Astron. Astrophys. 329, 451-481 (1998)
Appendix A: axial ratios for surfaces of equal density and triaxiality parameter
As mentioned in Sect. 3, the smooth density function of the
logarithmic potential has the following form:
![[EQUATION]](img385.gif)
where , ,
, are functions of
and :
![[EQUATION]](img388.gif)
It can be directly perceived from the form of Eq. (A1) that the
surfaces of equal density are not ellipsoidal. We may thus calculate
approximate axial ratios through the parameters
, and
defined as the points where the considered
isodensity cuts the corresponding x, y and z
axis, respectively. Unfortunately, these parameters are not unique,
i.e. they are parameterized by the value of the considered isodensity.
Indeed, by nullifying successively two out of three values of the
spatial variables we have the following equations for the density
function:
![[EQUATION]](img392.gif)
with and the convention that
. By resolving three quadratic equations with
respect to and excluding the negative
solutions, we have:
![[EQUATION]](img395.gif)
Thus, for each value of the density , we are
able to determine the density axial ratios. Now, we may notice that
the value of the density on each axis is minimised when the
corresponding spatial variable takes its maximum value. However, the
values of the density for maximum values of the variables
corresponding to the middle and short axes are always smaller than the
value of the last isodensity cutting the long axis. Hence, the smaller
possible value for an isodensity surface is given by the maximum value
of the variable labelling the long axis. More specifically, the
smaller possible value of the isodensity surfaces is given by the
equation:
![[EQUATION]](img396.gif)
and or 2 or 3 depending on which of the
spatial variables x, y or z labels the long axis
of the system, following the values of and
. In that way better estimates of the isodensity
axial ratios can be determined:
![[EQUATION]](img398.gif)
where the axial ratios are denoted as and
, with ,
labelling the short, middle and long axes,
respectively, with the previously employed convention for the
correspondence between the numbers and the spatial variables. Finally,
through these axial ratios we may determine a triaxiality parameter
(Eq. 28):
![[EQUATION]](img403.gif)
© European Southern Observatory (ESO) 1998
Online publication: December 8, 1997
helpdesk.link@springer.de  |