next previous
Up: Radiation pressure on

3. Basic equations

3.1. Radiation pressure force

3.1.1. Nonrotating grains

An azimuthal asymmetry of the geometry of the light scattering by non-spherical particles (see Fig. 1 (click here)) provokes the non-coincidence of the directions of the radiation pressure force and of the wave-vector of incident radiation. This fact was mentioned in the literature (e.g., Kerker 1981) but the detailed consideration has been made only for infinite cylinders (Voshchinnikov & Il'in 1983a) and spheroids (Voshchinnikov 1990).

  figure290
Figure 1: Geometry of light scattering by a prolate spheroid with a/b=4. The wave-vector of incident non-polarized radiation tex2html_wrap_inline2066 forms the angle tex2html_wrap_inline2068 with the rotation axis of spheroid. Short-dashed curve shows the angular distribution of the scattered radiation (phase function). The part of scattered radiation which is symmetric relative to the direction of incident radiation is plotted by long-dashed curve. The directions of the resulting radiation force as well as its components in the radial and transversal directions are indicated

The radiation pressure force on a non-spherical particle may be written as follows:
 equation295
where tex2html_wrap_inline2070 is the flux of non-polarized radiation of a point source at frequency tex2html_wrap_inline2072, c the velocity of light, tex2html_wrap_inline2076 and tex2html_wrap_inline2078 are the unit vectors in the radial and transversal directions (tex2html_wrap_inline2080, tex2html_wrap_inline2082, tex2html_wrap_inline2084 is the wave-vector of incident radiation), and tex2html_wrap_inline2086 the radiation pressure cross-sections for the corresponding directions.

We approximate the flux tex2html_wrap_inline2070 in Eq. (4 (click here)) by the Planck function with the effective temperature of the star tex2html_wrap_inline2090 and consider the Planck mean cross-sections and efficiency factors
 equation318

 eqnarray331
Here, tex2html_wrap_inline2092 is the Stephan-Boltzmann constant, tex2html_wrap_inline2094 the refractive index of the grain material, tex2html_wrap_inline2096 the size parameter, tex2html_wrap_inline2098 the angle between the rotation axis of the spheroid and the wave-vector (tex2html_wrap_inline2100), G the geometrical cross-section of a spheroid (the area of the particle shadow) that is
 equation345
for an prolate spheroid and
 equation356
for an oblate spheroid. The superscripts TM and TE are related to two cases of the polarization of incident radiation (TM and TE modes).

The efficiency factors tex2html_wrap_inline2104 are calculated from the solution to the light scattering problem for spheroids (see Voshchinnikov & Farafonov 1993 for details). For both modes, they can be expressed as follows:
 eqnarray370

 eqnarray377
where tex2html_wrap_inline2106 is the extinction efficiency factor, and the coefficients tex2html_wrap_inline2108 and tex2html_wrap_inline2110 are
 equation388

 equation399
In Eqs. (11 (click here)) - (12 (click here)), tex2html_wrap_inline2112 is a parameter, tex2html_wrap_inline2114 the dimensionless intensity of scattered radiation (phase function; Fig. 1 (click here)). From a symmetry consideration, it is clear that for spheroids tex2html_wrap_inline2116 if tex2html_wrap_inline2118 or tex2html_wrap_inline2120. The detailed expressions for calculations of the factors tex2html_wrap_inline2122 are given in the paper of Voshchinnikov (1990).

3.1.2. Rotating grains

For randomly aligned particles, the Planck mean cross-sections have to be averaged over all orientations
 equation418

equation434

For perfect dynamical alignment, the Planck mean cross-sections are averaged over all rotational angles tex2html_wrap_inline2126. This gives for prolate spheroids
 equation439

 eqnarray458
where the angle tex2html_wrap_inline2098 is connected with tex2html_wrap_inline2060 and tex2html_wrap_inline2126 (tex2html_wrap_inline2134). For oblate spheroids randomly aligned in a plane, we have tex2html_wrap_inline2136 and
 equation477

3.2. Equation of motion and drift velocity

The radiation pressure force tex2html_wrap_inline2138 is the main force acting on dust grains in the envelopes of late-type giants. Its ratio to the gravitational force tex2html_wrap_inline2140 is much larger than 1 in almost all cases (Voshchinnikov & Il'in 1983b). As weak magnetic fields and small electric charge on grains are expected in the envelopes, we can neglect the Coulomb drag and Lorentz forces (see Il'in 1994 for details). Thus, the equation of the stationary grain motion should be
 equation498
where tex2html_wrap_inline2142 is the drag force due to collisions of a grain with surrounding gas particles. It is proportional to the gas density, the grain velocity relative to the gas tex2html_wrap_inline2144, and the geometrical cross-section of the grain. For rotating grains, the averaged geometrical cross-section tex2html_wrap_inline2146 or tex2html_wrap_inline2148 should be used. The corresponding expressions are given in Appendix A.

The force tex2html_wrap_inline2142 is antiparallel to tex2html_wrap_inline2152 for spherical particles. For non-spherical grains and subsonic drift velocity, the angle between these two vectors depends on the type of reflection of the gas particles from the grain surface. If the mirror reflection occurs, then tex2html_wrap_inline2142 is perpendicular to the major semiaxis of a spheroidal grain; for diffuse reflection, tex2html_wrap_inline2142 forms a small angle with tex2html_wrap_inline2158, and the value of this angle depends on the difference between the gas and dust temperatures. The angle should be close to zero for any type of reflection in the case of supersonic drift velocity.

From the equation tex2html_wrap_inline2160, the mean drift velocity of a rotating grain is
eqnarray528
for subsonic motion and
equation555
for supersonic motion. Here, tex2html_wrap_inline2162 is the stellar luminosity, tex2html_wrap_inline2164 the mass loss rate, tex2html_wrap_inline2166 the sound velocity, tex2html_wrap_inline2168 are the relative abundances (by mass) of hydrogen (atoms and molecules), helium and heavy elements, respectively.

Thus, the ratio of the drift velocities for spheroidal and spherical particles of the same volume is determined only by the cross-sections of the particles
 equation570
where n = 1 for subsonic motion and n = 2 for supersonic motion.


next previous
Up: Radiation pressure on

Copyright by the European Southern Observatory (ESO)