5.2. Instrumentation at Long Infrared Wavelengths
5.2.1. Mid-InfraRed/Far InfraRed Wavelengths
Long Infrared wavelengths start in the Mid-IR (~ 10 µm), cover the Far-IR (~ 100 µm) and eventually extend to the Extreme-IR (~ 1000 µm; submillimeter) wavelengths.
The first Mid-IR imaging polarimeter in the world, called NIMPOL, has recently been commissioned on the 3.9m Anglo Australian Telescope, and very limited observational results include the Galactic Center and Carinae (Smith et al. 1997). The preliminary dust emission data show that the magnetic fields are going parallel to the three main filamentary arms within 1 parsec of the Galactic Center (their Fig. 5), and that the magnetic fields are doing a figure-8 loop centered on the accretion disk of the Carinae nebula (their Figure 6).
The first observations in Far-IR polarimetry was made by Harwitt and his students in the late 1970's, toward the BN/KL object in Orion (e.g., Hildebrand 1996).
Hildebrand et al. (1995) found measurable polarization > 1% in many clouds, but usually less than 10%. In many clouds, the polarization position angle stays relatively constant over the clouds. The maximum polarization decreases with increasing 'optical' depth more rapidly than expected due to opacity alone, suggestive of additional factors (grain properties, grain alignment, inclination of magnetic field, unpolarized clumps). Davidson et al. (1995) found that the observed magnetic field in OMC1 could be due to a simple compression of an external large scale magnetic field, where the contraction of the cloud was predominently along the direction of the large scale magnetic field lines, and that the estimated strength of the magnetic field was strong enough ~ 3 m Gauss to have severely decreased the star formation rate in OMC1, by preventing gravitational collapse and by preventing the [nearby] HII region from more severely compressing the molecular cloud.
In the long IR régime, absorption by dust scattering becomes weaker at all the longer wavelengths, since scattering varies with the wavelength as -4 (e.g., Tamura et al. 1988; Bhatt & Jain 1992). Thus there is no significant absorption by dust scattering at Far IR and Extreme IR, as well as at millimeter wavelengths. Goodman (1995) argues in favor of Far IR and Extreme IR polarimetry (aligned dust emission) and against background starlight optical polarimetry, in order to map the magnetic field inside dense regions of the interstellar medium, because the physical size and type of grains inside dense clouds or clumps (larger, rounder on average, fluffier) are different than those outside clumps (smaller, more elongated on average).
5.2.2. Submillimeter/Extreme InfraRed wavelengths
In the long IR, emission by thermal radiation from dust particles becomes stronger from Extreme-IR to Far-IR wavelengths (flux density proportional to -4). This emission allows the study of magnetic fields in objects of moderate thermal gas densities. Lazarian et al. (1997) claim that far-IR emission polarimetry selects warm grains near young massive stars, and these grains are in an environment far from equilibrium, possibly rotating suprathermally because of radiative torques from stellar photons.
The first successful observations of polarization in the submillimeter/Extreme IR régime was made by Murray (1991), Flett & Murray (1991), with the JCMT in Hawaii at 450 µ, 800 µm, and 1100 µm.
At submillimeter wavelengths, the instruments are as follows. (1) A few polarimetric observations have been done near 100 µm, with the Kuiper Airborne Observatory's Stokes polarimeter until 1996, with a nominal resolution of 35" (e.g., Davidson et al. 1995; Hildebrand et al. 1995; Goodman 1995). (2) The 15-m JCMT on Mauna Kea in Hawaii offers polarimetry at 450, 800, and 1300 microns (the old Aberdeen/QMW polarimeter was used from 1990 to 1996; it has been replaced with the SCUBAPOL in 1998). (3) The 10-m CalTech Submillimeter Observatory in Hawaii has recently used the Hertz submillimeter polarimeter, (e.g., Schleuning et al. 1997). (4) A polarimeter for the SOFIA (replacing the KAO plane) is being planned, and the telescope should start around 2002. It is optimized for 70 µmeters, although flight times will be limited so only moderately bright sources will be observable. (5) A BOL polarimeter is being planned for the Far InfraRed and Submillimeter Telescope (FIRST), to be launched in space around 2005. It is optimized for 250, 350 and 500 µmeters (Greaves et al. 1997).
The few Far IR and Extreme IR polarimetric data so far available hint at a complex interplay between magnetic field directions and strengths and other large disk and cloudlet variables, such as self-gravity, turbulence, radiation field, and thermal instabilities. Efforts have been made to understand the patterns of the projected magnetic field directions associated with these objects, with limited success so far owing to the lack of a large polarization survey at Extreme Infrared (e.g., Myers 1991).
