The following set of scientific and technical guidelines were adopted as
basis for design of the instrument. (i) It should be
capable of observing faint extended objects like reflection nebulae,
accretion disks, dusty active galaxies etc., with
accuracy limited by photon noise and resolution limited by
seeing. This demands a typical field of view of a few arcminutes, an optics
which is well-matched with the telescope to minimize loss of light
and a detector
of sufficient sensitivity. Also, the effects of atmospheric scintillation
should be eliminated and instrumental polarization minimized so that such
measurements of low level polarization could be made. (ii) Since the objects
of interest are very often quite faint, the instrument should have an
acquisition and guidance (A&G) system which allows pointing the telescope
with an accuracy of a few arcseconds and tracking to better than 1
over long periods.
(iii) Observations should be possible in various wavelength
bands in the optical and near-IR regions. (iv) The entire instrument
consisting of the optics, A&G unit, associated
electronics etc. should be a self-contained unit which can be easily mounted
on the telescope, with only electrically isolated communication links to
the computers for instrument control and data acquisition. (v) The cost
of the instrument should be minimized by using standard optical and
electronic components and by avoiding over-specification as far as possible.
Figure 1 (click here) is a schematic representation of the optical arrangement
of IMPOL. The basic idea behind this arrangement is to use a Wollaston prism
with its axis normal to the optical axis of the system, as the analyzer to
convert the linear polarization in the incoming light into relative
intensity of two orthogonally polarized beams (the ordinary and the
extraordinary), separated by a small angle of 0.5
. This measurement is
sufficient to define one of the Stoke's parameters Q or U.
A half-wave plate with its fast-axis normal to the optical axis of the system,
is kept before the Wollaston prism on a rotatable mounting.
When the half-wave plate is rotated through an angle 
,
the plane of polarization rotates through an angle 
. At this new
position of the half-wave plate another measurement on the orthogonally
polarized beams can be made to determine the second Stoke's parameter as well.
It is easily seen that, for this arrangement the intensities
of the extraordinary and ordinary beams 
are given by
where 
are the
unpolarized and polarized intensities respectively;
are the position angles of the polarization
vector and the half-wave plate fast-axis respectively, with reference to the
axis of the Wollaston prism. Since the angle 
is conventionally
measured with respect to the celestial north-south axis (
towards north
celestial pole and increasing counter-clockwise), the axis of the Wollaston
prism is kept aligned to it.
Figure 1: Schematic of the IMPOL optical layout. The field lens and
half-wave plate are from Karl Lambrecht (part Nos. 322305 and
WPAC 2-22-BB400/700 respectively). The Wollaston prism is from
Bernard Halle (PWQ 30.30) and the camera lens is Nikkor
AF Telephoto 85 mm, f/1.8
We define the ratio
where 
, is
the fraction of the total light in linearly polarized condition. This ratio
reduces to the normalized Stoke's parameters 
and 
for 
and 
. In practice, additional
measurements are made at two more values of , namely 45
and 67.5, for reasons explained in Sect. 5.
The half-wave plate and the Wollaston prism are placed in between a field lens-camera lens combination, which reimages the telescope focal plane on to the main CCD with a reduction factor of about 3.8; the field lens reimages the telescope aperture on the half-wave plate and the light reaches the camera lens without any vignetting. As shown in Fig. 1 (click here) each point in the telescope focal plane produces two images on the CCD, corresponding to the ordinary and the extraordinary beams. In order to avoid overlap of the images of adjacent points, for observations of extended objects, a grid of parallel obscuring strips is placed at the focal plane of the telescope. The width of and spacing between the strips are chosen in such a way as to avoid overlap of the ordinary and the extraordinary images on the CCD. Four 0.3 mm diameter holes are provided at the four corners of the grid and are used to focus the surface of the grid, which coincides with the focal plane of the telescope, onto the surface of the CCD. The grid is made of black dielectric material to avoid polarization of the stray light arising due to reflections from its edges. Besides, the edges are made slanted (Fig. 1 (click here)) to prevent vignetting of the telescope beam.
The A&G unit has a field of view of about 2
and can be positioned
anywhere within a 
area at the focal plane of the
telescope close to the main field. This area has been chosen on the basis of
the requirement that three stars brighter than mv = 15
should be available with 95% confidence level towards the poles of the
Galaxy.
Figure 2: Various blocks of the IMPOL control system. The block marked "T"
represents the XY stages of the A&G unit
A block schematic of the IMPOL control system is shown in Fig. 2 (click here). The electronics assembly of the instrument consists of three parts - (i) the positioning systems for the half-wave plate and the A&G unit probe, (ii) the CCD camera for the A&G unit and the associated electronics and (iii) the main CCD camera with its own control electronics and host computer (PC 486) for exposure control and data-acquisition. The important parameters of this CCD camera, which was also developed at IUCAA, are listed in Table 1 (click here) (for more details refer to Deshpande & Gadre 1994).
| Parameters | Value |
| CCD make | EEV CCD02-06 series |
| CCD chip size | 385(H) 578(V) |
| Pixel size |  |
| Active area |  |
| No. of amplifiers | 1 |
| Quantum efficiency | blue: 20% |
| yellow: 50% | |
| red: 60% | |
| Readout speed |  per pixel |
| Acquisition & Display time | 12 s for full frame |
| Read noise (total) | 8  rms |
| Gain | 5 / ADU |
|
|