A&A 507, 1649–1658 (2009) Astronomy DOI: 10.1051/0004-6361/200912581 & c ESO 2009 Astrophysics Tohoku-Hiroshima-Nagoya planetary spectra library: a method for characterizing planets in the visible to near infrared R. Lundock1,T.Ichikawa1, H. Okita1, K. Kurita1,K.S.Kawabata2,M.Uemura2, T. Yamashita2, T. Ohsugi2, S. Sato3, and M. Kino3 1 Astronomical Institute, Tohoku University, Aoba, Sendai 980-8578, Japan e-mail: [email protected] 2 Astrophysical Science Center, Hiroshima University, Kagamiyama 1-3-1, Higashi-Hiroshima 739-8526, Japan 3 Department of Physics, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan Received 27 May 2009 / Accepted 28 August 2009 ABSTRACT Context. A comprehensive framework for comparing spectral data from different planets has yet to be established. This framework is needed for the study of extrasolar planets and objects within the solar system. Aims. We completed observations to compile a library of planet spectra for all planets, some moons, and some dwarf planets in the solar system to study their general spectroscopic and photometric natures. Methods. During May and November of 2008, we acquired spectra for the planets using TRISPEC, which is capable of simultaneous three-band spectroscopy across a wide wavelength range of 0.45−2.5 μm with low resolving power (λ/Δλ ∼ 140−360). Results. Patterns emerge when comparing the spectra. By analyzing their general spectroscopic and photometric natures, we show that it is possible to distinguish between gas planets, soil planets, and ice planets. These methods can be applied to extrasolar observations acquired using low resolution spectrography or broad-band filters. Conclusions. The present planet spectral library is the first library to contain observational spectra for all of the solar system planets, based on simultaneous observations at visible and near infrared wavelengths. This library will be a useful reference for analyzing extrasolar planet spectra and calibrating planetary data sets. Key words. astronomical data bases: miscellaneous – planets and satellites: general 1. Introduction telescope (Fink et al. 1976; Fink & Larson 1979). The Pollack group using the Kuiper Airborne Observatory used an uncooled, Successive improvements to telescopes and probe missions have circular variable-filter wheel for their 1975 observations of provided an unprecedented amount of information about the Venus and Jupiter (Pollack et al. 1978a). They upgraded to a planets orbiting the Sun and other stars. The numbers of known cooled filter wheel for their 1976 observations of the Galilean planets and bodies that are “not quite planets” increase every satellites (Pollack et al. 1978b). Few observations have been per- year. Ceres, Pluto and Eris were reclassified as dwarf planets in formed for spectral libraries in the past two decades. Examples 2006. Haumea and Makemake were subsequently added to the of spectral data at a broad range of wavelengths are Earth’s spec- dwarf planet category (WGPSN 2009). It is an open question μ ff tra from 0.7 to 2.4 m(Turnbull et al. 2006) and the IRTF spec- whether small brown dwarves can be di erentiated from large tral library (Rayner et al. 2008), which includes four planets gas giants. Because of the enormous amount of data taken so (Jupiter, Saturn, Uranus, and Neptune) observed by SpeX in the far by many authors, we know a great deal about the nature of near-infrared (0.8−5.5 μm) each planet in the solar system. However, the framework for dis- Most data for the spectral libraries were taken using a unique cussing “planets in general” has lagged behind. combination of telescope and instrument. The resolution, re- Spectra exist for all of the solar system’s planets and many sponses, and wavelength coverage differed for each data set. of its smaller objects. To find spectral libraries in many cases it It would be difficult to combine published spectral data at vari- is necessary to return to the 1970’s and early 1980’s when the ous wavelengths, because the data were taken for different phase first quality spectra of planets were being obtained. Particularly and/or rotation of the planets. Few data simultaneously obtained prolific studies during this era were, e.g., McCord & Westphal across a wide wavelength range from optical to infrared, are (1971), Fink et al. (1976), McCord et al. (1979), Fink & Larson available. (1979), Clark (1980, 1981), and McCord (1981). However, this was also a time of rapid technological development and In the present study, we show the results of simultaneous even within a single research group, observational techniques spectroscopic observations at wavelengths from optical to in- changed. Between their 1976 and 1979 observations, the Fink frared for all planets, some moons, and some dwarf planets in group at Kitt Peak upgraded from a 90 inch telescope to a 4 m the solar system to study their general spectroscopic and photo- metric natures. Accounts of the observations and data reduction THN Planet spectra library is only available in electronic form at are presented in Sects. 