug 1999