Priorities in Space Science Enabled by Nuclear Power and Propulsion (2006)

Radio (λ >1 mm) interferometry on interplanetary baselines appears, at first glance, to be an ideal application of the ability of nuclear power and propulsion systems to deploy astronomical assets at great distances from Earth. However, the scattering of radio waves by the interstellar medium blurs radio images to angular sizes much greater than the resolution of a radio interferometer with a baseline of >1 AU.

The “warm, ionized” phase of the interstellar medium contains electron density fluctuations on a range of size scales, from smaller than ~10⁸ m to greater than ~10¹³ m. Evidence implies that the spectrum of density fluctuations is close to the Kolmogorov spectrum, familiar from characterizations of neutral turbulence. These fluctuations scatter radio waves from cosmic sources, causing frequency-dependent phase deviations that ultimately result in interference in the observer plane. This results in a variety of observed phenomena, including amplitude variations in time and frequency, akin to the twinkling of stars due to density inhomogeneities in Earth’s atmosphere. Multipath scattering makes point sources of radio emission appear to have finite angular extent, of full width half maximum (FWHM) θ_s, a result of averaging over short-time-scale image wander. The size of the scattering “disk” varies inversely with υ² and depends strongly on the line of sight.

The greatest scattering is seen toward the galactic center: Sag A* has an apparent angular extent of 1.3 arcseconds at 1 GHz. Lines of sight perpendicular to the galactic plane show the least scattering, typically ~5 milli-arcseconds at 1 GHz. Out of the plane, given the frequency dependence of θ_s, it is not hard to show that regions for which υ < 50 to 150 B (where υ is in hertz and B is the interferometer baseline in meters) have θ_s > λ/B, the effective interferometer resolution. This region is shown in Figure 8.1; radio interferometry with B ≥ 1 AU is clearly uninteresting.

The impact of natural and human radio emissions on radio astronomy is enormous. The entire radio spectrum from ~30 MHz is strongly affected by interfering manmade signals. The region above 300 MHz—even those bands that have long been allocated specifically for radio astronomers—continue to experience tremendous pressure for commercial use. Moreover, Earth’s ionosphere absorbs and refracts radiation below ~30 MHz. Additionally, natural sources of interference on Earth—such as auroral kilometric radiation, which produces very intense

radiation in the frequency range from 50 to 750 kHz, or lightning, which produces strong interference in the range from 1 to 30 MHz and above—preclude observations below 30 MHz (the very low frequency [VLF] range) except under exceptional circumstances, or at special locations and for limited amounts of time.

A variety of astronomical phenomena are expected to emit radiation at the wavelengths affected by terrestrial radio noise. These include non-thermal emission from the Milky Way galaxy, pulsars, interstellar scintillation, active galactic nuclei, and clusters of galaxies, as well as the Sun and Jupiter. Much higher up in frequency, neutral atmospheric gases—particularly atmospheric water vapor—attenuate cosmic radiation increasingly strongly above 10 GHz, with attenuation peaking around 22 GHz. Strong oxygen lines attenuate heavily near 60 and 120 GHz, and water lines around 183 GHz.

For observations at λ ≈ 0.2 to 3 μm, sunlight scattered by zodiacal dust grains is the dominant source of diffuse background emissions and can, hence, be the dominant noise source for observations of faint sources. Observations from Pioneer 10¹ and Helios 1 and 2² spacecraft suggest that zodiacal brightness declines with heliocentric distance as I_z α r^–2.3 or I_z α r^–2.5. An observatory at 5 AU could have ≈ 50× lower zodiacal background than current or planned ultraviolet/optical/infrared observatories in Earth-trailing or L2 orbits. Reducing the zodiacal background further is of limited use, as diffuse galactic emission and the mean extragalactic flux are ≈ 10⁻² of the 1-AU zodiacal background near 800 nm.