At present, there are few studies at far IR or extreme IR of the magnetic structures over a range of physical parameters, and they are biased towards regions with strong flux density values. Many of the present and future polarimetric programs aim to answer questions related to the magnetic field shape and directions in protostellar disks and interstellar cloudlets, by analysing the submm and mm emission from dust. Up to now, few polarimetric programs have run on submm telescopes, and new results are expected to add significantly to our knowledge.
Schleuning et al. (1997) made a 350 µm map of Orion OMC-1, finding that the percentage of polarization ranges from ~ 1% up to 7%, while the polarization angle is fairly the same across the map at ~ 30° (East of North). A 'polarization hole' (lower percentage of polarization) is seen at the peak of the total intensity emission (Stokes I).
5.2.3. Millimeter/Radio wavelengths
The first successful millimeter polarimetry was by Barvainis and his colleague in 1988 (e.g., Hildebrand 1996). At millimetric wavelengths, the instruments are as follows. (1) The portable Millipol instrument was used with the 12-m NRAO Kitt Peak telescope (Barvainis et al. 1988; Clemens et al. 1990), operating at 1.3 mm. (2) A Sapphire halfwave plate was attached to the FCRAO telescope (Novak et al. 1990) to operate at 1.3 mm wavelength. Both are owned by private groups (Millipol was built by a Boston-based group; the Sapphire half-wave plate was built by the Amherst-based group operating the 14m FCRAO telescope). (3) A polarimeter operating at 3000 microns at Kitt Peak has been provided for general use (e.g., Payne et al. 1993). (4) Polarimetry with an interferometer is also possible with the Owens Valley Array, which works at 1300 and 3000 microns.
Novak et al. (1990) found that the magnetic field near the bipolar outflow from the KL nebula in Orion is nearly parallel to the axis of the outflow. Cloud collapse is expected to occur along field lines, leading to a flattened gas disk; an outflow forming later in such a disk will expand more easily along the largest density gradient hence along the field lines. Barvainis et al. (1988) found at 1.3 mm wavelength a polarization percentasge near 3% in Orion, somewhat higher than the 2% found at 270 µm and 77 µm (but within the 2 rms error).
5.2.4. Typical Polarimetric Run
Figure 9. The primary 15-meter diameter surface of the JCMT on Mauna Kea, one of the few large facility designed to opeate at Extreme InfraRed wavelengths. Also shown are the small secondary surface as supported by legs, and at right the windscreen.
Figure 9 shows, as an example, one of the radio telescopes in Hawaii. The JCMT is chosen here because it is the largest (15m) facility in the world specifically designed to operate at Extreme IR wavelengths, The primary 15-meter surface (bottom left) collects incoming photons from space coming through the huge windscreen (top right). The photons thus collected on the 15-m surface are first re-directed to the secondary surface (supported by the spider-like legs, at top right), where now the photons are bounced downward through the gaping hole in the 15-m surface (bottom center). Inside the hole below the primary surface, several instruments are located to detect these photons, to analyse their polarization properties, and to make recordings for future uses. The polarimeter there has a rotating half-wave plate inside, followed by a wire-grid analyser. Thus the intensity I, as seen at any position angle of the half-wave plate, is given by (e.g., Vallée & Bastien 1996):
Rotating the wave plate angle through 360 degrees allows the detection of the unique angle source where the emission is maximum. This unique direction source is perpendicular to the direction of the magnetic field (perpendicular to the photon's polarization position angle source of maximum intensity). For computing, one often uses a quantitative representation known as the Stokes parameters I, Q, U, V, where
and V is the circular polarization. The Stokes parameters are linearly additive.
The observations of the polarization angle (PA) can provide a vital test of the model-predicted shape of the magnetic field. Let source be the photon's electric vector, coming from an interstellar source. To see if the source's magnetic field B is parallel or perpendicular to the elongation of an object (a 90° difference), at a 5 r.m.s. level (i.e., r.m.s. = 90/5 = 18°), one only needs to integrate until the 'error in polarization position angle' source is ±18°. In general, the probability function for source is approximatively gaussian and sharp-peaked, and the error in source follows the polarization percentage pobs according to the statistical law:
Thus for source = 18° , one needs to observe only long enough at the telescope until one reaches pobs / pobs = 1.6 (e.g., Greaves et al. 1995a; Greaves et al. 1995b; Greaves et al. 1994).