Some kinds of observations must be done above the Earth's atmosphere, and almost any kind may be done better there. Specifically, the reasons for doing space astronomy, despite all its difficulties and high cost, are
(1) Getting above the atmospheric absorption that blocks most kinds of radiation from the Earth's surface. For X-rays, this is tens of km altitude. All measurements of UV, far-IR, X-rays, and γ-rays below about 1 GeV are in this category. Dave Hanes' chart of atmospheric transmission
(2) Getting above atmospheric turbulence. This induces not only blurring, but intensity scintillations for small apertures, an important source of noise in photometry. Even tiny instruments in space can beat the photometric stability possible from the ground.
(3) Reduce the "sky" background by eliminating airglow and thermal emission. For locations near the Earth, visible-light observations still suffer from zodiacal light and starlight scattered off galactic dust. Sources of light in the sky include, beyond the Earth, sunlight scattered by dust in the inner solar system (zodiacal light, radially limited to within the outer edge ofthe main asteroid belt), thermal emission from these same grains in the far-infrared, starlight scattering off interstellar dust and thermal emission from these grains, unresolved emission from faint stars and galaxies, and the cosmic microwave background. The minimum sky background falls in the mid-UV, about 1500 Å , so that uniquely deep studies of star formation in galaxies and other high-temperature populations are possible. Sources of background light are reviewed in The Light of the Night Sky by F.E. Roach and J.L. Gordon (Reidel 1973). Specific contributions of various sources are given in Leinert et al. (1997), A&AS 127, 1 (a 99-page contribution).
(4) Eliminate problems in observing sources at times when they are near the horizon or otherwise inaccessible due to the geometry of Earth and Sun. Furthermore, there are no clouds. A satellite can get 40 hours' nonstop data on a variable source, or a measurement every 16 hours for two months, and not lose any data to weather.
Skipping for now such temporary platforms as balloons and sounding rockets, viable sites are orbital. We may distinguish low Earth orbit (LEO), high Earth orbit (HEO), geosynchronous Earth orbit (GEO), and others such as lunar surface, lunar orbit, and solar orbit.
LEO: the heaviest payloads can always be placed in lower orbits. The lowest reasonably long-lived orbits have periods near 90 minutes (period is given from Kepler's laws, or the Newtonian gravitation law plus centripetal acceleration, as P = 2 π (GM)-1/2 (R+h)3/2 for altitude above the surface h and Earth radius R, Earth mass M). If access for servicing, refuelling, or reboosting is required, LEO is the only option since current spacecraft are limited to h < 800 km, and payloads attached to space stations (such as the Mir/Kvant complex, or the all-sky MAXI X-ray system on the Japanese segemnt of ISS) are similarly limited.
Low orbits have some significant drawbacks. The Earth blots out nearly half the sky, and which half changes rapidly (perhaps too rapidly for the pointing system to retarget). The thermal background from Earth emission is a significant load on cooling systems for far-IR work - oddly enough, since one can reflect much of the Sun's radiation in the optical, the thermal IR contribution of the Earth can be dominant from LEO. There are, in addition, orbital lifetime considerations. The drag due to the upper layers of the atmosphere limits orbital lifetimes, and is important even above 500 km altitude. Keeping things there (such as HST and CGRO) requires periodic reboosting, from onboard propulsion or visits. Finally, the residual atoms at these altitudes (notably O) give off collisionally-excited line radiation upon encountering an object at orbital velocities (a phenomenon first known as shuttle glow). Thus, in the deep UV, one wants to avoid facing the instrument into the oncoming "wind" that produces this glow (into the so-called ram vector), whose direction in celestial coordinates changes constantly during an orbit.
HEO: anywhere from LEO to the orbit of the Moon, loosely defined. Periods of 3-100 hours have been used. Solar panels are more effective since the spacecraft is rarely or never shadowed. A high enough orbit keeps the spaceraft out of the van Allen belts, especially the enhanced low-altitude piece known as the South Atlantic Anomaly. The Earth obscures a smaller fraction of sky here, and the motion of the obscured region is more manageable. A particularly interesting case is the sun-synchronous orbit, useful for sky surveys as well as terrestrial observation systems. This is an orbit that keeps the same orientation relative to the Earth-Sun vector, using the oblateness of the Earth's mass distribution to precess the orbit once per year in the appropriate direction. This means that, for a spacecraft pointed outward above the terminator, fixed solar panels will always face the Sun with no need for large-capacity batteries to carry through orbital night, and the field of view (if wide enough) will sweep over the entire sky every 6 months. This has been used for the IRAS, Akari, and WISE survey missions.
For an oblate spheroidal planet with appropriate spherical-harmonic moment J, the orbit precesses by Δ ψ = 2 π J (R/r)2 cos i per orbit, where r is the semimajor axis, R the Earth's radius, and i the orbital inclination to the equator. For the Earth, J = 1.637 X 10-3, so this becomes numerically Δ ψ = 0.0121 (R/r)2 cos i. This was derived by Blitzer, Weisfield, and Wheelon, J. Appl. Phys. 27, 1141 (1956), and detailed by Thomson, Introduction to Space Dynamics (Dover, 1961, 1986). As a specific example, IRAS was in a 103-minute orbit at an altitude about 900 km, so (r/R)=1.23. This gives the required precession as 2 π radians per year = 2 π radians / 5106 orbits = 1.23 X 10-3 radians per orbit, so the inclination must be 98.85 º, slightly retrograde to have the direction of precession correct. Even though the orbit appears to be tied to the Earth's coordinate system, the wrapping around the celestial sphere occurs in ecliptic coordinates, since the orbital plane is instantaneously tied to the Earth-Sun vector.
