Report - School of Physics

Table 8: Detection capabilities: Earth at 10 pc, from Angel (2003). ∆θ = 0.1 arcsec, t int = 24 hr,
QE = 0.2, ∆λ/λ = 0.2. Mode = N corresponds to a nulling system, C to a coronograph. The
ground-based results assume that long-term averaging is realistic, with fast atmospheric correction.
For further details of assumptions, see Angel (2003)
Telescope Size (m) λ(µm) Mode S/N Comment
Darwin/TPF-I 4 × 2 11 N 8 See also Table 9
TPF-C 3.5 0.5 C 11 Typical launcher diameter
” 7 0.8 C 5–34
Antarctic 21 11 N 0.5
” 0.8 C 6
CELT, GMT 30 11 N 0.3 30 m [C] too small at 11 µm
” 0.8 C 4
OWL 100 11 C 4 Large Φ [C] for IR suppression
” 0.8 C 46 Optical spectroscopy possible
Antarctic OWL 100 11 C 17 Comparable to Darwin/TPF, but
O 3 (9.6 µm) and CO 2 (15 µm)
not accessible (atmosphere opaque)
” 0.8 C 90 Water bands at 1.1–1.4 µm
µm; 10 times the level of zodiacal dust than in the inner solar system; all of the
habitable zone is searched; Spitzer Si-As detectors. Times given are based on a 90%
confidence level for a non-detection based on 3 observations (a positive detection at
S/N = 5 then takes one third of the times given). To detect the Earth at 265 K
(instead of 290 K) around the Sun at 20 pc would take 9 hr at S/N = 5, and 36 hr
at S/N = 10. For a K5V star at 10 pc these times are about 1–4 hr respectively. For
spectroscopy, the S/N varies between 7 and significantly higher, depending on the
atmosphere. For an Earth atmosphere (but at 265 K) at 20 pc, an integration time
of 54 hr provides a S/N∼ 7 − 15.
One of the strongest noise sources is the leakage of photons from the resolved stellar
surface. This is a function of both stellar temperature and diameter (distance,
spectral type). It is most clearly seen for nearby stars where the detection time
drops strongly. For 20 pc detection times do not change much since the star does
not leak very much for any diameter. Detection times rise rapidly for K5V stars
at 30 pc, because of the 90% confidence requirement, which itself demands more
changes in array size because the habitable zone is very close to the star.
The number of candidates accessible and visible to Darwin (two ±45 ◦ caps near the
ecliptic poles are inaccessible) is estimated as follows. Considering only single stars,
there are 211 K and 82 G candidates out to 25 pc, with another 30 F-type stars
plus many M-dwarfs. Combined with the integration time estimates, Darwin should
survey more than 150 stars in 2 years. Within 5 years, all 293 single and accessible
solar-type stars stars out to 25 pc can be surveyed, with a spectrum obtained for
each system in the case of a planet prevalence of ∼10%. Up to 1000 systems in the
solar neighbourhood could be surveyed if the planetary fraction is even smaller.
50