1 Introduction
Several phenomena in tokamaks, such as the saw-tooth oscillations, plasma disruptions and confinement degradation, appear to involve tearing mode activity. Studies of the linear tearing stability for hot plasma have exploited the separation of scales between a narrow ‘inner’ radial region around a resonant surface where reconnection processes occur and the remaining ‘external’ region where a marginal ideal magnetohydrodynamic (MHD) model is adequate. The ideal MHD solution, characterised by a quantity
$\unicode[STIX]{x1D6E5}^{\prime }$
, is matched to a corresponding quantity
$\hat{\unicode[STIX]{x1D6E5}}(\unicode[STIX]{x1D714}+\text{i}\unicode[STIX]{x1D6FE})$
calculated from the inner solution in order to determine the mode frequency
$\unicode[STIX]{x1D714}$
and growth rate,
$\unicode[STIX]{x1D6FE}$
. Whereas early studies of linear stability used a simple resistive MHD model for the inner layer, present day, hot, tokamaks require a much more complete physics model.
In a previous paper (Connor, Hastie & Helander Reference Connor, Hastie and Helander2009) we presented layer equations for determining the stability of semi-collisional tearing modes in a toroidal plasma in the banana regime of collisionality (
$\unicode[STIX]{x1D708}_{j,\text{eff}}<\unicode[STIX]{x1D714}_{bj}$
where
$\unicode[STIX]{x1D708}_{j,\text{eff}}\sim \unicode[STIX]{x1D708}_{j}/f_{t}^{2}$
is the effective collision frequency and
$\unicode[STIX]{x1D714}_{bj}\sim f_{t}v_{\text{thj}}/L_{c}$
the bounce frequency of a particle of species
$j$
,
$f_{t}\ll 1$
being the fraction of trapped particles,
$\unicode[STIX]{x1D708}_{j}$
the frequency for 90 degree Coulomb collisions and
$L_{c}$
the connection length around the torus), therefore incorporating neoclassical physics. The semi-collisional ordering involves the balance:
$\unicode[STIX]{x1D714}\sim k_{\Vert }^{2}v_{\text{the}}^{2}/\unicode[STIX]{x1D708}_{e}$
, where
$k_{\Vert }=k_{\unicode[STIX]{x1D703}}x/L_{s}$
is the wavenumber parallel to the magnetic field (
$k_{\unicode[STIX]{x1D703}}$
is a poloidal wavenumber,
$x$
is the distance from a rational surface and
$L_{s}$
is the magnetic shear length),
$v_{\text{the}}$
is the electron thermal speed and
$\unicode[STIX]{x1D708}_{e}$
is the electron 90 degree collision frequency, so that parallel transport processes compete with the mode frequency,
$\unicode[STIX]{x1D714}$
. This balance serves to define the semi-collisional width,
$\unicode[STIX]{x1D6FF}_{e}=(\unicode[STIX]{x1D714}\unicode[STIX]{x1D708}_{e}L_{s}^{2}/k_{\unicode[STIX]{x1D703}}^{2}v_{e}^{2})^{1/2}$
.
These equations were formulated for general axisymmetric geometry, thus in this respect extending the work of Fitzpatrick (Reference Fitzpatrick1989). A consistent ordering for semi-collisional theory requires that, as well as the inclusion of parallel collisional transport processes, one should also incorporate collisional cross-field transport. The equations of Fitzpatrick (Reference Fitzpatrick1989) did indeed include both these transport processes, albeit using a simplified model collision operator. While Connor et al. (Reference Connor, Hastie and Helander2009) discussed the role of cross-field transport based on a Lorentz collision operator, thus ignoring like-particle collisions, the emphasis was on the basic semi-collisional physics and these effects were ignored in the bulk of the paper. The role of the present paper is to rectify this limitation by providing a general axisymmetric formulation, including cross-field transport which can be used as a semi-collisional tearing mode layer module for inclusion in numerical codes that asymptotically match the layer to toroidal calculations of the tearing mode stability index,
$\unicode[STIX]{x1D6E5}^{\prime }$
(Glasser, Greene & Johnson Reference Glasser, Greene and Johnson1975). We also include some additional, relatively small, effects in the electron continuity and thermal equations arising from the poloidal magnetic drift that were ignored by Connor et al. (Reference Connor, Hastie and Helander2009).
The model pitch-angle scattering collision operator used by Fitzpatrick (Reference Fitzpatrick1989) provides a good description for distribution functions localised in velocity space around the trapped particle regions, but electric fields, parallel pressure gradients and thermal forces due to parallel temperature gradients generate distortions of the whole passing particle region. Nevertheless, this can be circumvented by the use of the Spitzer functions (Cohen, Spitzer & McR. Routly Reference Cohen, Spitzer and McR. Routly1950; Spitzer & Harm Reference Spitzer and Harm1953) to account for these drives, as demonstrated by Rosenbluth, Hazeltine & Hinton (Reference Rosenbluth, Hazeltine and Hinton1972), Connor et al. (Reference Connor, Grimm, Hastie and Keeping1973) and Helander & Sigmar (Reference Helander and Sigmar2002); we shall also adopt this approach here. The calculation closely follows neoclassical transport theory but differs in one respect. In equilibrium the electron density and temperature are constant on a flux surface and the only parallel driving force is due to the toroidal electric field, resulting in a role for the Spitzer function related to electrical conductivity. In stability theory, however, parallel gradients of both density and electron temperature can persist, leading to the need to involve the Spitzer function describing the parallel heat flux. The calculation below makes one assumption, namely that the fraction of trapped particles is small. This can be relaxed, albeit leading to more complex algebra, but the distribution functions become less localised and the asymptotic accuracy of the approach is compromised. The derivation of the vorticity equation calls for an expression for the neoclassical radial transport of toroidal momentum and we extend previous work by Wong & Chan (Reference Wong and Chan2005) to cover more general geometry, though needing to use a model collision operator in order to determine the required adjoint function.
In § 2 we introduce the gyro-kinetic model for electrons and ions. An appropriate ordering scheme is used to obtain solutions for the ion and electron distribution functions in § 3. To complete these solutions, equations for the perturbed densities and temperature of the two species are required. These are obtained in § 4. The numerical values of various averages of collision frequencies that are required in §§ 3 and 4 are listed in appendix A. Section 5 develops Ampère’s law and the vorticity equation to complete the set of equations needed to calculate
$\hat{\unicode[STIX]{x1D6E5}}(\unicode[STIX]{x1D714}+\text{i}\unicode[STIX]{x1D6FE})$
. A calculation of the neoclassical radial transport of toroidal angular momentum for general geometry, required in the development of the vorticity equation, is performed in appendix B. Section 6 introduces a set of convenient normalisations for the set of equations. In § 7 we discuss an intermediate radial region needed to connect the solutions of this set of equations to the ideal MHD region where
$\unicode[STIX]{x1D6E5}^{\prime }$
is defined. Finally, we draw some conclusions in § 8.
