Polarimetry with IRCAM/TUFTI

IMPORTANT: Note that IRCAM3 (also know as TUFTI) has been de-commissioned. These pages are for reference only. Thermal imaging-polarimetry is now available with UIST.

L-band polarimetry is now available with TUFTI. Following the recent upgrades to IRCAM (addition of a cold snout and re-scaling of the pixels to 0.08″), IRCAM is now optimised for use in the thermal IR (L and M-bands) and has acquired a new name, “TUFTI”. Details of polarimetry options with TUFTI are given in the pages listed below.

Users of IRPOL2 with IRCAM/TUFTI may also find the IRCAM3 Polarimetry manual written by Ant Chrysostomou of some use. It contains a number of very helpful tips and comments concerning imaging polarimetry in general. A postscript version is available here.

The UKIRT and University of Hertfordshire would appreciate an acknowledgement in any publication which contains data obtained using IRPOL2. 

Polarimetry with IRCAM/TUFTI: Data Acquisition

Getting started – IRPOL setup

IRPOL2 comprises a half-wave retarder (the waveplate), a focal-plane mask and (internal to IRCAM/TUFTI) a Wollaston prism (the focal plane mask and half-wave retarder will be installed by the Telescope System Specialist, or T.S.S., at the beginning of the night). Separate waveplates are available for JHK, L and M-band polarimetry. The prism splits the incoming radiation into orthogonally-polarised e- and o-beams. With extended sources (greater than about 5 arcsec in declination), these beams would overlap on the array. Hence the need for the mask. At present a warm, focal-plane mask must be positioned in front of the IRCAM/TUFTI “snout”. For thermal imaging, the SILVERED masked should be carefully positioned at the spot marked on the bench.  IMPORTANT: Take a few exposures of blank sky or even the inside of the dome at the beginning of the night to check that the mask is properly positioned; a raw image with the mask properly positioned should look something like that shown in Fig.1 below. Any hairs or fibres on the mask will appear “resolved” in both e- and o-beam images (Fig.1. shows “nice” examples of this); these can be blown off the mask with a can of pressurised air.

Fig.1. A typical raw L-band IRCAM+Pol image with the mask well aligned.

An IRPOL image is split into 4 east-west strips. The two northern strips represent “extra-ordinary” and “ordinary” images of the same field viewed through the north slot in the mask; the two southern-most strips are o- and e-beam images of an adjacent field viewed through the south slot in the mask. Note, however, that on all image displays the IRCAM field is rotated CLOCKWISE through 89.0o. Consequently, the east-west o- and e-beam images are rotated, as in the example above. The reduction recipes assume that the star is located in the NORTHERN (or right hand) half of the array. The DR recipes extract object images from this area only; the southern e- and o-beam images are used to calculate sky values. The usable area of each E-W beam on the array (i.e. the size of each horizontal strip) measures approximately 20 x 4 arcsec.

Observing programmes – the ORAC-OT and ORAC-OM

IRCAM/TUFTI imaging-polarimetry data are acquired in much the same way that UFTI-pol data are obtained. A description of how to use the ORAC Observing Tool (OT) to prepare observing sequences, and how to observer via the ORAC Observing Manager (OM) and Sequencer are available in the UFTI + IRPOL Data Acquisition pages.

When preparing L-band sequences for IRCAM+IRPOL, please bear in mind that very short integration times will be needed, typically less that 1 second. The IRCAM field of view is also much smaller than that of UFTI. Recommended Template Sequences, specific to IRCAM/TUFTI, are available in the Template Library of the ORAC-OT.

Point Sources and Extended Objects

It is recommended that observers repeat observations of both point and extended sources several times. This will give the observer an indication, via internal statistics, of the accuracy of the final result. With IRCAM/TUFTI, a source that fits comfortably within the limited field of view of the “e-beam” and “o-beam” images on the array (see Fig.1 above) may be regarded as a point source. If the user wishes to retain spatial information and avoid bad pixels, then a small 3 point jitter pattern mosaic is recommended, with say 5 arcsec east-west offsets between each mosaic tile. The “Pol_Jitter then Angle (IRCAM)”, and “Pol_Angle then Jitter (IRCAM)” sequences in the Template Library are provided for such an observation.

For extended sources which are more than ~ 5 arcsec in declination, a “Pol_Extended” sequence is provided in the Template Library that allows the user to nod between target and blank sky (at each of the 4 waveplate positions).


With near-infrared astronomy, a good flatfield is of course always a high priority, although for polarimetry the flatfield is in principle independent of the final result, since it is the differences between observations which are measured. However, one can only safely assume this if the same pixels are compared each time. Image motion, for whatever reason (some degree of dithering is recommended to account for bad pixels) will mean that the images must be shifted to spatially register them, rendering the assumption of flatfield independency invalid. Once again, a good flatfield is – as in most cases in astronomy at UKIRT – crucial!

