Introduction
This article describes the astrophysical questions that can be addressed at infrared wavelengths, the advantages of pursuing infrared astronomy from space, the enabling technologies, and the missions that have been flown and are planned to exploit the unique potential of this wavelength range. The infrared band covers three decades—from ~1 mmto ~1000 mm—that encompass a very wide range of instrumental techniques and scientific issues. We have confined our discussion to rocket- and satellite-borne missions aimed primarily at astro-physical targets and have omitted the infrared instruments that have been carried on planetary probes. We draw the short-wavelength limit of our detailed discussion at 2.5 mm but note that the Near Infrared Camera and Multi-Object Spectrograph (NICMOS) instrument on the Hubble Space Telescope (HST) have operated very successfully in the 1 to 2.5-mm band; these wavelengths are also among those studied by the Diffuse Infrared Background Experiment (DIRBE) on the Cosmic Background Explorer (COBE). Wavelengths longer than ~200 mm have been probed extensively from space only by the highly successful COBE spacecraft. We discuss COBE observations of the nearby Universe but cannot do justice to its extraordinary spatial and spectral measurements of the cosmologically critical cosmic microwave background radiation (CMBR).
The organization of the article is as follows. An introductory section describes the astrophysical uniqueness of the infrared and the tremendous benefits to be gained by carrying infrared instrumentation above the atmosphere. The next section discusses three key technical areas of particular importance for infrared astronomy—detectors, cryogenics, and optics—and the way they have been adapted to the space environment. Next, comes a review of previous and near-term missions for infrared astronomy from space, emphasizing the evolution of both science and technology since the first rocket experiments of the 1960s. The final mission we discuss in detail is the Space Infrared Telescope Facility (SIRTF), scheduled for launch in 2003 (1). SIRTF’s advanced technology and great scientific potential will the end of the first phase of exploration of the Universe in the infrared band. But SIRTF is also a beginning, because it sets the stage for the missions of the next decade. We conclude with a summary of the technological challenges and scientific opportunities of these upcoming programs.
Uniqueness of Infrared
Infrared observations provide the following unique perspectives on the Universe: The Cold Universe. There is an inverse relationship between the temperature of an object and the peak wavelength 1 of its intrinsic or blackbody radiation: T(1) = 3700/1. Here T is measured in degrees kelvin (K) above absolute zero, and 1 in microns (mm). Objects with T<T(X) radiate very little at wavelengths less than 1. Observations at infrared wavelengths from 1-1000 mm are thus uniquely sensitive to astronomical objects whose temperatures are from ~3000K to ~3K. These include the coolest stars, planets and interplanetary dust, circumstellar and interstellar matter, and, at the longest wavelengths, the Universe itself.
The Dusty Universe. Interstellar dust—microscopic particles composed of ices, minerals, and common organic and inorganic materials—is a ubiquitous constituent of astrophysical environments. The properties of this material are such that a cloud that is totally opaque in the visible or ultraviolet can be virtually transparent in the infrared; thus infrared wavelengths can probe regions—such as the core of our galaxy—which are inaccessible at shorter wavelengths. Additionally, the dust particles are heated by the shorter wavelength radiation they absorb and reradiate the absorbed power at infrared wavelengths. The majority of the radiant energy from dense, dusty regions such as star-forming clouds—and in some cases from entire galaxies—lies at infrared wavelengths because of this efficient downconversion process. The Distant Universe. In the expanding Universe, the more distant an object is, the greater the velocity at which it recedes from us. This cosmic expansion shifts the starlight from distant galaxies into the infrared; the more distant the object, the farther out into the infrared. This expansion is characterized by the redshift parameter z:1 + z = (observed wavelength/emitted wavelength). The most distant known objects have z > 6, so that radiation from the middle of the visual band is shifted out beyond 3 mm. Because 1 + z is also equal to the factor by which the Universe has expanded between the times of emission and absorption of the radiation, objects at z = 5 are seen as they were at an epoch when the Universe was only one-sixth of its present size.
The Chemical Universe. The infrared band contains the spectral signatures of a variety of atoms, molecules, ions, and solid substances—some of which will be found in any astrophysical environment. Examples range from cool ices in the interstellar medium to highly excited ions in active galactic nuclei. Infrared spectroscopy can isolate these features, determine their absolute and relative strengths, and provide an important and often unique probe of the chemical and physical conditions in these systems.
The Advantages of Space
The space environment presents powerful advantages for conducting infrared astronomical observations, which motivate the technological developments discussed below. First, in space, one is free of the absorption by Earth’s atmosphere, which—even from the best mountaintop observatories—is totally opaque at wavelengths from ~ 30 to ~ 300 mm (2). Outside of this region, there are other bands of high and moderate opacity, and atmospheric absorption remains appreciable at aircraft and balloon altitudes. Only from space do we have access to the entire infrared band. A second, equally fundamental benefit is that a space observatory is free of the blackbody radiation of Earth’s atmosphere, and the space telescope can be cooled to low temperature to minimize its own blackbody radiation without fear of atmospheric condensation. Infrared observations from Earth are limited by very bright foreground radiation from the atmosphere and the ambient temperature telescope; in space using a sufficiently cold telescope, the limiting background—set by the faint glow of the interplanetary zodiacal dust cloud—is some six orders of magnitude fainter. This is about the same factor by which the night sky at new moon is fainter than the daytime sky at high noon; note that optical astronomy is practiced at night, not during the day. The impact of this million-fold background reduction, in space, is impossible to overestimate because it produces a thousand fold increase in sensitivity, or a million fold increase in the speed of observations. Thus the first major cryogenic infrared space observatory, the Infrared Astronomical Satellite (IRAS), revolutionized our knowledge of the infrared sky, even though it observed each point for less than about 20 seconds during its 10-month survey of the sky (3).