2 The gyro-kinetic equation
We describe the plasma species
$j$
by the gyro-kinetic equation (Tang, Connor & Hastie Reference Tang, Connor and Hastie1980):
$$\begin{eqnarray}\displaystyle & & \displaystyle (-\text{i}\unicode[STIX]{x1D714}+v_{\Vert }\boldsymbol{b}\boldsymbol{\cdot }\unicode[STIX]{x1D735}+\boldsymbol{v}_{dj}\boldsymbol{\cdot }\unicode[STIX]{x1D735})g_{j}-\overline{\exp (-\text{i}L_{j})C_{j}(g_{j}\exp (\text{i}L_{j}))}\nonumber\\ \displaystyle & & \displaystyle \quad =-\text{i}\frac{e_{j}f_{0j}}{T_{j}}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast j}^{T})\left((\unicode[STIX]{x1D6F7}-v_{\Vert }A_{\Vert })\text{J}_{0}(z_{j})+\frac{v_{\bot }}{k_{\bot }}\frac{\unicode[STIX]{x1D6FF}B_{\Vert }}{B}J_{1}(z_{j})\right),\end{eqnarray}$$
where
$\boldsymbol{b}$
is a unit vector along the magnetic field,
$v_{\Vert }$
is the particle velocity along the magnetic field,
$\unicode[STIX]{x1D6F7}$
is the perturbed electrostatic potential,
$A_{\Vert }$
is the perturbed parallel component of the vector potential,
$\unicode[STIX]{x1D6FF}B_{\Vert }$
is the parallel component of the perturbed magnetic field and we have written the perturbed distribution as
with perturbation time dependencies:
$\exp (-\text{i}\unicode[STIX]{x1D714}t)$
. Here,
$L_{j}=\boldsymbol{k}\times \boldsymbol{v}_{\bot }\boldsymbol{\cdot }\boldsymbol{b}/\unicode[STIX]{x1D6FA}_{j}$
, with
$\boldsymbol{k}$
the wavenumber and
$v_{\bot }$
the velocity perpendicular to the magnetic field, is the gyro-phase factor, the operation
$\overline{A}$
is a gyro-phase average over the quantity
$A$
,
$\text{J}_{0,1}$
are Bessel functions of argument:
$z_{j}=k_{\bot }v_{\bot }/\unicode[STIX]{x1D6FA}_{j}$
,
$f_{0j}=n_{j}(\unicode[STIX]{x1D712})(m_{j}/2\unicode[STIX]{x03C0}T_{j}(\unicode[STIX]{x1D712}))^{3/2}\exp (-m_{j}v^{2}/2T_{j}(\unicode[STIX]{x1D712}))$
are Maxwell distributions, with
$\unicode[STIX]{x1D712}$
the poloidal flux, and
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}\unicode[STIX]{x1D714}_{\ast j}^{T}=\unicode[STIX]{x1D714}_{\ast j}\left(1+\unicode[STIX]{x1D702}_{j}\left({\displaystyle \frac{m_{j}v^{2}}{2T_{j}}}-{\displaystyle \frac{3}{2}}\right)\right),\quad \unicode[STIX]{x1D714}_{\ast j}=\ell {\displaystyle \frac{T_{j}}{e_{j}}}{\displaystyle \frac{\unicode[STIX]{x2202}\ell nn_{j}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}},\quad \unicode[STIX]{x1D702}_{j}={\displaystyle \frac{\unicode[STIX]{x2202}\ell nT_{j}}{\unicode[STIX]{x2202}\ell nn_{j}}},\\ \boldsymbol{v}_{dj}={\displaystyle \frac{1}{\unicode[STIX]{x1D6FA}_{j}}}\boldsymbol{b}\times \left({\displaystyle \frac{\boldsymbol{v}_{\bot }^{2}\unicode[STIX]{x1D735}\ell nB}{2}}+\boldsymbol{v}_{\Vert }^{2}\unicode[STIX]{x1D73F}\right),\quad \unicode[STIX]{x1D6FA}_{j}={\displaystyle \frac{e_{j}B}{m_{j}}},\quad \unicode[STIX]{x1D73F}=\boldsymbol{b}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\boldsymbol{\cdot }\boldsymbol{b},\end{array}\right\}\end{eqnarray}$$
with
$v$
the particle speed,
$\unicode[STIX]{x1D6FA}_{j}$
the cyclotron frequency of species
$j$
,
$\unicode[STIX]{x1D73F}$
the curvature vector,
$\ell$
the toroidal mode number of the perturbations and all gradients are taken at constant
$(\unicode[STIX]{x1D707},v)$
, with
$\unicode[STIX]{x1D707}=m_{j}v_{\bot }^{2}/2B$
the magnetic moment, or
$(\unicode[STIX]{x1D706},v)$
with
$\unicode[STIX]{x1D706}=2\unicode[STIX]{x1D707}/m_{j}v^{2}$
(thus
$v_{\Vert }=\unicode[STIX]{x1D70E}v\sqrt{1-\unicode[STIX]{x1D706}B}$
, with
$\unicode[STIX]{x1D70E}=\text{sign}(v_{\Vert })$
). If
$\unicode[STIX]{x1D703}$
is defined so that the safety factor
$q=B\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D711}/B\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}$
, where
$B=I\unicode[STIX]{x1D735}\unicode[STIX]{x1D711}+\unicode[STIX]{x1D735}\unicode[STIX]{x1D711}\times \unicode[STIX]{x1D735}\unicode[STIX]{x1D712}$
with
$\unicode[STIX]{x1D711}$
the axisymmetric toroidal angle, is a flux function, then
If we let
etc., then
where
$\ell q-m=\ell q^{\prime }x$
;
$q^{\prime }=\text{d}q/\text{d}\unicode[STIX]{x1D712}$
with
$x=\unicode[STIX]{x1D712}-\unicode[STIX]{x1D712}_{s}$
,
$\unicode[STIX]{x1D712}_{s}$
being the resonant surface where
$m=\ell q(\unicode[STIX]{x1D712}_{s})$
. Here prime denotes a derivative with respect to
$\unicode[STIX]{x1D712}$
, or equivalently,
$x$
.
Thus
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{Iv_{\Vert }}{R^{2}Bq}\left(\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}+\text{i}\ell q^{\prime }x\right)g_{j}+\boldsymbol{v}_{dj}\boldsymbol{\cdot }\unicode[STIX]{x1D735}g_{j}-\text{i}\unicode[STIX]{x1D714}g_{j}-\overline{\exp (-\text{i}L_{j})C_{j}(g_{j}\exp (\text{i}L_{j}))}\nonumber\\ \displaystyle & & \displaystyle \quad =-\text{i}\frac{e_{j}f_{0j}}{T_{j}}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast j}^{T})\left((\unicode[STIX]{x1D6F7}-v_{\Vert }A_{\Vert })\text{J}_{0}(z_{j})+\frac{v_{\bot }}{k_{\bot }}\frac{\unicode[STIX]{x1D6FF}B_{\Vert }}{B}J_{1}(z_{j})\right).\end{eqnarray}$$
Assuming both species are magnetised with
$k_{\bot }v_{\bot }/\unicode[STIX]{x1D6FA}_{j}\ll 1$
, we expand
$\exp (iL_{j})\cong 1+\text{i}L_{j}$
and the Bessel functions for small
$z_{j}$
. We also introduce
$h_{j}$
:
where
$A_{\Vert }=-(\text{i}/\unicode[STIX]{x1D714})\unicode[STIX]{x1D735}_{\Vert }\unicode[STIX]{x1D6F9}$
, so that the parallel electric field,
$E_{\Vert }$
, is given by
Then the fundamental kinetic equations are
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{Iv_{\Vert }}{R^{2}Bq}\left(\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}+\text{i}\ell q^{\prime }x\right)h_{j}+\boldsymbol{v}_{dj}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{j}-\text{i}\unicode[STIX]{x1D714}h_{j}-C_{j}(h_{j})-\overline{L_{j}C_{j}(g_{j}L_{j})}\nonumber\\ \displaystyle & & \displaystyle \quad =-\,\text{i}\frac{e_{j}f_{0j}}{T_{j}}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast j}^{T})\left(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9}+\frac{v_{\bot }^{2}}{2\unicode[STIX]{x1D6FA}_{j}}\unicode[STIX]{x1D6FF}B_{\Vert }\right)-\frac{e_{j}f_{0j}}{T_{j}}\left(1-\frac{\unicode[STIX]{x1D714}_{\ast j}^{T}}{\unicode[STIX]{x1D714}}\right)\boldsymbol{v}_{dj}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D6F9}.\qquad\end{eqnarray}$$
We note that relations (2.2) and (2.8) imply the perturbed density
${\tilde{n}}_{j}$
is given by
where
$\hat{n}_{j}$
is the leading contribution to the density from
$h_{j}$
.
The perturbed and equilibrium quasi-neutrality conditions,
allow us to obtain
Similarly, from (2.2) and (2.8) we can obtain the perturbed temperatures:
where
$\hat{T}_{i}=(2T_{j}/3)\int \text{d}^{3}\boldsymbol{v}(m_{j}v^{2}/2T_{j}-3/2)h_{j}$
is the contribution from
$h_{j}$
etc.
