© 2005 Nature Publishing Group A large dust/ice ratio in the nucleus of comet 9P/Tempel 1 Michael Ku ¨ppers 1 , Ivano Bertini 2 , Sonia Fornasier 2 , Pedro J. Gutierrez 3 , Stubbe F. Hviid 1 , Laurent Jorda 4 , Horst Uwe Keller 1 , Jo ¨rg Knollenberg 5 , Detlef Koschny 6 , Rainer Kramm 1 , Luisa-Maria Lara 3 , Holger Sierks 1 , Nicolas Thomas 7 , Cesare Barbieri 2 , Philippe Lamy 4 , Hans Rickman 8 , Rafael Rodrigo 3 & the OSIRIS team* Comets spend most of their life in a low-temperature environment far from the Sun. They are therefore relatively unprocessed and maintain information about the formation conditions of the planetary system, but the structure and composition of their nuclei are poorly understood. Although in situ 1 and remote 2 measurements have derived the global properties of some come- tary nuclei, little is known about their interiors. The Deep Impact mission 3 shot a projectile into comet 9P/Tempel 1 in order to investigate its interior. Here we report the water vapour content (1.5 3 10 32 water molecules or 4.5 3 10 6 kg) and the cross-section of the dust (330 km 2 assuming an albedo of 0.1) created by the impact. The corresponding dust/ice mass ratio is probably larger than one, suggesting that comets are ‘icy dirtballs’ rather than ‘dirty snowballs’ as commonly believed 4 . High dust velocities (between 110 m s 21 and 300 m s 21 ) imply acceleration in the comet’s coma, probably by water molecules sublimated by solar radiation. We did not find evidence of enhanced activity of 9P/Tempel 1 in the days after the impact, suggesting that in general impacts of meteoroids are not the cause of cometary outbursts. On 4 July 2005, the impactor of the NASA Deep Impact mission hit the surface of comet 9P/Tempel 1 with a relative velocity of 10.2 km s 21 . The collision of the impactor with a mass of 362 kg was expected to generate a crater (predicted diameter ,100–125 m; ref. 3) and eject cometary material. Possibly it would also trigger an outburst and a new active area on the comet’s surface. The event was observed by many ground-based and some space- based telescopes. We used the scientific OSIRIS camera system on board the ESA Rosetta spacecraft to observe the comet before and after the impact. OSIRIS is composed of a narrow angle camera (NAC) and a wide angle camera (WAC). Both cameras are unob- structed mirror systems and are equipped with 2,048 £ 2,048 pixel backlit/back-thinned CCD detectors with a sensitivity range from 245 nm to 1,000 nm and a maximum quantum efficiency of 95%. At the time of impact Rosetta was located at a distance of 0.53 AU from 9P/Tempel 1 with a solar elongation angle of 918 and a phase angle of 698. The heliocentric distance of the comet was 1.51 AU. OSIRIS started to observe the comet on 28 June 2005 at 23:45 UTC and finished on 14 July 2005 15:00 UTC. The time resolution was generally 1.5–3 h and was increased (with the NAC) to approximately 1 min between 30 min before and 90 min after the impact. The NAC imaged the extended dust coma in different filters. The WAC was com- manded to take images through the OH and a neighbouring continuum filter. In addition Na, CN, and O I filters were used to monitor further gas emissions. An asymmetry of the ejected dust cloud is clearly visible for several days after the impact (Fig. 1). The analysis presented sepa- rates this debris from the background of the normal coma. The asymmetry of the gas (OH) is less visible because of the reduced spatial resolution (31,200 km) and the lower signal to noise ratio. Immediately after impact both cameras act like photometers until the impact-related dust and gas leaves the corresponding resolution element. The water (H 2 O) production rate of comet 9P/Tempel 1 was derived from the OH emission (308 nm). A scaled image of the ultraviolet dust continuum (at 375 nm) was subtracted from each OH image, assuming solar type reflectivity for the cometary dust. Pre-launch laboratory calibration and observations of Vega (a Lyrae) were used for conversion of data numbers into flux units. Toestimate the water production rate before the impact, the flux from OH molecules was added within circular areas with radii between 31,200 km and 156,000 km centred on the position of the cometary nucleus. A fluorescence efficiency of 1.5 £ 10 222 W molecule 21 at 1 AU (ref. 5) was scaled to the heliocentric distance of the comet and used to convert fluxes to number of molecules. The Haser formula 6,7 for the number of molecules within a circular aperture is used to convert the number of OH molecules to the water production rate of the comet. The underlying assumptions are that OH is derived from dissociation of cometary water, and that both parent (H 2 O) and daughter (OH) molecules are flowing radially away from the nucleus with a constant velocity. We used typical values of the photodissociation timescales for H 2 O (7.67 £ 10 4 s at 1 AU) and OH (1.32 £ 10 5 s) 8 , scaled to the helio- centric distance of comet 9P/Tempel 1. A fraction (86%) of all water molecules dissociate into OH þ H (ref. 9). The resulting pre-impact water production rate is (3.4 ^ 0.5) £ 10 27 molecules s 21 for a typical gas outflow velocity of 0.7 km s 21 at 1.5 AU from the Sun 10 , and (5.8 ^ 1) £ 10 27 molecules s 21 for an outflow velocity of 1 km s 21 , which is frequently used in the derivation of Haser parameters 11 . The higher value for the production rate is in good agreement with other estimates from near-ultraviolet measurements 12 . Figure 2 shows the number of OH molecules created by photo- dissociation of H 2 O in the impact cloud. We estimate that (1.5 ^ 0.5) £ 10 32 water molecules (or 4.6 £ 10 6 kg) were created by the impact. This value does not depend on the outflow velocity of the water or OH, and corresponds to approximately 20% of the water molecules that were present in the cometary coma owing to normal activity. LETTERS *Lists of participants and affiliations appear at the end of the paper. 1 Max-Planck Institut fu ¨r Sonnensystemforschung, Max-Planck-Str. 2, 37191 Katlenburg-Lindau, Germany. 2 Dipartimento di Astronomia and CISAS, Universita ` di Padova, Vicolo dell’Osservatorio 5, 35100 Padova, Italy. 3 Instituto de Astrofisica de Andalucı ´a-CSIC, C/Camino Bajo de Hue´tor, 50, 18008 Granada, Spain. 4 Laboratoire d’Astrophysique de Marseille, Traverse du Siphon, Les Trois Lucs BP 8, 13376 Marseille, France. 5 DLR Institute for Planetary Research, Rutherfordstr. 2, 12489 Berlin, Germany. 6 European Space Agency, ESTEC, SCI-SB, Keplerlaan 1, Postbus 299, 2200 AG Noordwijk, The Netherlands. 7 Physikalisches Institut, Abteilung Weltraumforschung und Planetologie, Universita ¨t Bern, Sidlerstr. 5, 3012 Bern, Switzerland. 8 Uppsala Astronomical Observatory, Box 515, 75120 Uppsala, Sweden. Vol 437|13 October 2005|doi:10.1038/nature04236 987