Complementary Approaches
Infrared astronomy is pursued very successfully from ground-, aircraft-, and balloon-borne platforms, and these sites present opportunities and capabilities complementary to the very high sensitivity and spectral access of space. The current state of the art for ground-based infrared astronomy is a series of 8- to 10-m diameter telescopes in Hawaii and Chile that provide ongoing scientific opportunities, much higher spatial resolution than achievable from space at present, and a greater variety of focal plane instrumentation—including complex spectroscopic instruments—than typically available in space observatories. Airborne observatories—exemplified by the imminent ~3-m-class Stratospheric Observatory for Infrared Astronomy (SOFIA)—provide capabilities similar to those of large ground-based telescopes in most of the wavelength bands that are inaccessible from the surface of Earth. Balloon-borne instruments have been successful for specialized measurements, most notably survey observations and studies of the CMBR, and will have an important niche in the upcoming era of long-duration balloon flights.
Detectors and Detector Arrays
Modern detectors fall into two classes, bolometers and photoconductors. Bolometers are devices that change resistance when heated by absorbed radiation; they respond to radiation across a wide wavelength band, as long as they effectively absorb across this entire band. Bolometer technology has advanced dramatically in the past few years due to the application of modern semiconductor processing techniques. Photoconductors are solid-state devices in which incident photons can excite electrons from a bound state—a valence band or an impurity level—into a conduction band to produce a current in response to an applied voltage. Because the valence or impurity levels and the conduction bands are separated by a well-defined energy gap, the energy or wavelength range within which photoconductors respond to radiation is restricted. To cover a broad infrared wavelength range, a combination of different photoconductors generally needs to be employed. The current materials of choice are InSb and HgCdTe photodiodes for wavelengths shorter than ~10 mm, extrinsic (doped) silicon photodetectors (Si:xx) for wavelengths from 5-40 mm, and extrinsic Germanium (Ge:xx) for ~40-200 mm. In this nomenclature “xx” denotes the specific dopant. The preferred dopants for silicon detectors are As, B, Ga, and Sb; Ga, Be, and Sb have been used as dopants for germanium. Bolometers are the detectors ofchoice for wavelengths longer than 200 mm and for some applications at shorter wavelengths as well. Infrared detectors for space astronomy are discussed in detail in two recent topics (4,5).
All detectors are inherently noisy. They register the incidence of arriving photons, generally referred to as “signal,” and also any number of other types of events classed as “noise.” Notable among the noise sources are thermally excited conduction electrons, or ”dark current” in photoconductors, and thermal fluctuations in bolometers. Both sources of noise can be reduced by cooling the detectors to temperatures so low that thermal effects become negligible. In real-life applications, both types of detectors can be degraded by the electronics required to operate them and read them out, although modern circuit design techniques and on-chip integration generally allow minimizing these effects. In space, noise can also be generated by high-energy cosmic rays that traverse the detectors. Effective shielding against such particles becomes a high priority; in addition, special fabrication techniques can be used to reduce the susceptibility of the detectors to this ionizing radiation, as has been done with extrinsic silicon pho-toconductors of the type to be used on SIRTF.
A recent major advance in infrared detectors is large arrays of many active elements, or pixels, bonded to a multiplexer that is used to sample and read out the pixels. The impact of this technology on space infrared astronomy—which is similar to the CCDs used in the visible band—will be very dramatic. Used with photoconductors in the low-background space environment, the technology permits on-chip integration, so that the signal can be accumulated on the detector array and read out only when it is large enough to overcome electronic noise. Clever schemes involving multiple, nondestructive readouts of the array have been devised to suppress electronic noise further. Detector arrays are equally applicable to imaging and to spectroscopic instruments—and to photoconductor and bolometer technology, and they will certainly be used very extensively, if not exclusively, for space infrared astronomy in the future. There will continue to be a push for larger format arrays, and the next generation of infrared space experiments should use arrays of at least 1024 x 1024 pixels, a substantial advance over the 256 x 256 pixel arrays to be used on SIRTF.
The ultimate performance goal for detectors for infrared astronomy is that they permit ”background-limited” observations, that is, the intrinsic detector and electronic noise should be less than the noise due to the statistical fluctuations in the rate of arrival of photons from ambient and astrophysical backgrounds. Modern infrared detectors achieve this readily in ground-based applications where the warm telescope and emissive atmosphere produce very high backgrounds. For space applications using cryogenic telescopes, the infrared background is that due to the zodiacal dust within the solar system, which is at least a million times fainter than the ground-based foreground sky. Achieving background-limited performance in this environment is quite challenging, even with the benefits of on-chip integration; for example, observations in the 3 to 5 mm window require that dark current and electronic noise contribute less (often much less) than the equivalent of one electron/second/pixel (6). Detector technologists and astronomers working together have responded to these challenges and improved the performance of infrared detectors by many orders of magnitude in the past two decades. The improvement has come from reducing the noise and also by improving the ”quantum efficiency”—the fraction of incident photons that is absorbed by the detector. As a result, the arrays to be used on SIRTF will achieve background-limited performance for both photometry and low-resolution spectroscopy at all wavelengths.
