Dust environment and nucleus spin axis of comet P/Tempel2 from models of the infrared dust tail observed by IRAS

The inverse dust tail model (Fulle 1989) is improved to take into account dust ejection according to the so-called fan effect (Sekanina 1987) and to analyse IR emission images as input data. The improved model is applied to dust tail images of the short period comet P/Tempel 2 obtained by IRAS (Campins et al. 1990). The dust ejection cone axis (i.e. the nucleus spin axis in the fan model) best fitting the input data is characterized by argument Phi, 60 degrees and obliquity I = 70 degrees. This direction differs by about 30 degrees with respect to that found by Sekanina (1987), which gives worse fits of the IRAS images. Intrinsic uncertainties of the fan model or a real nucleus precession may explain such a disagreement. The dust ejection velocity of 1 mm sized grains agrees with that required to fit the width of the trail associated to the same comet (Sykes et al. 1986) and varies within the values 12 +/- 5 m s(-1) from the comet activity onset at 2.9 AU to two months after perihelion. Its maximum occurs a month before perihelion and its size dependence power index is u > -0.2: both features are not explained by most current dust-gas drag models (a detailed review of these models can be found in Crifo 1991). The mass loss rate increases from 50 kg s(-1) at the activity onset to at least 500 kg s(-1) at perihelion, so that P/Tempel 2 is a significant source of interplanetary dust. With respect to the results from optical image analyses, this result has the great advantage to be independent of the poorly known albedo of large grains. The dust size spectrum points out a significant trend of the dominant size of the ejected dust, going from millimeters at the activity onset (at 2.9 AU from the Sun) to decimeters at perihelion (at 1.3 AU from the Sun). It might be related to the expected increasing dust drag efficiency for a decreasing Sun comet distance.