For background-limited observations of unresolved sources of specific flux f_υ, the signal-to-noise (S/N) ratio acquired in time T from a diffraction-limited telescope of diameter D scales as:

The last factor is the bandwidth of the observation. The lowered zodiacal background at 5 AU could increase observing efficiency by a factor of 50. This gain is realized only when the diffuse background is the dominant noise source. For brighter sources, shot noise in the source photons is dominant. For 2-meter-class visible telescopes at 1 AU, such as the Hubble Space Telescope (HST) or the proposed Supernova/Acceleration Probe, any source brighter than V ≈ 29 mag is brighter than the diffuse background—nearly every star within 10 kpc, for example. The zodiacal brightness in the near-infrared is similar to that in the visible and drops precipitously into the ultraviolet, so it is unlikely that observing beyond 1 AU would be of use in observations of stars in the Milky Way.

Study of stars beyond the Milky Way, for example in elliptical galaxies, requires reaching V > 29 mag. But such observations also require very high angular resolution, much better than that afforded by the HST, to eliminate crowding of stars and resolve the population. Hence an increase in D to improve resolution (and S/N) would be much more useful than a reduction in I_z. There is hence little S/N incentive to move beyond L2 for the observation of point sources at ultraviolet, optical, and infrared wavelengths.

A major thrust of the astronomy and astrophysics (AAp) decadal survey³ and NASA’s exploration initiative is the detection and study of extrasolar planets. For such observations, there is little S/N incentive to reducing the solar zodiacal background by going to >1-AU orbits, because most of the targets will be embedded in a dust disk about their host stars that is a significantly larger and unavoidable source of background photons. Thus, the first reconnaissance and characterization of extrasolar planets will be done from a near-Earth vantage point.

Once the telescope is large enough to resolve the astronomical source, the S/N for objects with surface brightness fainter than the zodiacal background becomes

A glance at the Hubble Ultra Deep Field shows that most of the faint, high-redshift galaxies in the universe are resolved by 2-m telescopes and are fainter than the zodiacal background in the visible. For observations of these very interesting sources, operation at 5 AU can perhaps be equivalent to a 50-fold increase in telescope area (or 7-fold increase in diameter). For a nuclear propulsion system to be useful, its cost and weight would have to be such that placing a 2-m telescope at 5 AU would be much cheaper than placing a 15-m telescope (or 50 2-m telescopes) at L2. Otherwise, one would choose the L2 observatories, which would offer superior S/N for other observations, as well as resolution.

Furthermore, a large telescope is only worth deploying at 5 AU if its purpose is limited to obtaining ultraviolet/optical/infrared spectra of high-redshift galaxies. A 0.1-arcsecond-diameter galaxy with surface brightness 10 times lower than the 1-AU zodiacal light will deliver only ≈10⁻³ photons per second to a 6-m aperture in a spectral element with R = Δυ/υ ≈ 1,000. Hence a measurement with S/N = 20 would take 400,000 s—even assuming no deleterious effects of detector noise or radiation events. Imaging observations (R ≈ 10) could more readily profit from the lower zodiacal background at 5 AU.

In terms purely of S/N, therefore, the value of nuclear power systems to ultraviolet/optical/infrared astronomy depends on the cost of the 5-AU location versus the cost of larger apertures at 1 AU, and in any case the S/N gains are likely to be limited to imaging of faint resolved galaxies. Certainly, the priorities of the ultraviolet/optical/infrared community will be served first by building a larger collecting area near 1 AU.