Cryogenics
Space infrared telescopes invariably require efficient cooling or cryogenic systems; the telescope and the surrounding structure are cooled to reduce their background radiation, and the detectors are cooled to reduce their intrinsic noise and increase their sensitivity. In many applications, these effects together require cooling below 10 K. In many of the rocket and satellite instruments built to date, the entire telescope has been cooled to temperatures as low as 2 K, where no part of the apparatus emits as much radiation as the 2.73 K CMBR. This has been done by placing the entire telescope structure in direct physical or thermal contact with a pumped-liquid-helium bath, and the vacuum pump is the natural vacuum in space. The IRAS and Infrared Space Observatory (ISO) systems used this architecture, as shown in Fig. 1 (left). Liquid helium is required to achieve temperatures below ~ 5 K, but other stored cryogens that provide more cooling power per unit mass are used in applications where higher temperatures are acceptable. In practice, this is equivalent to reducing the long-wavelength limit of the instrument. Other cryogens that have been used—and the approximate temperature that they provide—are solid hydrogen (8 K), liquid neon (30 K), solid nitrogen (50 K), and liquid nitrogen (75 K).
A primary design problem for the cryogenic engineer is to minimize the heat load on cooled surfaces of the apparatus. A first step is to shield these surfaces from the principal heat source, which is solar radiation, and to use suitable combinations of low- and high-emissivity materials to reduce heat transfer within the satellite. A next step is to blanket the container that holds the cryogen with dozens of layers of aluminized mylar loosely packed within a vacuum jacket to isolate the entire system from its ambient-temperature surroundings before launch. Such a vacuum-packed cryogenic system is referred to as a dewar, named for the nineteenth-century Scottish scientist J. Dewar. A well-designed cryogenic system also uses the cold effluent gas generated as the cryogen evaporates or sublimes to cool the surrounding structures and further reduce the heat load on the cryogen.
A second design challenge is the construction of apparatus sufficiently sturdy to survive launch and also to maintain optical alignment between the cooled telescope and the guide telescopes or gyroscopic components that typically operate at ambient temperature within the spacecraft bus. The mechanical rigidity required tends to go hand in hand with high thermal conductivity, which the design must avoid; this requires using low-thermal-conductivity materials that have high strength, such as epoxy-glass composites.
Considerable attention has also focused on minimizing heat loads on the cryogen by passively radiating intercepted heat into cold space; this is referred to as ”radiative cooling” (7). The SIRTF telescope, described later, exploits the favorable thermal environment of its heliocentric orbit by using a hybrid cryogenic system in which the instruments and detectors are cryogenically cooled, whereas the telescope is launched warm and is cooled by a combination of radiation, conduction, and effluent cryogen (Fig. 1 right). This approach has many advantages over that used in earlier missions such as IRAS and ISO, in which the entire telescope was placed within the cryostat. It leads to a lower mass cryogenic system for a fixed telescope size and decouples the size of the telescope from that of the cryostat. Thus this hybrid approach is certain to be adopted for large infrared telescopes in the future.
Figure 1. This figure compares the cold launch architecture used for the ISO and IRAS observatories (left) with the warm launch architecture to be used for SIRTF (right). Each is shown in cutaway view. Certain components, such as the spacecraft (SC), the solar panel (SP), and the startracker (ST) are common to both systems. In addition, the telescope (T) and instrument package (I) are identical in size for the two. Each also includes a cryostat (C), containing the liquid helium cryogen in a separate helium tank, which is shown shaded. In the cold launch system, the telescope is located within the cryostat and cooled by direct contact with the cryogen tank. In the warm launch system, the cryostat and cryogen tank can be much smaller, and the telescope is cooled by conduction and by the cold boil-off helium gas. This architecture works in the solar orbit because the cylindrical thermal shields that surround the telescope cool radiatively to 40 K or below, so there is very little parasitic heat diffusing inward toward the telescope. The cryostat must withstand atmospheric pressure, and it is much larger and more massive for the cold launch than for the warm launch system. In the former, it surrounds the entire telescope and supports a heavy vacuum cover (VC). In the warm launch system, the telescope is launched at ambient temperature and pressure, protected only by a lightweight dust cover (DC). The sawed-off conical sunshade at the top of the cryostat is required in an Earth-orbiting system by the Sun-Earth-orbit geometry. A much smaller sunshade is need in the solar orbit system because Earth is not a concern.
In a more extreme application of radiative cooling, it may be possible to cool the entire telescope to a temperature acceptable for many purposes without using cryogens. Just how low a temperature can be reached in practice is still not clear, but many designers now assume that equilibrium temperatures as low as 30 K could be within reach at 1 astronomical unit (the radius of Earth’s orbit) from the sun. Lower temperatures might be achieved by a telescope operating in the outer solar system. At such low temperatures, a well-designed telescope that has exceptionally low-emissivity mirrors might radiate at such low levels in the wavelength range shorter than 100 mm that the primary and secondary mirrors require no active cooling at all. Active cooling would be required primarily for the detector arrays and their immediate housings. The reduced cooling requirements of such a system might be satisfactorily met by an acceptable, though still substantial, charge of cryogen, or by closed-cycle refrigerators required only to pump heat at low rates.
