5.1 INSTRUMENT PERFORMANCE

5.1.1 SPATIAL RESOLUTION

5.1.2 AREAL COVERAGE

5.1.3 SPECTRAL REGION AND RESOLUTION

5.1.4 SENSITIVITY

5.2 INSTRUMENT CONCEPT

5.2.1 ILLUSTRATIVE OBSERVING PROGRAM

5.3 ILLUSTRATIVE EXNPS DESIGN AND ALTERNATIVES

5.4 KEY INTERFEROMETRIC REQUIREMENTS

The idea of using an interferometer to null out the stellar signal in the search for a weak planetary signal originated with Bracewell (1978), who demonstrated that an interferometer used with a -phase delay between the two arms of the interferometer would cause the Airy pattern for a source seen on-axis to vanish, whereas the signal for a planet, seen off-axis, would be transmitted. Conceptually, one can imagine an interferometer of baseline B imposing a fringe pattern of spacing /B onto the Airy pattern of width /D of a component telescope with diameter D. By rotating the interferometer around the line of sight to the star, one rotates the fringe pattern on the sky; the star remains nulled while a planet moves in and out of the fringe pattern (Figure 5-2). Both Angel (1990) and Shao (1990) suggested the use of a Bracewell interferometer for the detection of Earth-like planets.

There are two problems with the classical Bracewell interferometer. First, the null is neither deep nor broad enough for this application. The finite size of the stellar disk results in incomplete suppression of star light (~10^-3). Second, as the interferometer rotates, the exo-zodiacal emission in the beam has the same 2 modulation as the planet. Angel (1990) suggested a modified Bracewell interferometer with four telescopes. Leger et al. (1996) suggested that much smaller mirrors could be used in an interferometer operated far from the Sun because of reduced zodiacal emission. Finally, Angel and Woolf (1996) developed a scheme for the OASES mission concept (Figure 5-3) which gives a deep, broad null that results in suppression of starlight by a factor of 10⁶ and a 2 modulation of the exo-zodiacal cloud, which can be distinguished from the 4 signals due to a planet.

Extracting information from the interferometer data is conceptually straightforward (Figure 5-4). Data are recorded in 10 to 20 wavelength bands as the interferometer rotates every 1 to 2 hours around the line of sight to the star. Planets at different radii produce unique signatures as they move in and out of the rotating fringe pattern. Since the fringe pattern is different at each wavelength, the instrument provides good uv-plane coverage for broad-band sources. This data set can be inverted to give a sequence of images showing planet motion, i.e., an orrery, of the planetary system. Figure 5-4 shows the simulated raw data; Figure 5-5 shows the derived image of an illustrative planetary system. The data represent a 10-hr integration with an interferometer consisting of four 1.5-m telescopes on a 75-m baseline. The instrument efficiency has been assumed to be 20%. The image doubling resulting from the symmetry of the instrument has been removed by offsetting the star relative to the optical axis. The inner radius of the image corresponds to 0.06 arcsec, implying that the habitable zone around a G2 star would be resolvable out to 15 pc and that of a K2 star out to 7.5 pc. The spectroscopic performance of the system is such that the broad CO₂ line could be detected to the 5- level in a week of observing, while the weaker O₃ and H₂O lines would require 6 weeks of integration (Figure 5-6).

5.2.1 ILLUSTRATIVE OBSERVING PROGRAM

After 1 year of observation, one would be able to perform a broadband survey of approximately 50% of the sky and study 100 to 200 objects in the search for planetary candidates. In the subsequent 3 years, one would make follow-up spectroscopic observations of promising candidates. For example, one could look for CO₂ toward 50 stars for one week each, and for H₂O and O₃ toward 10 to 15 stars for 6 weeks each.

The baseline space IR instrument consists of four 1.5-m telescopes operating in a linear array with a 75-m baseline (Figure 5-7). From its vantage point at around 5 AU to take advantage of the dark sky, this system would be able to detect and characterize Earth-like systems out to 13 pc from the Sun. However, this configuration is not unique. The Darwin concept (Leger et al. 1995) proposed to the European Space Agency (ESA) for a similar planet-finding purpose uses a circular array of five 1-m telescopes on a 50-m-diameter circle. This arrangement produces a sharp distinction between the exo-zodiacal cloud and the planet signals. A six-element system proposed by JPL is arranged as three two-element arrays and allows for good suppression of star light and good rejection of smooth exo-zodiacal emission. Each configuration has its own strengths, which must be studied carefully over the next few years.

The nulling performance of one part in 10⁶ will be difficult, but not impossible to achieve. First, the interfering wavefronts must match in phase to /6000, or roughly 1.5 nm, and the amplitudes in the two arms of the interferometer must match to one part in 10³. In addition, the null must be achromatic. These requirements can be achieved by using a fully reflecting phase shifter and single-mode spatial filters (optical fiber or waveguide) to clean up the wavefront to the required accuracy. Other requirements, such as precision pointing and path-length control, are possible because of the copious number of short-wavelength photons from the parent star. The <5 µm radiation will be used to close the telescope pointing and interferometer servo systems with kHz bandwidths.

Back to Table of Contents
Next Chapter

• Back to Road Map Title Page
• List of Tables
• List of Figures