Telescopes And Detectors 141 Telescopes

The two main properties of a telescope are the diameter of the primary optical element (D), which defines both the light grasp and the limiting angular resolution, and the focal length, f, which determines the plate-scale, the number of arcsec per millimetre in the focal plane. The angular resolution, O, is set by the diffraction limit:

where A is the wavelength of the observations, and 206,265 is the number of arcseconds in a radian. Thus, the human eye, with an effective aperture of 5 mm, has an angular resolution of ~20 arcsec; an 8-inch (20-cm) telescope can resolve double stars separated by 1 arcsec; and the 200-inch (5-metre) Hale telescope at Mount Palomar has a diffraction limit of 0.02 arcsec at visual wavelengths, although atmospheric turbulence (seeing) prevents this resolution being attained in direct observations.

The focal length, f, of a telescope is the distance between the primary optical element and the principal focus. The plate-scale at that focus is given by

206,265 _1

The focal ratio of the system is the ratio of the focal length and the diameter of the primary, Rf = f /£. Thus, the 200-inch Hale telescope has a primary mirror with a focal ratio of 3.3 (written f/3.3), corresponding to a focal length of 55 feet (16.76 metres) and a plate-scale of 12.3 arcsecmm—

The majority of large telescopes built for astronomical research (and, indeed, for amateur work) are reflecting telescopes. The main optical element is a mirror made of low-expansion glass and coated with a thin, highly-reflective layer of aluminium or gold. The primary advantages of a reflector are ease of construction, and cost. The objective lens in a refracting telescope must be edge-supported, and therefore must be sufficiently rigid to minimise deformation as the telescope is slewed. The structural strength of glass sets an upper limit of ~ 1m to the diameter of refractors. In contrast, with reflecting optics the primary mirror is back-supported, and it is therefore possible to make larger-aperture telescopes - up to 8 m for monolithic mirrors (such as the Gemini telescopes and the ESO Very Large Telescopes) and 10 m for segmented-mirror telescopes (the Keck telescopes on Mauna Kea). Plans are being made for 30m+ segmented mirrors (e.g., T2).

Not only is it substantially more expensive to fabricate high-quality transmitting optics than to make reflecting optics of the same aperture, but mirrors can be ground to faster optical ratios, leading to shorter focal-length telescopes, and hence smaller enclosures (domes). Each of the Keck 10-m telescopes, for example, has a primary mirror with a focal ratio f/2, producing a focal length of only 20 m, and allowing the telescope to be accommodated within a dome little larger than that of the Palomar 200-inch. On the other hand, the largest refractor constructed, the 40-inch Yerkes telescope at Williams Bay, Wisconsin, has a focal ratio of f/19 and a focal length of 63 feet - 8 feet longer than the 200-inch telescope. Very few large refracting telescopes have been constructed since the end of the nineteenth century.

Until recently, most telescopes were mounted equatorially: that is, a rotation axis is aligned with the Earth's polar axis, so motion due to diurnal rotation can be eliminated by tracking in only one co-ordinate (RA). However, it is simpler mechanically to construct an alt-azimuth telescope, where a rotational axis is aligned with the local gravitational field. Advances in computer technology now permit accurate tracking in two co-ordinates, taking full account of effects such as telescope flexure and refraction, and most telescopes built in the last 15 years have alt-azimuth mounts.

The standard telescope design is based on the Cassegrain system: a secondary mirror (B) brings the light to a focus (C) at a point below the primary (A) (Figure 1.7). This provides a stable mount for moderate-sized instruments. In conventional telescopes, the distance AB is close to the focal length of the primary mirror, principally to minimise the diameter of the secondary mirror, and hence the size of the obstruction of the primary beam. This positioning requires that the secondary be suitably figured, which changes the effective focal ratio. Typical values are f/8, although ratios as high as f/400 are possible, with a consequent increase in the plate-scale. Alternatively, in many telescopes the secondary can be removed, and observations undertaken directly at the prime focus, P, at the primary focal ratio of f/2.5 or f/3. The latter focal station gives the largest plate-scale (most arcsec/mm), and thus provides easy coverage of large solid angles, but only limited space for p p

Figure 1.7. Schematics of optical telescopes: {left), a conventional Cassegrain reflector; {right), a Schmidt telescope.