A stored cryogenic system always has a limited lifetime: unless replenished (an approach which has not been adopted for any astronomical mission), the cryogen eventually is fully depleted, the system warms up, and the mission comes to an end. To increase mission life spans, a variety of recyclable cryocoolers—both mechanical and electrochemical—have been under intense study (8). In principle, they could extend lifetimes indefinitely. In practice, a nagging long-term problem has been the limited reliability of closed-cycle, low-temperature refrigerators designed to operate in the vacuum of space. In the laboratory, such systems have often failed catastrophically after only a few months. No refrigerator of this type has ever operated in the laboratory continuously for a 10-year span. Yet this is the expected mission lifetime of many infrared astronomical space facilities now on the drawing boards. In 1998, NASA successfully tested a particular type of mechanical cooler on a Shuttle mission—a reverse Brayton-cycle cryocooler that can cool detectors to temperatures as low as 60-70 K. This test showed that operation under weightless conditions was not a problem for this type of cooler, but long-term reliability is still an open question, though the same coolers have a good record in the laboratory and run reliably for many months to a few years. Reliable closed-cycle cryocoolers are certain to affect critically the design and life spans of future infrared astronomical missions in space. The reverse Brayton-cycle cryocooler described before has been retrofitted to the NICMOS instrument on HST to extend its useful lifetime beyond the almost 2 years achieved with the initial charge of solid nitrogen.
Some types of highly sensitive infrared detectors now being planned for future missions operate effectively only at temperatures in the millikelvin range. Additional cooling beyond that achievable with liquid 4He must be provided for them. In the laboratory, a variety of techniques has already been developed to reach such low temperatures; often, they require a succession of stages that might employ combinations of thermoelectric, liquid 3He, 3He/4He dilution, ad-iabatic demagnetization, or other refrigerators. For long-duration astronomical space missions, reliable refrigerators will be required to provide these low temperatures continuously or cyclically. Again, these devices are often used in tandem; on the ESA/NASA Planck mission to study the CMBR, a hydrogen sorption refrigeration provides an 18-K heat station for a mechanical cooler which, in turn, provides a 4.5 K stage for a dilution refrigerator that cools the bolometer detectors to ~ 100 mK (9).
Light Collectors
With few exceptions, light collectors for the infrared are all-reflecting telescopes whose optical components may be aluminized, or gold-coated, depending on the wavelength range. Conventional telescopes image a portion of the sky onto a focal plane to provide accurate maps. Occasionally, however, the astronomer is interested in diffuse radiation that is not localized but arrives from all over the sky. For such observations, a carefully designed horn, an all-reflecting funnel, is generally employed to gather radiation from a large but well-defined field of view in the sky onto the smallest possible detector. These two types of light collectors were used, respectively, on the ISO/IRAS and COBE spacecraft.
For space applications, there is a premium on lightweight optics because the mass of the entire satellite scales with the mass of the optical system it must support, and, in turn, a more massive satellite requires a larger and more expensive launch vehicle. For an infrared mission, there is the added complication of the increased thermal conductivity of the beefier structure required to support the more massive optical system. Both IRAS and SIRTF, as discussed later, used all-beryllium optical systems because of the favorable strength-to-mass ratio of this material. The 85-cm diameter SIRTF primary, for example, has a mass of 15 kg and an areal density of 26 kg/m2. By comparison, the Hubble Space Telescope primary mirror has an areal density of 180 kg/m2. As telescope apertures beyond ~4m diameter are considered for future missions, another launch vehicle limitation, set by the physical size of the payload shroud, is encountered. Thus planning for the 8-m diameter Next Generation Space Telescope (NGST) is based on ultralight weight panels of glass, beryllium, or composite materials, whose areal density is no greater than 15 kg/m2. It would deploy after launch to achieve the desired aperture (6).
Filters
Filters isolate wavelength ranges of particular interest to the astronomer. For imaging and photometry, a well-defined, broad wavelength range needs to be isolated. Carefully designed transmission filters are usually used for this purpose. For spectroscopy, different types of spectrometers that select numerous narrow-wavelength intervals are inserted between the light collector and the detector or detector arrays. The most common types of spectrometers for infrared are prism or grating ”dispersive systems” that separate out radiation direction-ally, according to wavelength, and interferometers. Fabry-Perot interferometers select one narrow-wavelength range at a time; Michelson and other two-beam, multiplex interferometers transmit many wavelengths simultaneously but have to be swept through a range of settings to encode unambiguously and register the flux detected at each wavelength. Later, we describe spectrometers by their spectral resolving power R, defined as 1/81, where 1 is the operating wavelength and 81 is the finest discernible spectral detail. The unique requirements placed on filters and spectrometers for space applications are largely environmental and have to do with surviving launch or the ionizing radiation in space, or achieving low mass or volume, rather than with the device’s functionality or performance.
Early Rocket Instrumentation
In the mid-1960s, a collaborative effort between Cornell University and the Naval Research Laboratory led to the design of liquid-nitrogen-cooled and eventually liquid-helium-cooled, rocket-borne infrared telescopes. Early Cornell designs incorporated a parabolic primary mirror with an 18-cm aperture and focal ratio length ratio //0.9. The entire telescope, except for the entrance aperture, was surrounded by the liquid. Four different types of detectors were flown on many of these flights to sample the spectral range from 5 mm to 1.6 mm (10). Using this apparatus, the total flux in a field of view roughly 1° in diameter was first successfully measured for the galactic center and four other regions in the central portions of the Milky Way, at 5, 13, 20, and 100 mm. A first spectral measure of the radiation emitted by the solar system’s zodiacal dust was also obtained (11).