2 and 3, respectively. The spectra of all the CDS via anonymous ftp to targets that we observed are compared in Sect. 4 along with the cdsarc.u-strasbg.fr (130.79.128.5) or via broad-band colors obtained by integrating the spectra. The col- http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/507/1649 ors of other objects in the solar system, which are available in Article published by EDP Sciences 1650 R. Lundock et al.: THN-PSL the literature, are compared with our results. In Sect. 5, we dis- For earthshine observations, the lunar disk fills the entire cuss our results and relevant literature in conjunction with the spectrograph slit, so that we had to take dedicated sky back- classification of gas, soil, and ice planets. ground frames to the east and west of Luna. These frames were taken using the same exposure times and procedures as obser- vations of the dark portions of Luna. The illuminated portion 2. Observations of Luna is bright enough for sky glow not to contribute notice- ably, and can be safely neglected (Turnbull et al. 2006). We took All observations for the present compilation of the spectral li- sky background frames to the east and west. Our background brary were performed at Higashi-Hiroshima Observatory us- frames were placed 30 arcmin both in front of and behind Luna’s ing the 1.5 m Kanata telescope and TRISPEC (Triple Range center point. Our earthshine observation was made at 10 arcmin Imager and SPECtrometer with Polarimetry), a simultaneous from the center on the dark side and our observations for Luna optical and near-infrared imager, spectrograph, and polarimeter were made at 10 arcmin on the illuminated side. Thus, we had (Watanabe et al. 2005). Two dichroic mirrors split the incoming equal spacing of 20 arcmin between the exposure points. This light from the telescope into three beams, one optical channel allowed us to account for light from the bright surface that had (0.45−0.90 μm) and two infrared channels IR1 (0.90−1.85 μm) been scattered 20 degrees and would thus show up in our dark and IR2 (1.85−2.5 μm). The instrument is capable of simul- side exposure. taneous three-band spectroscopy across a wavelength range of Bright targets: for bright targets such as Jupiter and crescent 0.45−2.5 μm. We selected the slit of 300 μm width, which gives Luna, detector saturation became a problem. We partially closed the resolving power (R = λ/Δλ)ofR = 138, 142, and 360 for the telescope dome slit to reduce the light gathering power of optical, IR1, and IR2, respectively. The TRISPEC slit length is the telescope. We used this technique only with bright objects, 28 mm, which subtends 7 arcmin on the sky. so that thermal noise from the dome did not have a noticeable The observations were carried out during May 2−13, and effect on the signal quality even near 2.5 μm. November 19−27 of 2008. While the overall data quality is Saturn and its ring: because of the north/south orientation good, shorter than 0.47 μm and between 0.9 and 1.0 μm the spec- of the slit, most of the signals shown for Saturn include con- tra are not reliable because of the dichroic coating problem with tamination from the planet’s strong ring system. We expect that the beam splitters. There is the possibility of thermal contamina- the ring spectrum and gas planets spectrum do not resemble tion from Earth’s atmosphere and the telescope beyond 2.4 μm. each other. Therefore, we obtained a spectrum for Saturn with- Around 1.4 and 1.8 μm, Earth’s atmosphere has strong water ab- out its ring during the November observations, by rotating the sorption bands, where the signal quality is low. spectrograph slit by 90 degrees and placing it on Saturn’s north The bodies observed were: Mercury, Venus, Earth, Luna, pole. As stated above, we are aware that sampling a small part Mars, Ceres, Jupiter, Io, Europa, Ganymede, Callisto, Saturn, of the surface does not necessarily reflect the integrated spec- Dione, Rhea, Titan, Uranus, Neptune, Pluto, and Saturn’s ring trum of the planet, but we will use the resultant spectrum as (Table 1). We obtained Earth’s spectra by observing earthshine, an approximation. the light from the Earth which is reflected by the dark side of the moon. We followed the procedure of Woolf et al. (2002) for ac- quiring and analyzing earthshine. In May, we acquired a Lunar 3. Data analysis spectrum corresponding to a bright crater, while the bulk of the We followed the standard procedure for flat fielding, sky sub- surface (mare) was observed in November.