A space telescope located 1 AU from the Sun will have an ambient temperature of approximately 300 K. Observations at wavelengths longer than 1 μm will be strongly background limited by the telescope’s own thermal emissions. Cooling the telescope’s optical system is clearly highly advantageous. Small, 1-meter-class telescopes can achieve operating temperatures of order of between 4 and 8 K by the use of onboard expendable cryogens. Applying such a cooling strategy to a large astronomical telescope is more problematic. The 6.5-m James Webb Space Telescope (JWST) will use a multilayer sunshield to passively achieve an operating temperature of 40 K at the Sun-Earth L2 point. This temperature is low enough to allow background-limited performance at λ < 20 μm, but for observations at longer wavelengths, still lower temperatures will be required. Current models for the proposed Single Aperture Far-Infrared Observatory (SAFIR) mission—envisioned as a colder, somewhat larger (~10-m-class), far-infrared version of JWST—indicate total residual heat loads of ~1 W for a telescope at 5 to 10 K. Such a heat load could, in principle, be addressed with expendable cryogens, although this approach would require some 30 liters of liquid helium per day. Such a consumption rate would lead to unreasonable masses of cryogen, and active cooling with onboard refrigerators is considered to be the best way to ensure long observatory lifetime. As described below, SAFIR is taken to be representative of the requirements for a class of large thermal-infrared space telescopes.

Space-qualified cryocoolers have been developed for both infrared and x-ray applications. Such low-temperature refrigeration systems do not rely on consumables and can be understood to provide a failure-limited observatory lifetime. These systems do, however, have substantial power requirements, and nuclear power sources can be considered as potentially enabling for such missions. For the low temperatures required by SAFIR, cryocooler efficiency is low, and a ratio of compressor input power to cooling power on the order of 1,500 is expected. Using the residual heat load referred to above, this would require on the order of 1.5 kW of

electric power to maintain SAFIR at optimal temperature, which would mean that these cryocoolers would dominate the power budget of the observatory.

To supply just the cryocooler power budget for SAFIR at 1 AU, at least five Cassini-class RPS units (see Table 1.2) would be required. The power needs can also be met, however, by ~4 m² of solar panels, which are substantially smaller than the ~100 m² sunshield required for the observatory and are much less expensive than the RPSs. This sunshield would maintain an orientation perpendicular to the Sun, which is the orientation that is also most efficient for the solar panels. Thus, both the solar panels and the sunshield could be conveniently engineered together as part of an observatory structure. The advantages of solar panels relative to RPSs for SAFIR are multiplied by the more favorable power-to-mass ratio of solar panels. Although RPSs would not be as deployment-dependent as solar panels, they would add to the weight of the observatory about twice the weight that the solar panels would add.

Because of the high minimum power (~0.5 to 1 MWt) required for reactors with competitive power-to-mass ratios, fission systems appear a poor choice to power an observatory that would otherwise require only a few kWe. Dumping most of the thermal power from an onboard fission generator into free space would be difficult to do without adding large heat loads to the observatory itself. In addition, the strong gamma-ray flux from the fission system would seriously affect the performance of the observatory sensors.

Under what circumstances might nuclear power systems offer value to cryogenic infrared observatories? Although the power needs for such telescopes seem to be met economically by solar arrays at 1 AU, telescopes at larger heliocentric distances might benefit. Large heliocentric distances offer lower zodiacal background and increased operational efficiency. While both cooling power requirements (which are determined largely by solar insolation) and photovoltaic-power generation together decrease with distance, the benefit equation changes when the distance is large enough that active cooling no longer dominates the observatory power budget. In addition, communication power requirements rise even faster with distance for a given bandwidth. For such cases (at, for example, 3 to 5 AU) RPS power might be highly enabling. For the reasons noted above, however, onboard fission systems would remain problematic.

1. M.S. Hanner, J.L. Weinberg, D.E. Beeson, and J.G. Sparrow, “Pioneer 10 Observations of Zodiacal Light Brightness Near the Ecliptic—Changes with Heliocentric Distance,” pp. 29–35 in Interplanetary Dust and Zodiacal Light: Proceedings of the IAU Colloquium No. 31, Springer-Verlag, Berlin and New York, 1975.

2. C. Leinert, I. Richter, E. Pitz, and B. Planck, “The Zodiacal Light from 1.0 to 0.3 A.U. as Observed by the Helios Space Probes,” Astronomy and Astrophysics 103: 177–188, 1981.

3. Board on Physics and Astronomy–Space Studies Board, National Research Council, Astronomy and Astrophysics in the New Millennium, National Academy Press, Washington, D.C., 2001.