Some years later, results from a survey conducted in a series of rocket flights were published by the U.S. Air Force Cambridge Research Laboratories (now the Air Force Geophysical Laboratories). The group initially flew liquid-neon-cooled telescopes that had 10-cm apertures and detectors sensitive to radiation at 12-14 mm. Each of six detectors in a linear array surveyed a 10′ x 10′ field of view. Later, the group began all-sky surveys using satellite-borne instrumentation and also began observations across wider spectral ranges. Early results of one of these surveys at 4.2, 11.0, 19.8, and 27 mm were cataloged and published by Price and Walker (12).
Although rockets have not been extensively used for infrared astronomy in recent years, large-format infrared detector arrays may enable significant science in the limited duration of a rocket flight. For example, a 16.5-cm rocket-borne telescope instrumented with a 256 x 256 InSb array and cooled by supercritical helium has been flown to search for a faint halo of low-mass stars enveloping a nearby edge-on spiral galaxy (13).
The Infrared Astronomical Satellite
The first true infrared survey of the sky from space was carried out by the Infrared Astronomical Satellite (IRAS), jointly sponsored by the United States, the Netherlands, and Great Britain (3). Approximately two-thirds of the 300-day mission that lasted from January to November 1983 was devoted to an unbiased survey of the sky that succeeded in charting 98% of the celestial sphere in four broad wavelength bands. IRAS was launched into a polar orbit at the day-night terminator which precessed about 1° per day. In this ”Sun-synchronous” orbit the Earth/Sun/spacecraft geometry varied only slowly, so that the survey could be executed by a simple scanning strategy. Observations were carried out with an all-beryllium 57-cm aperture, //9.6 Richey-Chretien telescope whose focal plane was cooled to 3K and featured a total of ~60 Si:As, Si:Sb, and Ge:Ga discrete photoconductors; each had a separate JFET amplifier readout. The detectors covered, respectively, the 12-, 25-, 60-, and 100-mm bands, using Ge:Ga appropriately filtered to cover the last two. A low-resolution spectrometer covered the wavelength range from 7.5-23 mm.
A measure of the mission’s success was the cataloging of some 250,000 celestial sources; the vast majority had never before been detected in the infrared. No area of modern astrophysics was untouched by IRAS. A few of the many scientific highlights include
1. The discovery of galaxies that emit up to fifty times more energy at far-infrared wavelengths than in the optical domain and also emit from 100 to 1000 times as much total power as our own galaxy, the Milky Way. The existence of such highly luminous infrared galaxies came as a huge surprise.
2. The discovery of disks composed of fine dust grains that orbit around a number of stars that, in many ways, were reminiscent of our own Sun. This dust, it was conjectured, is the remnant of an originally far more massive circumstellar cloud of gas and dust from which a system of planets had already formed and initiated further astronomical searches for signs of planets around these stars. A sharply defined, though far fainter, set of dust rings was also found orbiting our own Sun, as were enduring trails of dust left by the passage of solar system comets.
3. The successful measurement of spectra for planetary nebulae and a variety of other sources at wavelengths previously inaccessible due to telluric absorption. IRAS also identified patchy infrared emission from the diffuse interstellar medium, referred to as ”infrared cirrus” because of its similarity to the thin, streaky clouds in Earth’s atmosphere. Infrared cirrus is important as a tracer of matter within our galaxy and as a potential source of interference in observations of distant galaxies.
The Cosmic Background Explorer, COBE
COBE, built by NASA and launched into a polar orbit identical to that of IRAS in 1989, was dedicated to the study of the microwave and infrared background radiation in space (14). It carried three instruments; two of them, the Diffuse Microwave Radiometer (DMR) and the Far Infrared Absolute Spectrophotometer (FIRAS) were used to study the CMBR—the isotropic blackbody radiation whose temperature is ~2.73 K and is believed to be a relic of the Big Bang in which the Universe was born. The third experiment, the Diffuse Infrared Background Experiment (DIRBE), measured the background at infrared wavelengths from 1200 mm. DIRBE and FIRAS were cooled by liquid helium and mapped the entire sky repeatedly during the ~ 10-month cryogenic lifetime of COBE. The critical components of DMR were cooled passively to ~ 140 K; this instrument operated for about 4 years.
The CMBR carries important cosmological information, and DMR and FIRAS were extremely successful, respectively, in measuring the spatial structure in that radiative field and in establishing its blackbody nature at a very high degree of precision. These important cosmological experiments will not be discussed further here. Of course, both FIRAS and DMR also measured the foreground radiation from our own galaxy. FIRAS was a polarizing Michelson interferometer instrumented with helium-cooled bolometers as detectors. It obtained spectra of the galactic emission from ~ 100 mmto ~ 3 mm using a 7° field of view. Of particular interest was its detection of emission from oxygen, carbon, nitrogen, and carbon monoxide from gas in the Galaxy (15).
DIRBE made measurements in ~ 15 wavelength bands, using a variety of discrete photodiodes and photoconductors from 1-100 mm and helium-cooled bolometers at 140 and 240 mm. All measurements were referenced to an internal cold, black reference surface so that the absolute sky brightness was determined.
The principal scientific results from DIRBE include (16)
1. An improved determination of the distribution of the infrared radiation from the zodiacal dust cloud within the solar system, which has led to improved models of the dust cloud and its infrared emission. DIRBE also confirmed IRAS’ discovery of a modest enhancement of emission in Earth-trailing direction, which is attributed to temporary gravitational trapping by Earth of zodiacal dust particles that are spiraling inward towards the Sun.