There are a variety of methods which the observer can employ to obtain a flat field. At present we provide two Template Sequences that can be used to obtain flat field frames; the first combines images taken at different waveplate angles to make a “master-flat”, while the other – which is more time consuming, though is possibly more accurate – computes flats at each of the waveplate angles separately. In either case, with IRCAM/TUFTI at L-band, getting sufficient signal-to-noise on the array will not be a problem!

Note also that, because a focal plane mask is used with TUFTI, the corresponding flatfield frames must also be obtained with the focal plane mask in place. This is because the flatfield response of the array is sensitive to the state of polarisation of the radiation, and since the prism and mask always ensure that only e- or o-beam radiation is transmitted onto a given area of the array, then the flatfield response of that same area of the array must likewise only be measured in either e- or o-beam radiation. Once again, the purpose of the mask is to ensure that the two beams are kept separate on the array.

Changing Waveplates During the Night

For L-band polarimetry a zero-order waveplate is available. Note that, if near-IR polarimetric imaging is also required for your project, there are a limited number of near-IR filters installed in IRCAM/TUFTI at any one time. Consequently, JHK polarimetry can be obtained with IRCAM/TUFTI, provided the filters are not in the same filter wheel as the prism. This might be a more efficient option than switching between IRCAM/TUFTI and UFTI, although note the much larger field of view associated with the latter.

The L-band waveplate is warm (it sits in an extendable arm within the instrument support unit – ISU2). We also have an M-band waveplate, though tests suggest that it is unusable with IRCAM/TUFTI because of the high background. The waveplates are large (95mm diameter) to accommodate the full IRCAM/TUFTI field of view. This makes them VERY EXPENSIVE! Therefore, only certain UKIRT staff are allowed to handle the plates. If your observing programme calls for a change of waveplates during the night, then you should inform your support scientist of this. All the TSSs know the procedure and safety aspects of changing waveplates, and they should be the ones to do this during the night if required. Observers should NOT attempt to change waveplates!

Polarimetry with TUFTI: Data Reduction

ORAC-DR for Imaging Polarimetry

ORAC-DR pipeline recipes are now available for use at the telescope (and at your home institute) with IRCAM (and UFTI) imaging polarimetry. Since IRCAM reduction follows essentially the same rules as UFTI reduction, we direct the reader to the IRPOL+UFTI Data Reduction web page.  Note: the oracdr_ufti command should of course be replaced with oracdr_ircam. All ORAC-DR recipes for Polarimetry are common to both IRCAM and UFTI.

The STARLINK software package POLPACK may also be of use for (re)reduction.

e-o Beam Separations with IRCAM/TUFTI 

The Wollaston prism in TUFTI acts as the polarimetric analyser, splitting the incoming radiation into orthogonally polarised e- and o-beams. The divergence of the beams is dependent on the refractive index of the prism material and is therefore wavelength dependent. 

The beam separations at L were measured through various filters on a bright (6th magnitude) star. All IRCAM images are rotated clockwise so that North is at a position angle of -89.0 degrees (see this example image) and Eastincreases to the top. The values given below are the offsets of the far-right e-beam-star from the centre-right o-beam-star; we assume that the e-beam is nearest the northern/right-hand edge of the array. 

The IRCAM/TUFTI pixel scale is 0.0814″.

BANDOffset North/to the right (arcsec)Offset East/to the top (arcsec)
3.28um “dust”5.01-0.12
3.5um mbL4.91-0.15

Essentially, the beam separation in the L-band is 5 arcseconds (north-south).

Polarimetry with IRCAM/TUFTI: 
Instrumental Polarisation and Efficiency 

A bright, high-proper-motion (unpolarised) standard was observed through various filters in November 1999 during commissioning of IRPOL with IRCAM/TUFTI. These indicate that the instrumental + background polarisation may be a little higher than at shorter wavelengths (see the UFTI + Pol web pages). Notably, P increases with wavelength in the table below, and is heighest through the broad band Lp98 filter. This additional “thermal” contribution is probably constant with time. Users of IRPOL with IRCAM/TUFTI should clearly dedicate some time to observing an unpolarised standard during their run with their chosen instrument configuration.

3.28um “dust”0.62(+/-0.24)76(+/-30)
3.4um nbL0.85(+/-0.2)75(+/-10)
3.5um mbL0.82(+/-0.3)119(+/-13)

The polarisation efficiency of IRPOL at 3-4 microns has been measured with CGS4, and found to be wavelength dependent though in excess of 90%. Users should consult these pages for further details.  

Position Angle Calibration 

To calibrate measurements of polarisation position angle with IRCAM + IRPOL, deep observations should be obtained of a polarised standard. As a guideline, note that from recent observations of the polarised standard HD38563C with UFTI we measured the degree of polarisation and position angle at K to be 2.10(+/-0.09)% and 87(+/-6)o. These compare with catalogue values of 2.21(+/-0.55)% and 78(+/-17)o (Whittet et al. 1992, ApJ, 386, 562). The data suggest that a correction of -9o should be applied to p.a. measurements. Measurements with IRCAM3 — prior to it conversion to IRCAM/TUFTI in summer 1999 — indicated the need for a similar correction of -6o.