3 The ion and electron solutions
3.1 The ion solution
We consider the ‘collisional’ case:
$\unicode[STIX]{x1D708}_{ii}>\unicode[STIX]{x1D714}$
. Introducing the proton charge,
$e$
, we have
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{Iv_{\Vert }}{R^{2}Bq}\left(\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}+\text{i}\ell q^{\prime }x\right)\,h_{i}+\boldsymbol{v}_{di}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{i}-\text{i}\unicode[STIX]{x1D714}h_{i}-C_{i}(h_{i})-\overline{L_{i}C_{i}(g_{i}L_{i})}\nonumber\\ \displaystyle & & \displaystyle \quad =-\text{i}\frac{ef_{0i}}{T_{i}}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast i}^{T})\left(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9}+\frac{v_{\bot }^{2}}{2\unicode[STIX]{x1D6FA}_{i}}\unicode[STIX]{x1D6FF}B_{\Vert }\right)-\frac{ef_{0i}}{T_{i}}\left(1-\frac{\unicode[STIX]{x1D714}_{\ast i}^{T}}{\unicode[STIX]{x1D714}}\right)\boldsymbol{v}_{di}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D6F9}.\end{eqnarray}$$
We solve this equation by introducing an ordering scheme in terms of a small parameter,
$\unicode[STIX]{x1D700}_{i}$
, where
$\unicode[STIX]{x1D714}_{dri}\sim \unicode[STIX]{x1D708}_{i}\sim \unicode[STIX]{x1D700}_{i}\unicode[STIX]{x1D714}_{bi}$
;
$\unicode[STIX]{x1D714}\sim \unicode[STIX]{x1D714}_{dr\unicode[STIX]{x1D703}}\sim \unicode[STIX]{x1D700}_{i}^{2}\unicode[STIX]{x1D714}_{bi}$
; we also order
$z_{i}\sim \unicode[STIX]{x1D700}_{i}$
. Although we shall later assume a small number of trapped particles,
$f_{t}\ll 1$
, we do not order it in
$\unicode[STIX]{x1D700}_{i}$
. The radial magnetic drift frequency,
$\unicode[STIX]{x1D714}_{dri}$
, exceeds the azimuthal drift frequency,
$\unicode[STIX]{x1D714}_{dr\unicode[STIX]{x1D703}}$
, because of the narrow radial width of the semi-collisional layer.
Writing
$h_{i}=h_{0i}+\unicode[STIX]{x1D700}_{i}h_{1i}+\unicode[STIX]{x1D700}_{i}^{2}h_{2i}+\cdots \,$
, the lowest-order solution satisfies
while in first order we obtain
where we note
Thus
$h_{1i}$
for passing particles can be annihilated by applying the operation:
$\langle B(\ldots )/v_{\Vert }\rangle$
:
since
For trapped particles, we integrate along the bounce orbit, summing over
$\unicode[STIX]{x1D70E}$
in the usual way, to obtain the constraint
where the integration is now between bounce points. This determines
$h_{0i}$
, yielding:
where
$u^{2}=m_{j}v^{2}/2T_{j}$
with
$j=i$
.
The equation for
$h_{1i}$
can then be integrated to give
where
$H_{1i}$
remains to be determined. This can be rewritten as
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}h_{1i}={\displaystyle \frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{i}}}\left[A_{1}^{(i)}+\left(u^{2}-{\displaystyle \frac{5}{2}}\right)A_{2}^{(i)}\right]f_{0i}+H_{1i}(v,\unicode[STIX]{x1D706},x,\unicode[STIX]{x1D70E}),\\ A_{1}^{(i)}=-{\displaystyle \frac{e\unicode[STIX]{x1D6F9}^{\prime }}{T_{i}}}\left(1-{\displaystyle \frac{\unicode[STIX]{x1D714}_{\ast i}}{\unicode[STIX]{x1D714}}}(1+\unicode[STIX]{x1D702}_{i})\right)-\left({\displaystyle \frac{\hat{n}_{i}^{\prime }}{n_{0}}}+{\displaystyle \frac{\hat{T}_{i}^{\prime }}{T_{i}}}\right)\equiv -\left({\displaystyle \frac{\tilde{p}_{i}^{\prime }}{p_{0i}}}+{\displaystyle \frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{i}}}\right),\\ A_{2}^{(i)}=-\left(-{\displaystyle \frac{\unicode[STIX]{x1D714}_{\ast i}\unicode[STIX]{x1D702}_{i}}{\unicode[STIX]{x1D714}}}{\displaystyle \frac{e\unicode[STIX]{x1D6F9}^{\prime }}{T_{i}}}+{\displaystyle \frac{\hat{T}_{i}^{\prime }}{T_{i}}}\right)\equiv -{\displaystyle \frac{\tilde{T}_{i}^{\prime }}{T_{i}}},\end{array}\right\}\end{eqnarray}$$
where
$p_{0i}=n_{0}T_{i}$
, on using (2.13) and (2.14). Note that the quantities
$\tilde{p}_{i}^{\prime }$
and
$\tilde{T}_{i}^{\prime }$
thus defined, depend on
$\unicode[STIX]{x1D714}$
.
The equation for
$h_{2i}$
is
(where, for most purposes, we can ignore the small ion–electron collisional term, but see § 7 later), which provides the constraint
For the ions, we take the model pitch-angle scattering collision operator that conserves momentum (Rutherford et al. Reference Rutherford, Kovrizhnykh, Rosenbluth and Hinton1970):
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}C_{ii}(h)=\unicode[STIX]{x1D708}_{ii}(u)\left(L(h)+{\displaystyle \frac{m_{i}}{T_{i}}}v_{\Vert }U_{\Vert i}^{\ast }(h)f_{0i}\right),\quad L(h)={\displaystyle \frac{2v_{\Vert }}{v^{2}B}}{\displaystyle \frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D706}}}\left(\unicode[STIX]{x1D706}v_{\Vert }{\displaystyle \frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D706}}}h\right),\\ U_{\Vert i}^{\ast }(h)=\left.\displaystyle \int \,\text{d}^{3}v\unicode[STIX]{x1D708}_{ii}(u)v_{\Vert }h\right/\displaystyle \int \,\text{d}^{3}vf_{0i}\unicode[STIX]{x1D708}_{ii}(u)m_{i}v_{\Vert }^{2}/2T_{i},\quad \unicode[STIX]{x1D708}_{ii}(u)=\unicode[STIX]{x1D708}_{0i}{\displaystyle \frac{\unicode[STIX]{x1D711}(u)}{u^{3}}},\\ \unicode[STIX]{x1D708}_{0i}={\displaystyle \frac{n_{0}\text{e}^{4}\ln \unicode[STIX]{x1D6EC}}{4\unicode[STIX]{x03C0}\unicode[STIX]{x1D700}_{0}^{2}m_{i}^{2}}}\left({\displaystyle \frac{m_{i}}{2T_{i}}}\right)^{3/2},\quad \unicode[STIX]{x1D711}(u)=\left(1-{\displaystyle \frac{1}{2u^{2}}}\right)\unicode[STIX]{x1D702}(u)+{\displaystyle \frac{\unicode[STIX]{x1D702}^{\prime }(u)}{2u}},\\ \unicode[STIX]{x1D702}(u)={\displaystyle \frac{2}{\sqrt{\unicode[STIX]{x03C0}}}}\displaystyle \int _{0}^{u}\,\text{d}t\text{e}^{-t^{2}}.\end{array}\right\}\end{eqnarray}$$
The constraint (3.12) yields an expression for
$H_{1i}$
so that:
where
$H(\unicode[STIX]{x1D706}-\unicode[STIX]{x1D706}_{c})$
is the Heaviside function and
Using (3.14) and integrating over velocity space:
we can calculate
$U_{\Vert \,i}^{\ast }$
self-consistently. Thus
where we have defined the symbol
$\left\{\ldots ..\right\}$
by
and
where
$f_{c}=1-f_{t}$
, with
$f_{t}$
the trapped particle fraction defined by
Solving self-consistently for
$\langle BU_{\Vert i}^{\ast }\rangle$
, substituting for
$A_{1}^{(i)}$
and
$A_{2}^{(i)}$
from (3.10) and using the values of the collisional integrals from appendix A, we obtain:
Then one can calculate the ion parallel flux from (3.14), to give
It remains to address the determination of
$\hat{n}_{i}$
and
$\hat{T}_{i}$
;
$\hat{n}_{i}$
is already given in terms of
${\tilde{n}}$
by (2.13), whereas
$\hat{T}_{i}$
is obtained by applying the operator
$\int (\ldots )(m_{i}v^{2}/2)\,\text{d}^{3}\boldsymbol{v}$
to the next-order equation,
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}h_{3i}+\frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{i}}\right)\frac{\unicode[STIX]{x2202}h_{2i}}{\unicode[STIX]{x2202}x}+\boldsymbol{v}_{d\unicode[STIX]{x1D703}i}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0i}-\text{i}\unicode[STIX]{x1D714}h_{0i}-C_{i}(h_{2i})-\overline{L_{i}C_{i}(g_{0i}L_{i})}\nonumber\\ \displaystyle & & \displaystyle \quad =-\frac{\text{i}e}{T_{i}}f_{0i}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast i}^{T})\left(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9}+\frac{v_{\bot }^{2}}{2\unicode[STIX]{x1D6FA}_{i}}\unicode[STIX]{x1D6FF}B_{\Vert }\right),\end{eqnarray}$$
since it annihilates the first term on the left by means of the flux surface average and the collision term,
$C_{i}\left(h_{2i}\right)$
, due to its conservation properties. This will be elaborated in § 4.2.