2. Measurements of the large-scale distribution of infrared radiation from the Galaxy, including both the far infrared radiation from 25-200 mm that samples the distribution of heated dust, and the near-infrared radiation from 1-25 mm that is indicative of the large-scale distribution of stars in the Galaxy.
3. Detection of an isotropic background of infrared radiation at 140 and 240 mm that arises from outside the Galaxy and may be attributable to the integrated effects of star-forming galaxies at redshifts z~ 1to2.
The Infrared Space Observatory (ISO)
The Infrared Space Observatory, built and launched by the European Space Agency (ESA), was the first comprehensive infrared astronomical space observatory. NASA and the Japanese Space Agency (ISAS), provided important technical, operational, and scientific support. ISO mapped celestial sources and analyzed them through spectroscopy, photometry, and linear polarization studies (17).
On the night of 16-17 November 1995, an Ariane 4 rocket launched ISO into a highly elliptical, 24-hour, circumterrestrial orbit, where the observatory operated with great success until its helium ran out and instruments began warming up in April 1998. The spacecraft in orbit was 5.3 m long, 2.3 m wide, and its mass was approximately 2500 kg. At launch, it carried a superfluid helium charge of 2300 liters, which maintained the Ritchey-Chretien telescope, the scientific instruments, and the optical baffles at temperatures of 2-8 K. The diameter of the telescope’s fused silica primary mirror was 60 cm. A three-axis-stabilization system provided an absolute pointing accuracy of a few seconds of arc and stability of a fraction of an arc second in both jitter and long-term drift. The telescope was diffraction-limited down to wavelengths of roughly 5 mm. Four instruments formed the core of the scientific payload:
1. A camera containing two 32 x 32 pixel arrays: In Sb for the wavelength range 2.5-5.5 mm, and Si:Ga for the range 4-18 mm. Each array could be operated with a selection of filters for broadband spectrophotometry or continuously variable filters (CVF) for low-resolution (R~40) imaging spectroscopy and could view sources through three linear polarizers oriented relative to each other at angles of 60°.
2. A photometer covered the entire wavelength range from 2.5-240 mm. It employed Si:Ga detectors that gave peak response at 15 mm, Si:B detectors that gave peak response at 25 mm, unstressed Ge:Ga detectors that gave peak response at 100 mm, and stressed Ge:Ga detectors that gave peak response at 180 mm. Stressed detectors are mounted in a miniature clamp or vise that applies high mechanical pressure to the crystal, thereby extending their wavelength response. At 100 and 200 mm, the instrument housed, respectively, 3 x 3 and 2 x 2 arrays of unstressed and stressed Ge:Ga to facilitate mapping. Multiple apertures, multiple filters, and polarizers were used for photometric and photopolarimetric measurements in each range. Scanning and mapping operations were carried out at all wavelengths. Two grating spectrophotometers, each with a 64-element linear Si:Ga detector array, provided spectra with resolving power R~ 100 at 2.5-5 and 6-12 mm.
3. A short-wavelength spectrometer included both grating and Fabry-Perot (FP) instruments. Grating spectra were available for the entire wavelength range from 2.38 to 45.2 mm, with resolving power R~ 1000-2000. The FP mode covered the 11.4- to 44.5-mm range and gave resolving power of a factor of 20 higher. For the grating mode, the detectors were InSb at 2.38-4.08 mm, Si:Ga at 4.08-29 mm, and Ge:Be at 29-45.2 mm. For the FP mode, Si:Sb was used out to 26 mm, and Ge:Be from 26-44.5 mm.
4. A long-wavelength spectrometer provided coverage from 43-196.9 mm. A grating provided resolving power R~ 150-200. A FP mode permitted observations at Rb 6800-9700. The 10 detectors, arranged in a linear array on a curved surface, were Ge:Be at 43-50 mm, unstressed Ge:Ga at 50-100 mm, and stressed Ge:Ga beyond 110 mm.
Among the scientific highlights of ISO were
1. The detection of water vapor throughout the interstellar medium of the Galaxy. Before ISO, the infrared emission from interstellar water vapor could not be detected because telluric water vapor absorbs at precisely the emission wavelengths. Water vapor, however, can be one of the primary coolants of interstellar clouds, and the extent of this cooling needed to be understood to assess the extent to which it facilitates protostellar collapse.
2. Detection of polycyclic aromatic hydrocarbons in the spectra of galaxies. These large molecules were well known in our galaxy, but ISO had the sensitivity needed to show that their emission dominates the 5- to 12-mm emission from nearby spiral galaxies as well. The emission from these molecules is due to radiative fluorescence: a molecule is excited by optical or ultraviolet radiation, quickly rereadiates the energy of a single absorbed photon, in the infrared and returns to the ground state.
3. Inventories of extragalactic source-counts at wavelengths ranging from 4175 mm. These are of particular value in understanding the origins of the extragalactic diffuse infrared radiation detected by the COBE mission. Many of the randomly observed galaxies appear to be ultraluminous, indicating that they contain substantial regions of massive star formation or that they harbor an active galactic nucleus that possibly surrounds a massive central black hole.
The Space Infrared Telescope Facility (SIRTF)
NASA is developing SIRTF for launch in 2003; it has a projected cryogenic lifetime greater than 5 years (1). SIRTF will be an observatory for infrared astronomy from space and will complete NASA’s family of Great Observatories (the other members are the Hubble Space Telescope, the Compton Gamma Ray Observatory, and the Chandra X-ray Observatory).