GEO: a 24-hour orbit will always be accessible from a single ground station (though except for a circular orbit on the equator there will be apparent motion in a figure-8). This means that interactive contact is always possible, allowing maximum flexibility (as in IUE). The spacecraft passes through the Eath's shadow only during two periods per year (possibly requiring power management and interrupting operations). Distance from the Earth is advantageous at some UV wavelengths, since we are surrounded by a corona of H and He that emit strongly in their UV resonance lines. The deepest sensitivity is obtained by looking down the Earth's shadow, reducing the line emission to what comes from the interplanetary medium and heliopause (a strategy employed by EUVE and to some extent by GALEX).
Other locations: in the far-IR, distance from the Earth improves thermal loading so much that a solar orbit with P ~ 1.01 year is used for Spitzer to keep the spacecraft within useful range but away from the Earth. This also has energy advantages, being easier to reach than some high Earth orbits. Lunar orbit provides the only nearby possibility of doing low-frequency radio measurements, since aurorae and the van Allen belts are strong sources of km-wavelength emission. The lunar surface has advantages of stability and ease of engineering, at the cost of losing half the sky to each site (described by Dan Lester as "gravity and dirt"). Plans for lunar observatories have proceeded on the theory that if someone is determined to build a lunar base, they might as well do some astronomy.
Space astronomy provides stringent requirements on equipment properties and performance. Instruments must be made lightweight, yet able to withstand the stresses and vibrations of launch. They must be highly reliable, since repairs are impossible or extremely expensive; designers are understandably allergic to moving parts (although I am horrified at how often they are willing to cycle power on their systems). Attitude determination and control must be accurate in the absence of human intervention. There is a premium on low consumption of fuels or reaction gases, since these often limit the lifetime of a mission, so cryogenic cooling is often impossible. Provision must be made for adequate power, usually with solar panels and battery backup systems, and for dealing with a variable thermal environment. Communication systems may be complex, requiring relay satellites and high-speed data dumps in brief windows. The orbital environment is rich in radiation and charged particles, so that hardening and anticoincidence are needed to detect as nearly as possible photons alone. Strict shielding has a huge weight penalty, and may produce as many secondary events as it blocks in primaries.
It is always crucial to know where the telescope is pointed, and often crucial to be able to control its pointing to similar accuracy. Gyroscopes can give arcminute or better pointing accuracy, but to update their frame in an inertial one, to effect finer-scale pointing, and as reference for fine-scale corrections, the stars themselves must be used. Star trackers on small telescopes are frequently used in initial pointing and attitude updates, with accuracies of an arcsecond possible over fields of a degree or more by use of centroiding and detector-element edges; a sample system is described by McQuerry, Deters, and Radovich 1990 (Advances Astronautical Sci. 72, 83. Tracking through the telescope (for optical and UV instruments) can be used for fine attitude control. X-ray and EUV detectors register photons individually without substantial noise, so that an after-the-fact reconstruction requires precise attitude determination but less precise control. An illustration of this is the Einstein slew survey; the detectors were left on as the spacecraft moved between targets, so data were collected over long strips of sky. More recently, Chandra and GALEX routinely execute small motions during an exposure, to smooth out residual calibration effects as well as exposure to strong sources. Control of attitude may use reaction wheels against the gyroscope platform; reaction jets are a last resort since they impart a large impulse and use a consumable substance. Near Earth, pointing and stability may use electromagnetic torque against the Earth's magnetic feld, needing no consumables (MOST on purpse, FUSE as a last resort).
Target acquisition is usually by position alone, perhaps assisted by some onboard peak-finding or offset routine. A few satellites have manual feedback capability for difficult cases or to check the inertial system orientation.
The interstellar medium imposes limits on the propagation of radiation that restrict what we can see, even from space. These include reddening, photoelectric absorption during ionization of hydrogen, and low-frequency plasma absorption. Interstellar grains absorb UV radiation most efficiently, so that much of our galaxy is obscured in the ultraviolet. More important, at energies above 13.6 eV (wavelengths less than 912 Å), ionization of hydrogen removes radiation from space with a cross-section varying approximately as ν-3 near the Lyman limit. Column densities of less than 1020 atoms/cm2 will block radiation down to 50-100 Å , effectively separating the UV and X-ray regimes. Fortunately for astronomy, there turn out to be many "holes" in the galactic H I layer through which we can see considerable distances at wavelengths of a few hundred Angstroms. A few extragalactic objects are even found in the extreme UV. Finally, at wavelengths longer than 1 km or so, plasma absorption in the ISM soaks up radiation very effectively.
There are some initial test efforts worthy of notice. The first UV spectrum of a star beyond the Sun was obtained by putting a diffraction grating in front of a camera and holding it out the open hatch of a Gemini capsule. Notable test exposures in the EUV were done on the Apollo-Soyuz flight. A UV Schmidt camera was used on the lunar surface during Apollo 16. More recently, space VLBI was demonstrated using a TDRSS relay satellite as the extratarrestrial station, reaching baseline of several Earth diameters and demonstrating the feasibility of what became the HALCA program. Many of the first test observations were made from sounding rockets, during flights spending only a few minutes at the appropriate altitudes.
Last changes: 2/2011 © 2000-2011