3.2 The electron solution
For the electrons, equation (2.10) becomes
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{Iv_{\Vert }}{R^{2}Bq}\left(\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}+\text{i}\ell q^{\prime }x\right)h_{e}+\boldsymbol{v}_{de}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{e}-\text{i}\unicode[STIX]{x1D714}h_{e}-C_{e}(h_{e})-\overline{L_{e}C_{e}(g_{e}L_{e})}\nonumber\\ \displaystyle & & \displaystyle \quad =\text{i}\frac{ef_{0e}}{T_{j}}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast e}^{T})\left(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9}+\frac{v_{\bot }^{2}}{2\unicode[STIX]{x1D6FA}_{e}}\unicode[STIX]{x1D6FF}B_{\Vert }\right)+\frac{ef_{0e}}{T_{e}}\left(1-\frac{\unicode[STIX]{x1D714}_{\ast e}^{T}}{\unicode[STIX]{x1D714}}\right)\boldsymbol{v}_{de}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D6F9}.\qquad\end{eqnarray}$$
We employ a related ordering scheme to that of the ions to solve (3.24) for the electrons, introducing another small parameter,
$\unicode[STIX]{x1D700}_{e}$
, where
$\unicode[STIX]{x1D714}_{dre}\sim \unicode[STIX]{x1D708}_{e}\sim \unicode[STIX]{x1D700}_{e}\unicode[STIX]{x1D714}_{be}$
;
$\unicode[STIX]{x1D714}\sim \unicode[STIX]{x1D714}_{d\unicode[STIX]{x1D703}e}\sim \unicode[STIX]{x1D700}_{e}^{2}\unicode[STIX]{x1D714}_{be}$
. However, the semi-collisional electron model requires the additional ordering:
$k_{\Vert }v_{\text{the}}\sim \unicode[STIX]{x1D700}_{e}^{2}v_{\text{the}}/L_{c}$
, where
$k_{\Vert }=\ell q^{\prime }x/Rq$
. With
$h_{e}=h_{0e}+\unicode[STIX]{x1D700}_{e}h_{1e}+\unicode[STIX]{x1D700}_{e}^{2}h_{2e}+\cdots \,$
, the lowest-order equation is
while in first order we obtain
As in the case of the ions,
$h_{1e}$
can be annihilated for passing particles by applying the operation:
$\langle B(\ldots )/v_{\Vert }\rangle$
. For trapped particles, we again integrate along the bounce orbit, summing over
$\unicode[STIX]{x1D70E}$
in the usual way. Thus, we obtain the constraint
which determines
$h_{0e}$
, yielding:
where now
$u^{2}=m_{e}v^{2}/2T_{e}$
. The energy exchange term in the electron–ion collision operator is neglected (see (3.33) below), which makes the perturbation in the electron temperature independent of that in the ion temperature.
The equation for
$h_{1e}$
can then be integrated to give
where the function
$H_{1e}$
remains to be determined. This result can be rewritten as
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}h_{1e}={\displaystyle \frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{e}}}\left[A_{1}^{(e)}+\left(u^{2}-{\displaystyle \frac{5}{2}}\right)A_{2}^{(e)}\right]f_{0e}+H_{1e}(v,\unicode[STIX]{x1D706},x,\unicode[STIX]{x1D70E}),\\ A_{1}^{(e)}={\displaystyle \frac{e\unicode[STIX]{x1D6F9}^{\prime }}{T_{e}}}\left(1-{\displaystyle \frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}}(1+\unicode[STIX]{x1D702}_{e})\right)-\left({\displaystyle \frac{\hat{n}_{e}^{\prime }}{n_{0}}}+{\displaystyle \frac{\hat{T}_{e}^{\prime }}{T_{e}}}\right)\equiv -\left({\displaystyle \frac{\tilde{p}_{e}^{\prime }}{p_{0e}}}-{\displaystyle \frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{e}}}\right),\\ A_{2}^{(e)}=-\left({\displaystyle \frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}}{\displaystyle \frac{e\unicode[STIX]{x1D6F9}^{\prime }}{T_{e}}}+{\displaystyle \frac{\hat{T}_{e}^{\prime }}{T_{e}}}\right)\equiv -{\displaystyle \frac{\tilde{T}_{e}^{\prime }}{T_{e}}},\end{array}\right\}\end{eqnarray}$$
with
$p_{0e}=n_{0}T_{e}$
, where again we have used (2.13) and (2.14). The next-order equation is
leading to the constraint
which determines
$H_{1e}(v,\unicode[STIX]{x1D706},x,\unicode[STIX]{x1D70E})$
.
The electron collision operator takes the form
where
$C_{ee}(h)$
is the electron–electron collision operator,
$U_{\Vert i}$
is the ion bulk parallel velocity,
$\unicode[STIX]{x1D708}_{ei}(u)=\unicode[STIX]{x1D708}_{0e}/u^{3}$
with
$\unicode[STIX]{x1D708}_{0e}=n_{0}\text{e}^{4}\ln \unicode[STIX]{x1D6EC}(m_{e}/2T_{e})^{3/2}/(4\unicode[STIX]{x03C0}\unicode[STIX]{x1D700}_{0}^{2}m_{e}^{2})$
.
At this point we introduce the two Spitzer functions (Spitzer & Harm Reference Spitzer and Harm1953):
$h_{es}^{(1)}$
and
$h_{es}^{(2)}$
, where
so that the constraint equation becomes
with
$h_{es}=h_{es}^{(1)}+h_{es}^{(2)}$
. Now that the collision operator is acting on a localised (to the trapped and barely passing region) distribution function we can use a model pitch-angle scattering operator for the electron–electron collisions (Rutherford et al.
Reference Rutherford, Kovrizhnykh, Rosenbluth and Hinton1970).