SIRTF culminates the four decades of technology development and scientific progress described above. SIRTF will be the first space mission to use exclusively the imaging and spectroscopic power of large format infrared detector arrays. SIRTF’s all-beryllium telescope that incorporates an 85-cm diameter primary mirror and is diffraction-limited down to 6.5 mm, defines the state of the art for ultralight weight cryogenic optics. The SIRTF telescope and cryogenic system will be carried on a fairly standard spacecraft bus that provides pointing control, power, data storage, and communication. The pointing system is built around an external autonomous star tracker that controls and reports the spacecraft orientation at more than 2″ accuracy and has a reaction wheel/gyro control system. Visible light sensors in the cold focal plane can sense stars simultaneously using the external star tracker to track the relative orientation of the telescope and star tracker lines of sight. This pointing system architecture was used on ISO as well. The SIRTF spacecraft also incorporates a nitrogen gas system that is used to unload the reaction wheels if they accumulate too much angular momentum; the magnetic torquer bars used for this function in Earth-orbiting spacecraft would not work on SIRTF because it is far outside Earth’s magnetosphere.
Unlike the missions previously described, all of which operated in Earth orbit, SIRTF will be placed into an Earth-trailing heliocentric orbit, drifting slowly away to reach a distance of ~0.5 AU from Earth after 5 years. In this orbit, SIRTF is free of the heat load from Earth and provides good access to the sky for target selection and scheduling. SIRTF will launch with the telescope warm and the instruments at helium temperature. In space, the telescope and its surrounding thermal shields cool radiatively to ~ 40 K, and the effluent helium from the cryogenic tank cools the telescope down to its operating temperature of ~5.5K (see Fig. 1b). The thermal shields remain at 40 K and below throughout the mission, so that the parasitic heat conducted into the telescope is very small. As a result, the heat load that dissipates the SIRTF cryogen and determines the lifetime of the mission comes largely from the focal plane instruments.
This hybrid, radiative, cryogenic cooling system is facilitated because SIR-TF can maintain an attitude in its solar orbit in which the solar panel is always oriented toward the Sun and shades the thermal shields that control the telescope temperature. These structures, in turn, are optimized to minimize the heat transferred from the solar panel and to radiate to space any heat that is transferred. This approach would not work in near-Earth orbit because the heat load from Earth would occasionally be incident on the thermal shields and turn the radiator into an absorber.
This optimized cryogenic system, together with the low-power dissipation of its instruments and the elimination of parasitic heat loads, makes SIRTF a much more efficient system cryogenically than any of its predecessors. SIRTF carries 350 liters of helium at launch, and a lifetime of 5 + years is predicted, based on an average instrument power dissipation of ~5mW. By comparison, ISO was launched with ~2300 liters of helium and achieved a lifetime of ~2.5 years with an average instrument power dissipation of ~ 10 mW.
SIRTF will carry three array-based focal plane instruments:
1. A near-infrared camera that provides imaging simultaneously in four bands at 3.6, 4.5, 5.8, and 8 mm. Both of the 3.6- and 5.8-mm channels image the same field of view in the sky; this is made possible by a dichroic filter that transmits 5.8 mm and reflects 3.6 mm. An adjacent field of view is imaged at 4.6 and 8 mm in a similar fashion. Each band uses a 256 x 256 pixel array hybridized to a 256 x 256 MOSFET multiplexer. The detector material is InSb in the 3.6- and 4.5-mm bands and Si:As in the 5.8- and 8-mm bands.
2. A spectrometer that provides low resolving power (R ~ 60-120) spectroscopy from 5-40 mm and higher resolution spectroscopy (R~600) from 10-38 mm. The spectrometer consists offour physically distinct modules; each contains a 128 x 128 array (Si:Ga for the shorter wavelengths, Si:Sb for the longer wavelengths) illuminated by an optical train of mirrors and gratings. The use of detector arrays allows these modules to be very compact and efficient and obviates the need for moving parts. The higher resolution modules use two diffraction gratings so that, an entire octave of the spectrum can be cross-dispersed across the entire array and measured simultaneously. In the lower resolution modules, a long entrance slit is used to permit obtaining spectra simultaneously at many spatial points. A portion of the array in one of these modules is also used for the precision target acquisition required to place a source on a narrow spectrograph slit, thereby alleviating the absolute pointing requirements placed on the spacecraft.
3. An imager/photometer that provides imaging and low-resolution spectrophotometry at wavelengths between 25 and 160 mm. This instrument uses a 128 x 128 Si:Ga array at 25 mm but incorporates two Ge:Ga arrays for longer wavelength measurements. The 32 x 32 Ge:Ga array used by SIRTF at 70 mm is composed of eight 4 x 32 submodules; each in turn consists of four 1 x 32 linear arrays, coupled to a 1 x 32 amplifier/multiplexer. The 2 x 20 array used at 160 mm is similarly built up of four 2 x 5 pixel modules with the added complication that the module construction allows for the mechanical stress needed to extend the long-wavelength response of Ge:Ga from 120 to beyond 160 mm. These arrays represent substantial advances in the state of the art and point the way toward still larger arrays for future applications. Note that although these arrays have fewer pixels than those described before, they are to be used at longer wavelengths where the diffraction-limited image size is larger compared to the field of view. Thus they provide comparable sampling of the telescope’s focal plane; in fact, all three arrays are designed to sample the image fully to allow numerical post processing of the data to enhance the spatial resolution.
