3. Stability or Instability of a Convex Meniscus
A meniscus is convex if
Remark first that in case of a convex meniscus the function
is increasing. This means that the angle between the tangent line to meniscus in every point , and the OX axis
is decreasing. In particular, it follows that
.Since
and
we obtain inequality
>0 (see
Figure 2).
It should be noted that, the convex meniscus stability (the static one) is different from the dynamic stability of the crystallization process. For statically stable convex meniscus, not only necessary first order but also second order sufficient conditions of functional (2.6) should be satisfied. These second order sufficient conditions for the minimum of functional (2.6) are the Legendre condition and the Jacobi condition [
5].
The Legendre condition is
Computing we find . Therefore the Legendre condition is satisfied.
The Jacoby condition concern the so-called Jacoby equation:
for which a Sturm type upper bound has to be find [
5].
In case of the functional (2.6) Equation (3.2) become:
Remark that for the coefficients of (3.3) the following inequalities hold:
Hence
is a Sturm –type upper bound for (3.3)
Since every non zero solution of the equation
vanishes at most once on the interval
the solution
of the initial value problem
has only one zero on the interval
. Hence the stability condition of Jacobi is verified.
This result can be surprising and create the impression that a convex meniscus is stable. In fact, the result is that if a convex meniscus exist, then it is stable. For this reason in the following, we will establish necessary conditions for the existence of convex meniscus (see
Figure 2).
Starting from Equations (2.3) and (2.5) it is easy to see that in hydrostatic approximation the pressure difference
verify equalities:
Using
, the boundary conditions
,
, with the Lagrange mean value theorem we obtain that there exists
in the interval
such that
Since
is strictly decreasing on the interval
the following inequalities hold
Using equality (3.8) and inequalities (3.9)–(3.11) in hydrostatic approximation, in case of the existence of convex static meniscus, for the pressure difference
the following inequalities hold
Therefore in case of convex meniscus the values of the pressure difference
has to be researched in the interval
where :
For
convex meniscus like in
Figure 2 does not exit. These regions of the pressure difference are regions of static instability. A meniscus obtained with
in this region collapse. For
satisfying one of the inequalities (3.14) it is impossible to create experimentally a convex meniscus like in
Figure 2.
Retain that the pressure difference can be controlled by the gas pressure and the parameter .
In the following we will illustrate first the existence of convex static meniscus in case of Germanium (Ge) assuming that
In case of Ge
[Pa] and
[
Pa] and a convex meniscus is obtained for
. This meniscus is obtained by integrating the initial value problem:
The numerical values of are:, , , [ ] ,=[rad], The pressure difference was found by trial solving (3.15) for different values of in the range .
The obtained results concerning the meniscus shape and the variation of are presented in the next figures:
Figure 4.
Ge convex meniscus shape for
Figure 4.
Ge convex meniscus shape for
Figure 5.
.

,
Figure 5.
.

,
According to the above results concerning stability of the static meniscus that of obtained here numerically is stable. It can be realized experimentally!
We continue analyzing what happens numerically if or .
First we solve (3.15) for
. The obtained result is presented on
Figure 6 and
Figure 7.
This computation shows that for
the meniscus shape is not convex. The purpose to create a convex meniscus, as in
Figure 2 is not realizable experimentally!
Now we solve (3.15) for
. The obtained result is presented on
Figure 8 and
Figure 9.
This computation shows that for
the meniscus shape is not convex. The purpose to create a convex meniscus, as in
Figure 2 is not realizable experimentally!
Remember that in hydrostatic approximation . This relation can be used for the control of via and .
For example if then and for a given gas pressure we can find , namely . For instance if we have = 0.02411897481[m]. This means that the crucible melt level has to be under the shaper top level with 0.02411897481[m].
If and then . This means that the crucible melt level is under the shaper top level with [m].
If and then [m]. This means that the shaper top level is under the crucible melt level with .