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}C_{ee}(h)=\unicode[STIX]{x1D708}_{ee}(u)\left(L(h)+{\displaystyle \frac{m_{e}}{T_{e}}}v_{\Vert }U_{\Vert e}^{\ast }(h)f_{0e}\right)\\ U_{\Vert e}^{\ast }(h)=\left.\displaystyle \int \text{d}^{3}\boldsymbol{v}\unicode[STIX]{x1D708}_{ee}(u)v_{\Vert }h\right/\displaystyle \int \text{d}^{3}vf_{0e}\unicode[STIX]{x1D708}_{ee}(u)m_{e}v_{\Vert }^{2}/2T_{e},\\ \unicode[STIX]{x1D708}_{ee}(u)=\unicode[STIX]{x1D708}_{0e}{\displaystyle \frac{\unicode[STIX]{x1D711}(u)}{u^{3}}}.\end{array}\right\}\end{eqnarray}$$
To remove the ion flow we write
The functions
$h_{s}^{(1,2)}$
are given by Cohen et al. (Reference Cohen, Spitzer and McR. Routly1950) and Spitzer & Harm (Reference Spitzer and Harm1953):
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}h_{s}^{(1)}={\displaystyle \frac{v_{\Vert }}{R^{2}B}}S^{(1)}{\displaystyle \frac{D^{(1)}(u)}{u}}f_{0e},\quad h_{s}^{(2)}={\displaystyle \frac{v_{\Vert }}{R^{2}B}}S^{(2)}{\displaystyle \frac{D^{(2)}(u)}{u}}f_{0e},\\ S^{(1)}={\displaystyle \frac{4\text{i}\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}}{\displaystyle \frac{I\ell q^{\prime }x}{q}}\left({\displaystyle \frac{\hat{n}_{e}}{n_{0}}}+{\displaystyle \frac{\hat{T}_{e}}{T_{e}}}\right)\equiv {\displaystyle \frac{4\text{i}\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}}{\displaystyle \frac{I\ell q^{\prime }x}{q}}\left({\displaystyle \frac{\tilde{p}_{e}}{p_{0e}}}-{\displaystyle \frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}}-{\displaystyle \frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}}{\displaystyle \frac{e\unicode[STIX]{x1D6F9}}{T_{e}}}\right),\\ S^{(2)}={\displaystyle \frac{4i\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}}{\displaystyle \frac{I\ell q^{\prime }x}{q}}{\displaystyle \frac{\hat{T}_{e}}{T_{e}}}\equiv {\displaystyle \frac{4i\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}}{\displaystyle \frac{I\ell q^{\prime }x}{q}}\left({\displaystyle \frac{\tilde{T}_{e}}{T_{e}}}-{\displaystyle \frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}}{\displaystyle \frac{e\unicode[STIX]{x1D6F9}}{T_{e}}}\right),\end{array}\right\}\end{eqnarray}$$
where
$\unicode[STIX]{x1D70F}_{ei}=3\sqrt{\unicode[STIX]{x03C0}}/4\unicode[STIX]{x1D708}_{0e}$
and
$D^{(1)}$
and
$D^{(2)}$
are related to the normalised responses to a parallel electric field and thermal force, tabulated in Spitzer & Harm (Reference Spitzer and Harm1953) (to be precise,
$D^{(1)}=-D/A$
and
$D^{(2)}=2D/B$
as given in Spitzer & Harm (Reference Spitzer and Harm1953, Tables I and II)).
So, on defining
$\unicode[STIX]{x1D708}_{e}=\unicode[STIX]{x1D708}_{ee}+\unicode[STIX]{x1D708}_{ei}$
, the constraint equation becomes
where
$h_{s}=h_{s}^{(1)}+h_{s}^{(2)}$
. Thus
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D706}}\left(\unicode[STIX]{x1D706}\langle v_{\Vert }\rangle \frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D706}}H_{1e}\right)=\frac{IB_{0}v^{2}}{2\unicode[STIX]{x1D6FA}_{e}}\left(A_{1}^{(e)}+\left(u^{2}-\frac{5}{2}\right)A_{2}^{(e)}\right)f_{0e}\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{m_{e}v^{2}}{2T_{e}}\left\langle \frac{B}{\unicode[STIX]{x1D708}_{e}}(\unicode[STIX]{x1D708}_{ee}U_{\Vert e}^{\ast }+\unicode[STIX]{x1D708}_{ei}U_{\Vert i})f_{0e}\right\rangle -\frac{(S^{(1)}D^{(1)}(u)+S^{(2)}D^{(2)}(u))v^{2}}{2u}\left\langle \frac{1}{R^{2}}\right\rangle f_{0e}.\qquad\end{eqnarray}$$
Integrating (3.40) in
$\unicode[STIX]{x1D706}$
, we find
$$\begin{eqnarray}\displaystyle h_{1e} & = & \displaystyle \frac{I}{\unicode[STIX]{x1D6FA}_{e}}\left(v_{\Vert }-H\hat{V}_{\Vert }\frac{B}{B_{0}}\right)\left(A_{1}^{(e)}+\left(u^{2}-\frac{5}{2}\right)A_{2}^{(e)}\right)f_{0e}\nonumber\\ \displaystyle & & \displaystyle +\,\left[\frac{m_{e}}{T_{e}}\left\langle \frac{B}{B_{0}\unicode[STIX]{x1D708}_{e}}(\unicode[STIX]{x1D708}_{ee}U_{\Vert e}^{\ast }+\unicode[STIX]{x1D708}_{ei}U_{\Vert i})\right\rangle +\frac{(S^{(1)}D^{(1)}(u)+S^{(2)}D^{(2)}(u))}{B_{0}u}\left\langle \frac{1}{R^{2}}\right\rangle \right]H\hat{V}_{\Vert }\,f_{0e}.\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
Using the definition of
$U_{\Vert e}^{\ast }$
in (3.36) and integrating over velocity space, we can calculate
$U_{\Vert e}^{\ast }(h_{1e}-h_{s})$
self-consistently. However, as argued by Helander & Sigmar (Reference Helander and Sigmar2002),
$\langle BU_{\Vert e}^{\ast }\rangle =\langle BU_{\Vert i}\rangle +0(f_{t})$
, provided
$f_{t}\ll 0(\unicode[STIX]{x1D708}_{ei}/\unicode[STIX]{x1D708}_{e})$
. Thus result (3.41) simplifies to
$$\begin{eqnarray}\displaystyle h_{1e} & = & \displaystyle \frac{I}{\unicode[STIX]{x1D6FA}_{e}}\left(v_{\Vert }-H\hat{V}_{\Vert }\frac{B}{B_{0}}\right)\left(A_{1}^{(e)}+\left(u^{2}-\frac{5}{2}\right)A_{2}^{(e)}\right)f_{0e}\nonumber\\ \displaystyle & & \displaystyle +\,\left[\frac{m_{e}\langle BU_{\Vert i}\rangle }{T_{e}}+\frac{(S^{(1)}D^{(1)}(u)+S^{(2)}D^{(2)}(u))}{u}\left\langle \frac{1}{R^{2}}\right\rangle \right]\frac{H\hat{V}_{\Vert }}{B_{0}}f_{0e}.\end{eqnarray}$$
It remains to determine
$\hat{n}_{e}$
and
$\hat{T}_{e}$
; these are obtained in § 4.1 by applying the operators
$\langle \int (\ldots )\text{d}^{3}\boldsymbol{v}\rangle$
and
$\int (\ldots )(m_{e}v^{2}/2)\text{d}^{3}\boldsymbol{v}$
to the next-order equation:
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}h_{3e}+\frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{e}}\right)\frac{\unicode[STIX]{x2202}h_{2e}}{\unicode[STIX]{x2202}x}+\frac{\text{i}I\ell q^{\prime }x}{qR^{2}B}v_{\Vert }h_{1e}+\boldsymbol{v}_{d\unicode[STIX]{x1D703}e}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0e}\nonumber\\ \displaystyle & & \displaystyle \quad -\,\text{i}\unicode[STIX]{x1D714}h_{0e}-C_{e}(h_{2e})-\overline{L_{e}C_{e}(g_{0e}L_{e})}=\frac{\text{i}e}{T_{e}}f_{0e}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast e}^{T})\left(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9}+\frac{v_{\bot }^{2}}{2\unicode[STIX]{x1D6FA}_{e}}\unicode[STIX]{x1D6FF}B_{\Vert }\right).\qquad\end{eqnarray}$$
4 The perturbed density and temperatures
In this section we derive equations for the perturbed electron density and temperature and ion temperature in terms of the perturbed potentials
$\unicode[STIX]{x1D6F7}$
and
$\unicode[STIX]{x1D6F9}$
. The ion density perturbation can be obtained from the leading-order quasi-neutrality condition, see (2.12).