Figure 1.7. Schematics of optical telescopes: {left), a conventional Cassegrain reflector; {right), a Schmidt telescope.

mounting instruments. Large instruments obstruct the primary mirror and require a mechanically rigid telescope mount.

The classical Cassegrain design has a parabolic primary mirror. This has the disadvantage of producing off-axis images with significant optical aberrations: coma and spherical aberration [J1], [W2]. These can be corrected to some extent using a lens {corrector plate), but for only a limited field of view {^ 0?3 in diameter). The Ritchey-Chretien modified Cassegrain design has a hyperbolic primary which, combined with the appropriate secondary, results in less distortion and a larger useable field of view {up to 1 square degree). Most telescopes built since 1950 have Ritchey-Chretien primary mirrors {see [W2] for an historical review and thorough discussion of contemporary designs).

Two other foci deserve mention. Both require the introduction of a third mirror at D {usually with additional optics) and direct the beam to larger instruments placed at stationary locations. In equatorial telescopes, the tertiary mirror sends the beam down the polar axis to the coude room. With a typical focal ratio of f/45, the plate-scale is usually ~ 1"/mm, enabling very high spectral resolution {see Section 1.7). In alt-azimuth telescopes, the equivalent focal positions are the Nasmyth platforms, mounted alongside the altitude rotational axis.

Most conventional telescopes, even with Ritchey-Chretien primaries, are limited to relatively small fields of view, covering no more than square degree {or 0.0025% of the sky) at the prime focus. For most of the twentieth century, those telescopes were ill-suited to surveying large fractions of the celestial sphere - the survey would simply take too long using the only available detectors, photographic plates. (Circumstances have changed recently, as outlined further in Section 1.6.3.) Specialised wide-field telescopes were designed to deal with this issue. The most successful of these survey telescopes is the Schmidt telescope, named after Bernhard Schmidt [O3]. The Schmidt primary (A in Figure 1.7, right) has a spherical surface, which would normally introduce spherical aberration. However, a corrector lens5 (C) placed at the centre of curvature of the primary, can 'pre-correct' the beam, leading to high-quality images over a field of view several degrees in diameter. The focal plane, F, is a spherically curved surface, and detectors (principally photographic plates) must be shaped to conform with that surface.

Finally, with the advent of satellite observatories in the early 1960s, it has become possible to survey the sky at wavelengths which are absorbed or scattered by the Earth's atmosphere. Longer wavelength (infrared, far-infrared) and ultraviolet telescopes (such as IRAS, ISO and IUE) are similar in design to their ground-based optical counterparts. X-rays, however, are not reflected when they strike a mirror at normal incidence (perpendicular to the plane of the mirror) but are reflected when they strike at grazing incidence (that is, angles of only 1 or 2 degrees to the surface). If the mirrors have suitable parabolic and/or hyperbolic figures, then the radiation can be collected and focused as in an optical telescope [C2], [T1]. Unlike optical telescopes, these mirrors are often made from (or at least coated with) high atomic-number metals. Thus, the primary reflector in an X-ray telescope consists of a metal sleeve at the top-end of the telescope (Figure 1.8), rather than a conventional mirror. Recent X-ray missions include Einstein, ROSAT and Chandra.

1.4.2 Detectors

Astronomical observations require the detection of light at extremely low intensity levels. This in turn demands detectors which have both high quantum efficiency (QE)6 and low noise characteristics, so that a high fraction of the photons striking the detector are recorded, and the detector does not degrade the signal significantly in the act of recording. Photography was a mainstay of astronomical imaging and spectroscopy for more than 150 years after its invention. Photographic emulsions are sensitive primarily to radiation in the wavelength regime 2,000-9,000 A, while photographic plates can be manufactured at a size large enough to take full advantage of the wide field of view offered by Schmidt survey telescopes. The overall QE, however, is only ^2-5%, and over the last two decades other more sensitive detectors have superseded photography in essentially all other roles.

