Tearing Modes
The layer equations, (9.32) and (9.33), can be solved in a fairly straightforward manner
in Fourier space. Let
where
, and
.
Equations (9.32) and (9.33) can be Fourier transformed, and the results combined, to
give
 |
(9.45) |
where
 |
(9.46) |
The most general small-
asymptotic solution of Equation (9.45) is written
 |
(9.47) |
where
and
are independent of
, and it is assumed that
.
When inverse Fourier transformed, the previous expression leads to the
following expression for the asymptotic behavior of
at the edge of
the layer (Erdéyli 1954):
![$\displaystyle \skew{3}\hat{\phi}(\hat{x})= 2\,{\rm i}\left[\frac{a_0}{S^{1/3}\,\hat{x}} + \frac{\pi\,a_{-1}}{2}\,{\rm sgn}(\hat{x})
+{\cal O}(\hat{x}^2)\right].$](img3392.png) |
(9.48) |
It follows from a comparison with Equations (9.40) and (9.41) that
 |
(9.49) |
Thus, the layer matching parameter,
, is determined from the
small-
asymptotic behavior of the Fourier-transformed layer solution.
Let us search for an unstable perturbation characterized by
. It is
convenient to assume that
 |
(9.50) |
This ordering, which is known as the constant-
approximation [because
it implies that
is approximately constant across the layer]
will be justified later on.
In the limit
, Equation (9.45)
reduces to
 |
(9.51) |
The solution to this equation that is well behaved in the limit
is written
, where
is a standard parabolic cylinder function (Abramowitz and Stegun 1965). In the limit
 |
(9.52) |
we can make use of the standard small-argument expansion
of
to write the most general solution to Equation (9.45) in the
form (Abramowitz and Stegun 1965)
![$\displaystyle \skew{3}\bar{\phi}(p) = A\left[1- \frac{2\,\Gamma(3/4)}{\Gamma(1/4)}\, Q^{1/4}\,p + {\cal O}(p^{2})\right].$](img3403.png) |
(9.53) |
Here,
is an arbitrary constant, and
is a gamma function (Abramowitz and Stegun 1965).
In the limit
 |
(9.54) |
Equation (9.45) reduces to
 |
(9.55) |
The most general solution to this equation is written
 |
(9.56) |
where
and
are arbitrary constants.
Matching coefficients between Equations (9.53) and (9.56) in the range of
satisfying the inequality (9.52) yields the following expression
for the most general solution to Equation (9.45) in the limit
:
![$\displaystyle \skew{3}\bar{\phi} = A\,\left[\frac{2\,\Gamma(3/4)}{\Gamma(1/4)}\, \frac{Q^{5/4}}{p} + 1 + {\cal O}(p)\right].$](img3409.png) |
(9.57) |
Finally, a comparison of Equations (9.47), (9.49), and (9.57)
gives the result
 |
(9.58) |
The asymptotic matching condition (9.42) can be combined with the previous
expression for
to give (Furth, Killeen, and Rosenbluth 1963)
![$\displaystyle \hat{\gamma} = \left[\frac{\Gamma(1/4)}{2\pi\,\Gamma(3/4)}\right]^{4/5}
{\mit\Delta}'^{\,4/5}\,S^{-3/5},$](img3411.png) |
(9.59) |
or
![$\displaystyle \gamma = \left[\frac{\Gamma(1/4)}{2\pi\,\Gamma(3/4)}\right]^{4/5}
\frac{{\mit\Delta}'^{\,4/5}}{\tau_H^{2/5}\,\tau_R^{3/5}}.$](img3412.png) |
(9.60) |
Here, use has been made of the definitions of
,
, and
. According
to the previous equation, the perturbation, which is known as a tearing mode, is unstable whenever
, and grows on the hybrid timescale
. [This hybrid growth time is consistent with our initial assumption (9.29), provided that
.]
It is easily demonstrated that the tearing mode is stable whenever
. Thus, we can now appreciate that
the solid curve in Figure 9.2, which is indented at the top (because
), is the outer solution of
an unstable tearing mode, whereas the dashed curve (which is not indented) is the outer solution of a stable
tearing mode.
Note, finally, that
as
. In other words, the instability of the
current sheet when
is only made possible by finite plasma resistivity.
According to Equations (9.42), (9.50), and (9.58), the constant-
approximation holds provided that
 |
(9.61) |
that is, provided that the tearing mode does not become too
unstable.
Equation (9.51) implies that thickness of the layer in
-space
is
 |
(9.62) |
It follows from Equations (9.43), (9.44), and (9.46) that the thickness of the layer in
-space
is
 |
(9.63) |
When
then
, according to
Equation (9.59), giving
. It is clear, therefore, that if
the Lundquist number,
, is very large then the resistive layer centered
on the shear-Alfvén resonance,
, is extremely narrow compared to the width of the current sheet.
The timescale for magnetic flux to diffuse across a layer of thickness
(in
-space) is [see Equation (9.25)]
 |
(9.64) |
If
 |
(9.65) |
then the tearing mode grows on a timescale that is far longer than the
timescale on which magnetic flux diffuses across the layer. In this case,
we would expect the normalized flux,
, to be approximately
constant across the layer, because any non-uniformities in
would be
smoothed out via resistive diffusion. It follows from Equations (9.63) and (9.64)
that the constant-
approximation holds provided that
 |
(9.66) |
(i.e.,
), which is in agreement with Equation (9.50).