The other perturbed field that these quantities depend on is
$\unicode[STIX]{x1D6FF}B_{\Vert }$
, which is obtained from the perpendicular component of Ampère’s law. The perpendicular current can be calculated from the first order in a Larmor radius expansion of the distribution function (2.2), again recalling (2.8), (3.8) and (3.28). Introducing this result into the perpendicular component of Ampère’s law yields (Tang et al.
Reference Tang, Connor and Hastie1980)
In § 5 we will discuss the parallel component of Ampère’s law and find that in leading order,
$\unicode[STIX]{x1D6F9}$
is independent of the poloidal angle,
$\unicode[STIX]{x1D703}$
. Since
$\hat{n}_{j}$
and
$\hat{T}_{j}$
are also independent of
$\unicode[STIX]{x1D703}$
, it follows from (2.13) and (2.14) that
${\tilde{n}}_{j}$
,
$\tilde{T}_{j}$
and
$\unicode[STIX]{x1D6F7}$
are also independent of
$\unicode[STIX]{x1D703}$
.
4.1 The perturbed electron density and temperature
As mentioned earlier, the determination of
$\hat{n}_{e}$
and
$\hat{T}_{e}$
is accomplished by applying the annihilators
$\langle \int (\ldots )\text{d}^{3}\boldsymbol{v}\rangle$
and
$\int (\ldots )(m_{e}v^{2}/2)\text{d}^{3}\boldsymbol{v}$
to the third-order equation (3.43), which both eliminate
$h_{3e}$
.
The first operation results in
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{e}}\right)\frac{\unicode[STIX]{x2202}h_{2e}}{\unicode[STIX]{x2202}x}\right\rangle +\left\langle \frac{\text{i}\ell Iq^{\prime }x}{R^{2}Bq}\int \,\text{d}^{3}\boldsymbol{v}v_{\Vert }h_{1e}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \qquad +\,\left\langle \int \,\text{d}^{3}\boldsymbol{v}\boldsymbol{v}_{d\unicode[STIX]{x1D703}e}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0e}\right\rangle -\left\langle \int \,\text{d}^{3}\boldsymbol{v}\overline{L_{e}C_{e}(g_{0e}L_{e})}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\text{i}\unicode[STIX]{x1D714}\hat{n}_{e}+\text{i}n_{0}\left\langle \frac{e}{T_{e}}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast e})(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})-\frac{\unicode[STIX]{x1D6FF}B_{\Vert }}{B}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast e}(1+\unicode[STIX]{x1D702}_{e}))\right\rangle .\end{eqnarray}$$
Here one can recognise the first three terms on the left-hand side as representing, respectively, the surface-averaged contributions to the electron continuity equation of the divergences of the radial flux,
$\unicode[STIX]{x1D6E4}_{re}$
(which we will see below is related to Pfirsch–Schlüter and neoclassical radial transport), the parallel flux,
$\unicode[STIX]{x1D6E4}_{\Vert e}$
, and the ‘azimuthal flux’,
$\unicode[STIX]{x1D6E4}_{e}^{Az}$
(Fitzpatrick Reference Fitzpatrick1989), while the fourth corresponds to the contribution from classical radial transport.
The second, the energy moment, takes the form
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{e}}\right)\frac{m_{e}v^{2}}{2}\frac{\unicode[STIX]{x2202}h_{2e}}{\unicode[STIX]{x2202}x}\right\rangle +\left\langle \frac{\text{i}\ell Iq^{\prime }x}{R^{2}Bq}\int \int \,\text{d}^{3}\boldsymbol{v}v_{\Vert }\frac{m_{e}v^{2}}{2}h_{1e}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \qquad +\,\left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{m_{e}v^{2}}{2}\boldsymbol{v}_{d\unicode[STIX]{x1D703}e}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0e}\right\rangle -\left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{m_{e}v^{2}}{2}\overline{L_{e}C_{e}(g_{0e}L_{e})}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\frac{3}{2}\text{i}\unicode[STIX]{x1D714}\left(\hat{n}_{e}+\frac{n_{0}\hat{T}_{e}}{T_{e}}\right)+\text{i}n_{0}\left\langle \frac{3}{2}\frac{e}{T_{e}}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast e}(1+\unicode[STIX]{x1D702}_{e}))(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})\right.\nonumber\\ \displaystyle & & \displaystyle \qquad -\,\left.\frac{5}{2}\frac{\unicode[STIX]{x1D6FF}B_{\Vert }}{B}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast e}(1+2\unicode[STIX]{x1D702}_{e}))\right\rangle ,\end{eqnarray}$$
with a similar interpretation in terms of fluxes.
It is helpful to separate the electron and ion distributions into a Pfirsch–Schlüter-like,
$h_{e}^{PS}$
, and a banana neoclassical part,
$h_{e}^{\text{Ban}}$
, and express the quantities
$\hat{n}_{j}$
and
$\hat{T}_{j}$
in terms of
${\tilde{n}}_{j}$
and
$\tilde{T}_{j}$
. Thus, for the electrons we write
where
$$\begin{eqnarray}\displaystyle h_{e}^{PS} & = & \displaystyle \frac{Iv_{\Vert }m_{e}}{eB}\left(1-\frac{B^{2}}{\langle B^{2}\rangle }\right)\left[\frac{\tilde{p}^{\prime }}{p_{0e}}+\left(u^{2}-\frac{5}{2}\right)\frac{\tilde{T}_{e}^{\prime }}{T_{e}}\right]f_{0e}+\frac{Iv_{\Vert }m_{e}}{eB}\left(1-\frac{B^{2}}{\langle B^{2}\rangle }\right)\nonumber\\ \displaystyle & & \displaystyle \times \,\left[\frac{\tilde{p}_{i}^{\prime }}{p_{0e}}+\frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{e}}\right]f_{0e}-\frac{4\text{i}\unicode[STIX]{x1D70F}_{ei}v_{\Vert }}{3\sqrt{\unicode[STIX]{x03C0}}}\frac{I\ell q^{\prime }x}{q}\left(\frac{1}{R^{2}B}-\frac{B}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\nonumber\\ \displaystyle & & \displaystyle \times \,\left\{\frac{D^{(1)}(u)}{u}\left[\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}(1+\unicode[STIX]{x1D702}_{e})}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right]\right.\nonumber\\ \displaystyle & & \displaystyle +\,\left.\frac{D^{(2)}(u)}{u}\left[\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right]\right\}f_{0e}\end{eqnarray}$$
and
$$\begin{eqnarray}\displaystyle h_{e}^{\text{Ban}} & = & \displaystyle \frac{Im_{e}}{e}\left(\frac{B}{\langle B^{2}\rangle }v_{\Vert }-\frac{H\hat{V}_{\Vert }}{B_{0}}\right)\left[\frac{\tilde{p}^{\prime }}{p_{0e}}+\left(u^{2}-\frac{5}{2}\right)\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-1.17\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]f_{0e}\nonumber\\ \displaystyle & & \displaystyle -\,\frac{4\text{i}\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}\frac{I\ell q^{\prime }x}{q}\left(\frac{B}{\langle B^{2}\rangle }v_{\Vert }-\frac{H\hat{V}_{\Vert }}{B_{0}}\right)\left\langle \frac{1}{R^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \times \,\left\{\frac{D^{(1)}(u)}{u}\left[\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}(1+\unicode[STIX]{x1D702}_{e})}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right]\right.\nonumber\\ \displaystyle & & \displaystyle +\,\left.\frac{D^{(2)}(u)}{u}\left[\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right]\right\}f_{0e},\end{eqnarray}$$
where we have substituted for
$U_{\Vert i}$
from (3.22).