Most modern optical and infrared instruments use semiconductors as detectors. Those devices rely on the photoelectric effect to detect radiation. As mentioned in

5 The aperture of a Schmidt telescope is the diameter of the corrector plate, not the primary mirror.

6 The QE of a detector is the fraction of incident photons which are detected as photoelectrons. Thus, if 100 photons fall on a detector and only five are detected, the QE is 5%.

Figure 1.8. A schematic of an X-ray telescope. P1 and P2 are grazing-incidence paraboloidal mirrors, and H1 and H2 are hyperboloids.

Section 1.2, a photon of frequency v has energy hv. If this energy exceeds the binding energy of an electron of an atom in the semiconductor, then the electron is released. Counting the number of free electrons (photoelectrons) emitted within a given time provides a measure of the number of photons striking the photosensitive material; that is, the brightness of the light source. In the original photometers, a voltage was placed across the detector, and the signal detected as a current. Even the brightest sources generate very low currents, however. At optical wavelengths, the signal can be amplified using a photomultiplier tube. A photosensitive material, such as Cs3 Sb or GaAs, held at a high negative potential, serves as the detector, the photocathode. The potential gradient (typically ~ 1 kV) leads to photoelectrons striking a series of electrodes (dynodes) in a vacuum tube (Figure 1.9), releasing many secondary electrons at each strike and producing a pulse at the anode. This amplification results in a detectable current spike.

Proportional counters operate in a similar fashion in detecting X-rays, although the primary detector is a high atomic-number gas rather than a photocathode. Again, a high-energy photon interacts with an atom, leading to the emission of a photoelectron, which is accelerated in a potential field, colliding with other atoms to produce a cascade of secondary electrons. Unlike the optical photomultiplier, proportional counters can provide some information on the frequency of the photons detected. The higher the frequency, the greater the energy and the larger the number of photoelectrons produced. Comparable detectors, with limited spectral resolution, are being developed for work at optical wavelengths.

The first photon-counting detectors provided no positional information, only the total photon rate within an area of sky defined by either the field of view of the telescope or by an aperture mask in the focal plane. Two-dimensional imaging detectors were developed in the 1970s: notably, the Imaging Photon Counting System [B5], but also reticon arrays and multichannel-plate detectors [E1]. These photocathode photons photoelecti __

dynodes anode


focusing electrodes

Figure 1.9. A schematic of a photon-counting detector.

provided spatial information either by accurate timing or by subdividing the photocathode into separate regions, each with its own photomultiplier chain. As with single-channel devices, each photon is detected separately, and these devices are very efficient in working at very low light-levels, particularly at blue wavelengths where the photocathodes are most efficient (20-30%). However, detecting each pulse of secondary electrons takes a finite time (typically a few microseconds), and the system cannot identify separate pulses due to other (coincident) photons arriving within that deadtime. Thus, the photon count-rate for bright sources is underestimated; indeed, if the flux level is too high, the resultant high current can damage the detector. This coincidence problem means that photon-counting devices are ill-suited for broadband, wide-field imaging, where multiple sources inevitably lead to high photon-rates, while the low efficiency of photocathodes at wavelengths longer than ~ 6,000 A limits their utility at red spectral regions. Given these limitations, recent instrumental development has centred on charge-coupled devices (CCDs) - array detectors which possess neither of these shortcomings, and which have become the standard detector in optical and infrared astronomical equipment.

A CCD consists of an insulating material bonded onto a doped silicon semiconductor (Figure 1.10). A grid of electrodes is embedded within the insulator, with each electrode held at a positive potential. When the device is exposed to light, freely-moving photoelectrons are produced in the doped semiconductor, and these (or the positive 'holes') migrate towards the local electrode. The efficiency is increased, and the noise contribution from random motion of electrons (dark current) reduced, if the device is cooled to temperatures of ~ 100-170 K, usually using liquid nitrogen or liquid helium as a refrigerant. The total charge collected by each electrode is a measure of the photon flux incident on that 'pixel' (picture element) of the detector. By manipulating the voltages appropriately, the charge collected in each pixel can be read out at the end of an exposure, and hence the image is reconstructed. Reading out the CCD chip contributes additional noise to the counts detected in each pixel (readout noise), but this is a small electrodes

Was this article helpful?

0 0

Post a comment