The scientific return of SIRTF cannot be forecast because its capabilities represent such a great advance beyond what has been possible in the past and also because the bulk of the observing time on SIRTF will be dedicated to programs to be proposed and carried out by the general scientific community. However, based on the science return from the other missions described before, we anticipate that SIRTF will lead to great advances in our understanding of such problems as
1. the formation and early evolution of galaxies, stars, and planets;
2. the physical processes that power the objects of highest luminosity in the Universe;
3. the chemical composition of interstellar and circumstellar matter;
4. the nature of the coolest, lowest luminosity stars and star-like objects in the solar neighborhood; and
5. the properties and interrelationships of comets, asteroids, interplanetary dust, and other small bodies in the solar system.
In addition, SIRTF’s large arrays, very high sensitivity, and long lifetime give this mission great potential for discovering new phenomena.
Other Infrared Missions Already Flown
Other significant infrared space astronomy missions are described briefly here:
1. The Spacelab II Infrared Telescope—1985. A 15-cm diameter helium-cooled telescope was flown on Spacelab-2 and made infrared measurements between 2 and 120 mm. It provided data about the structure of the Galaxy (18) and about the infrared background environment on the Space Shuttle.
2. The Midcourse Space Experiment (MSX)—1995. MSX carried a 35-cm aperture off-axis telescope and five linear Si:As arrays that mapped the sky in a push-broom fashion in bands from 4-22 mm. Although it was primarily designed to scan Earth’s limb, it carried out a number of astrophysical experiments and produced excellent images of the entire galactic plane at ~ 18 arcsec resolution (19).
3. The Infrared Telescope In Space (IRTS)—1995. IRTS was an ISAS program with significant NASA participation. IRTS had a 15-cm diameter liquid-helium-cooled telescope and four varied focal plane instruments covering wavelengths from 3-800 mm. It was carried on a Japanese satellite called the Space Flyer Unit and surveyed ~6% of the sky in a 5-week lifetime (20).
4. The Near Infrared Camera and Multi-Object Spectrograph (NICMOS)— 1996. NICMOS is a replacement focal plane instrument, which was installed on HST. It was instrumented with three 256 x 256 HgCdTe arrays that carried a range of filters that covered the 1- to 2.5-mm spectral band and was optimized for high spatial resolution imaging. NICMOS was cooled by solid nitrogen and achieved a lifetime of slightly less than 2 years. This was somewhat less than expected because of a partial failure of the cryogenic system in orbit (21).
5. The Submillimeter Wave Astronomy Satellite (SWAS)—1998. SWAS was the first space mission to carry radio-type (heterodyne) receivers for spectroscopic exploration. SWAS has a 55 x 71cm near-optical quality off-axis primary mirror and two heterodyne radiometers with Schottky barrier diode mixers and a single acousto-optical spectrometer. SWAS is surveying the galactic plane in the emission of atomic carbon, molecular oxygen, water vapor, and carbon monoxide in five transitions between 538 and 615 mm (22).
Future Missions
A number of missions are planned or proposed for the next two decades to go beyond even the great scientific and technical accomplishments described before. These include
1. The Infrared Imaging Surveyor (IRIS) is a Japanese mission that will employ a 70-cm telescope cooled to 6 K, using a hybrid cryogenic system that incorporates both liquid helium and mechanical coolers to achieve a lifetime in excess of 1 year. Its primary mission is to conduct an all-sky survey at wavelengths of 50-200 mm at an angular resolution of 30-50 arc seconds. The detectors to be used for this purpose are stressed and unstressed Ge:Ga. IRIS is likely to be launched in the first half of the decade 20002010 (23).
2. The Far Infrared Space Telescope (FIRST) is a mission sponsored by the European Space Agency with substantial participation by NASA. Its primary mission is to provide detailed spectroscopy and imaging for the 80- to 670-mm spectral range. It features a 3.5-m passively cooled, primary mirror that illuminates three different liquid-helium-cooled instruments. FIRST is to be launched by an Ariane 5 rocket into a Lagrangian point L2 orbit in the second half of the decade 2000-2010 (24). FIRST will share its launch vehicle with the Planck mission to study the CMBR mentioned earlier. FIRST has been renamed Herschel by ESA.
3. The Next Generation Space Telescope (NGST) is an international mission led by NASA, that will use a radiatively cooled, ultralightweight 8-m diameter telescope that will deploy following launch (6). A suite of focal plane instruments using very large detector arrays will observe at wavelengths~ 0.6 mm to beyond 10 mm. NGST will greatly extend the scientific results of both HST and SIRTF in this wavelength region. NGST will be launched by an Atlas rocket into a Lagrangian point L2 orbit toward the end of the decade 2000-2010 and will have a scientific lifetime greater than 5 years.
4. Interferometers in Space (25). All of the missions described before were built around a single telescope. To achieve much higher resolution than the ~ 0.1-1 arcsec achievable with an ~8-m telescope, it will be necessary to use the techniques of interferometry, in which infrared radiation collected by two widely spaced telescopes can be brought together to achieve angular resolution comparable to that which would be provided by a single telescope whose aperture is equal to the separation of the two telescopes. Ultimately, this technique will be employed in NASA’s Terrestrial Planet Finder, scheduled to launch in the 2010-2015 time frame to image Earth-like planets around nearby stars.