Similarly, it is convenient to separate
$h_{1i}$
in (3.14) into Pfirsch–Schlüter,
$h_{i}^{PS}$
, and banana contributions,
$h_{i}^{\text{Ban}}$
, writing:
where
$h_{i}^{\text{Flow}}$
describes the mean flow:
while
and
The first term represents a drifting Maxwellian and therefore does not contribute to the ion–ion collision term.
4.1.1 Electron radial fluxes
First, we consider the radial fluxes arising from classical transport. Recalling the definitions of
$h_{0j}$
and
$L_{j}$
and using the conservation of momentum in like-particle collisions, we obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\overline{L_{e}C_{e}(g_{0e}L_{e})}\right\rangle \equiv \frac{\text{d}}{\text{d}r}\langle \unicode[STIX]{x1D6E4}_{r,e}^{C}\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\frac{m_{e}T_{e}}{e^{2}}\frac{\text{d}^{2}}{\text{d}x^{2}}\left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{B^{2}}\right\rangle \int \,\text{d}^{3}\boldsymbol{v}\frac{m_{e}v_{r}}{T_{e}}C_{ei}\left[v_{r}\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e\unicode[STIX]{x1D6F7}}{T_{e}}+\left(u^{2}-\frac{5}{2}\right)\frac{\tilde{T}_{e}}{T_{e}}\right)\right],\qquad\end{eqnarray}$$
where
$r$
represents a ‘radial-like’ coordinate labelling flux surfaces,
$v_{r}$
being the corresponding radial component of the velocity. We note the gyro-correction to the scattering target ion distribution has the form
The velocity integral can be evaluated by observing that it involves the matrix elements of the collision operator between Laguerre polynomials given by Helander & Sigmar (Reference Helander and Sigmar2002, § 4.5). The resultant contribution from (4.11) is the familiar Braginskii expression:
A similar calculation for the classical heat flux results in
so that the classical energy flux,
$\langle Q_{r,e}^{C}\rangle =\langle q_{r,e}^{C}\rangle +(5/2)\langle \unicode[STIX]{x1D6E4}_{r,e}^{C}\rangle$
, satisfies
Turning to the leading term in (4.2), we can eliminate
$h_{2e}$
by integrating by parts in
$\unicode[STIX]{x1D703}$
and using (3.31) to obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{e}}\right)h_{2e}\right\rangle \equiv \langle \unicode[STIX]{x1D6E4}_{r,e}^{NC}\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\left\langle \frac{I}{\unicode[STIX]{x1D6FA}_{e}}\int \,\text{d}^{3}\boldsymbol{v}\left[\frac{v_{\Vert }^{2}I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{I}{\unicode[STIX]{x1D6FA}_{e}}\right)\frac{\text{d}h_{1e}}{\text{d}x}+\frac{\text{i}\ell Iq^{\prime }x}{R^{2}Bq}v_{\Vert }^{2}h_{0e}-v_{\Vert }C_{e}(h_{1e})\right]\right\rangle ,\end{eqnarray}$$
where the first term vanishes if the equilibrium is up–down symmetric, which we take to be the case here. Thus, finally, we recognise this contribution as the divergence of the neoclassical radial flux
Likewise, for the thermal equation, we find
To evaluate the velocity integrals in (4.17) and (4.18), we recall the result (3.19).
Considering the Pfirsch–Schlüter and banana contributions separately, we introduce their respective distribution functions. From momentum conservation, it is again clear that only
$e{-}i$
collisions contribute to (4.17) in both cases. For the Pfirsch–Schlüter contribution the calculation has the same structure as the classical case, with the substitution
$v_{r}\rightarrow v_{\Vert }$
in the distribution function. However, since we are now considering the difference
$h_{1e}-h_{es}$
, the scattering ‘ion distribution’ is effectively at rest when evaluating the collisional matrix elements. Recalling (4.5), the result is
$$\begin{eqnarray}\displaystyle \frac{1}{n_{0}}\frac{\text{d}}{\text{d}r}\langle \unicode[STIX]{x1D6E4}_{r,e}^{PS}\rangle & = & \displaystyle -\frac{m_{e}T_{e}}{\text{e}^{2}\unicode[STIX]{x1D70F}_{ei}}I^{2}\left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}}{p_{0e}}-\frac{3}{2}\frac{\tilde{T}_{e}}{T_{e}}\right)\nonumber\\ \displaystyle & & \displaystyle -\,\frac{\text{i}I^{2}\ell q^{\prime }T_{e}}{eq}\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\frac{\text{d}}{\text{d}x}\nonumber\\ \displaystyle & & \displaystyle \times \,\left[x\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e\unicode[STIX]{x1D6F7}}{T_{e}}+\left(1-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\right)\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right].\end{eqnarray}$$
For the heat flux, we obtain
$$\begin{eqnarray}\displaystyle \frac{1}{n_{0}T_{e}}\frac{\text{d}}{\text{d}r}\langle Q_{r,e}^{PS}\rangle & = & \displaystyle -\frac{m_{e}T_{e}I^{2}}{\text{e}^{2}\unicode[STIX]{x1D70F}_{ei}}\left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}}{p_{0e}}+0.91\frac{\tilde{T}_{e}}{T_{e}}\right)\nonumber\\ \displaystyle & & \displaystyle -\,\frac{\text{i}I^{2}\ell q^{\prime }T_{e}}{eq}\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\frac{\text{d}}{\text{d}x}\nonumber\\ \displaystyle & & \displaystyle \times \,\left[\frac{5}{2}x\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e\unicode[STIX]{x1D6F7}}{T_{e}}+\frac{\tilde{T}_{e}}{T_{e}}+\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\left(1-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{e})\right)\right)\right].\qquad\end{eqnarray}$$
Finally, we consider the banana contribution. For the classical and Pfirsch–Schlüter contributions we used the exact collision operator, since the distributions were not localised in velocity space. However, the banana contribution to
$h_{1e}-h_{ss}$
is localised and it is sufficient to use the model collision operator (3.36), which leads to:
Substituting expression (4.6), using result (3.22) and approximating
$\langle BU_{\Vert e}^{\ast }\rangle \cong \langle BU_{\Vert i}\rangle$
as discussed below (3.41), we obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1}{n_{0}}\frac{\text{d}}{\text{d}r}\langle \unicode[STIX]{x1D6E4}_{r,e}^{\text{Ban}}\rangle =-f_{t}\frac{I^{2}m_{e}T_{e}\{\unicode[STIX]{x1D708}_{e}\}}{\text{e}^{2}\langle B^{2}\rangle }\frac{\text{d}^{2}}{\text{d}x^{2}}\left[\frac{\tilde{p}}{p_{0e}}+\left(\frac{\{u^{2}\unicode[STIX]{x1D708}_{e}\}}{\{\unicode[STIX]{x1D708}_{e}\}}-\frac{5}{2}\right)\frac{\tilde{T}_{e}}{T_{e}}-1.17\frac{\tilde{T}_{i}}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{4\unicode[STIX]{x1D70F}_{ei}\text{i}\ell q^{\prime }}{3\sqrt{\unicode[STIX]{x03C0}}q}\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{T_{e}}{e}\frac{\text{d}}{\text{d}x}\!\left(x\left[\!\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\!\left\{\frac{\unicode[STIX]{x1D708}_{e}D^{(1)}}{u}\right\}\right.\right.\nonumber\\ \displaystyle & & \displaystyle \quad +\,\left.\left.\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\left\{\frac{\unicode[STIX]{x1D708}_{e}D^{(2)}}{u}\right\}\right]\right).\end{eqnarray}$$
Introducing the numerical values of the various collisional averages given by the integrals in appendix A, this becomes:
