Astron. Astrophys. 336, 654-661 (1998)

## 6. Dust mass-loss rate deduced from the IRAS 60 µm flux

The dust mass-loss rate is obtained from the following formula, based on the results of Jura (1987) and adopted in the form given by Olofsson et al. (1993),

where is given in Jy, the dust velocity in km/s, D in pc, the luminosity L in , and a grain emissivity at 60 µm, , of 150 cmg has been adopted. For our low mass-loss rate objects the maximum of the spectral energy distribution (in Hz) can be calculated from Wien's law, , with (in K) taken from blackbody fits, corrected according to SRIII as .

Following the approach of paper SRIV we corrected for the fraction of the 60µm flux that originates in the stellar photosphere by fitting blackbodies to the broad-band spectra (see SRIII). Corrections of about 10 % were typical within a total range of 5-25 %.

In low mass-loss rate objects the drift velocity between the gas and the dust can be substantial. According to (Kwok 1975) the dust velocity is given by

where Q is the absorption efficiency of the dust grains (assumed to be according to Huggins et al. 1988). We assume that Eq. (3) gives a reasonable upper limit to the dust velocity. For the lowest mass-loss rate objects this formula may result in drift velocities exceeding the limit when substantial grain sputtering is expected to occur, i.e.,  kms (Kwok 1975). In a few cases one consequently has to limit the drift velocity to this value, i.e, let . The resulting dust mass loss rate estimates are given in Table 2. The range between 10 and 300, with a median of 40. This value is so low that one suspects that either the gas mass loss rate is underestimated or else the dust mass loss rate is overestimated. See Sect. 3.5 of SRIV for a short discussion of the problem. Detailed modelling of both dust and gas mass-loss will be needed to arrive at final results.

In our sample it turns out that a typical is around 16 kms. We used this value for large samples of visual O-rich Miras, SRVs and Lbs in order not to be limited to the small number of objects were can be derived from an observed .

Figure 7 shows the resulting distributions in dust mass-loss rates derived in the way described above. The result is comparable to that from the CO observations thus indicating that our small CO sample was not affected by too strong biases. A relatively small range of mass-loss rates is found for all three groups. This further emphasizes that the mass-loss properties appear not to be strongly influenced by the pulsational regularity or even the mode.

 Fig. 7. Distribution of dust mass-loss rates

© European Southern Observatory (ESO) 1998

Online publication: July 20, 1998