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1}{n_{0}}\frac{\text{d}}{\text{d}r}\langle \unicode[STIX]{x1D6E4}_{r,e}^{\text{Ban}}\rangle =-1.53f_{t}\frac{I^{2}m_{e}T_{e}}{\text{e}^{2}\langle B^{2}\rangle \unicode[STIX]{x1D70F}_{ei}}\frac{\text{d}^{2}}{\text{d}x^{2}}\left[\frac{\tilde{p}}{p_{0e}}+1.39\frac{\tilde{T}_{e}}{T_{e}}-1.17\frac{\tilde{T}_{i}}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,1.67\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{i}\ell q^{\prime }T_{e}}{eq}\frac{\text{d}}{\text{d}x}\left(x\left[\vphantom{\frac{\tilde{T}_{e}}{T_{e}}}\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+1.02\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right.\right.\nonumber\\ \displaystyle & & \displaystyle \quad +\,\left.\left.0.02\frac{\tilde{T}_{e}}{T_{e}}\right]\right).\end{eqnarray}$$
Similarly, we find
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1}{n_{0}T_{e}}\frac{\text{d}}{\text{d}r}\langle Q_{r,e}^{\text{Ban}}\rangle =-f_{t}\frac{I^{2}m_{e}T_{e}\{u^{2}\unicode[STIX]{x1D708}_{e}\}}{\text{e}^{2}\langle B^{2}\rangle }\frac{\text{d}^{2}}{\text{d}x^{2}}\!\!\left[\frac{\tilde{p}}{p_{0e}}+\left(\frac{\{u^{4}\unicode[STIX]{x1D708}_{e}\}}{\{u^{2}\unicode[STIX]{x1D708}_{e}\}}-\frac{5}{2}\right)\frac{\tilde{T}_{e}}{T_{e}}+\left(\frac{\{u^{2}\unicode[STIX]{x1D708}_{ii}\}}{\{\unicode[STIX]{x1D708}_{ii}\}}-\frac{5}{2}\right)\frac{\tilde{T}_{i}}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{4\unicode[STIX]{x1D70F}_{ei}\text{i}\ell q^{\prime }}{3\sqrt{\unicode[STIX]{x03C0}}q}\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{T_{e}}{e}\frac{\text{d}}{\text{d}x}\left(x\left[\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\{u\unicode[STIX]{x1D708}_{e}D^{(1)}\}\right.\right.\nonumber\\ \displaystyle & & \displaystyle \quad +\,\left.\left.\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\{u\unicode[STIX]{x1D708}_{e}D^{(2)}\}\right]\right),\end{eqnarray}$$
leading to
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1}{n_{0}T_{e}}\frac{\text{d}}{\text{d}r}\langle Q_{r,e}^{\text{Ban}}\rangle =-1.70f_{t}\frac{I^{2}m_{e}T_{e}}{\text{e}^{2}\langle B^{2}\rangle \unicode[STIX]{x1D70F}_{ei}}\frac{\text{d}^{2}}{\text{d}x^{2}}\left[\frac{\tilde{p}}{p_{0e}}-0.40\frac{\tilde{T}_{e}}{T_{e}}-1.17\frac{\tilde{T}_{i}}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,2.97\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{i\ell q^{\prime }T_{e}}{eq}\frac{\text{d}}{\text{d}x}\left(x\left[\vphantom{\frac{\tilde{T}_{e}}{T_{e}}}\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+1.25\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right.\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.\left.+\,0.25\frac{\tilde{T}_{e}}{T_{e}}\right]\right).\end{eqnarray}$$
4.1.2 Parallel electron fluxes and plasma current
The contribution to (4.2) from the parallel flux
$\unicode[STIX]{x1D6E4}_{\Vert e}$
is
The part arising from the Pfirsch–Schlüter-like part of
$h_{1e}$
can be calculated directly, but for the banana contribution we take advantage of the self-adjointness of the collision operator to circumvent the fact that
$h_{1e}$
is not localised in pitch angle, while still allowing us to use the model collision operator (Helander & Sigmar Reference Helander and Sigmar2002).
We first consider the parallel current, which is needed for Ampère’s equation, but also provides an expression for
$U_{\Vert e}$
. We can readily calculate
$j_{\Vert }^{PS}$
, the current arising from the Pfirsch–Schlüter-like contributions to the electron and ion distribution functions given in (4.5) and (4.9), obtaining:
$$\begin{eqnarray}\displaystyle j_{\Vert }^{PS} & = & \displaystyle -I\tilde{p}^{\prime }\left(\frac{1}{B}-\frac{B}{\langle B^{2}\rangle }\right)+\frac{4\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}\frac{\text{i}I\ell q^{\prime }xp_{0e}e}{m_{e}q}\nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{1}{R^{2}B}-\frac{B}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\left\{\frac{D^{(1)}}{u}\right\}\right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\left\{\frac{D^{(2)}}{u}\right\}\right],\end{eqnarray}$$
which becomes
$$\begin{eqnarray}\displaystyle j_{\Vert }^{PS} & = & \displaystyle -I\tilde{p}^{\prime }\left(\frac{1}{B}-\frac{B}{\langle B^{2}\rangle }\right)-1.97\unicode[STIX]{x1D70F}_{ei}\frac{\text{i}I\ell q^{\prime }xp_{0e}e}{m_{e}q}\nonumber\\ \displaystyle & & \displaystyle \times \left(\frac{1}{R^{2}B}-\frac{B}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \!\right)\!\left[\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}+\frac{0.34\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1\!+\!1.34\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\!\right]\!.\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
To calculate the banana regime neoclassical contribution, we write
where
so that
$$\begin{eqnarray}\displaystyle j_{\Vert s} & = & \displaystyle -\frac{4\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}n_{0}e\frac{\text{i}\ell Iq^{\prime }x}{R^{2}Bq}\frac{T_{e}}{m_{e}}\nonumber\\ \displaystyle & & \displaystyle \times \,\left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}\left\{\frac{D^{(1)}(u)}{u}\right\}\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,\left\{\frac{D^{(2)}(u)}{u}\right\}\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right],\end{eqnarray}$$
which reduces to
Now, recalling (3.34), we can write
Using the self-adjointness property of
$C_{e}(h_{s}^{(1)})$
, we obtain
so that
$$\begin{eqnarray}\displaystyle (j_{\Vert }-j_{\Vert s}) & = & \displaystyle -\frac{4\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}e\int \,\text{d}^{3}\boldsymbol{v}\frac{D^{(1)}(u)}{u}v_{\Vert }C_{e}(h_{1e}-h_{es}^{(1)}-h_{es}^{(2)})\nonumber\\ \displaystyle & = & \displaystyle \frac{4\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}e\int \,\text{d}^{3}\boldsymbol{v}\frac{D^{(1)}(u)}{u}v_{\Vert }\unicode[STIX]{x1D708}_{e}\left(h_{1e}-h_{s}^{(1)}-h_{s}^{(2)}-\frac{m_{e}v_{\Vert }}{T_{e}}U_{\Vert i}\,f_{0e}\right),\end{eqnarray}$$
where we have used the definitions (3.38) and the model electron–electron collision operator (3.36), since
$(h_{1e}-h_{es}^{(1)}-h_{es}^{(2)})$
is now localised. Evaluating this expression using result (4.6)
$$\begin{eqnarray}\displaystyle (j_{\Vert }-j_{\Vert s})^{\text{Ban}} & = & \displaystyle \frac{4\unicode[STIX]{x1D70F}_{ei}n_{0}}{3\sqrt{\unicode[STIX]{x03C0}}}\frac{f_{t}BIT_{e}}{\langle B^{2}\rangle }\left[\frac{\tilde{p}^{\prime }}{p_{0e}}\left\{\frac{\unicode[STIX]{x1D708}_{e}D^{(1)}}{u}\right\}-\left\{\left(\frac{5}{2}-u^{2}\right)\frac{\unicode[STIX]{x1D708}_{e}D^{(1)}}{u}\right\}\frac{\tilde{T}_{e}^{\prime }}{T_{e}}\right.\nonumber\\ \displaystyle & & \displaystyle \left.-\,\left(\frac{5}{2}-\frac{\{u^{2}\unicode[STIX]{x1D708}_{ii}\}}{\{\unicode[STIX]{x1D708}_{ii}\}}\right)\left\{\frac{\unicode[STIX]{x1D708}_{e}D^{(1)}}{u}\right\}\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]-\left(\frac{4\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}\right)^{2}\frac{f_{t}B\text{i}I\ell q^{\prime }xn_{0}}{\langle B^{2}\rangle q}\left\langle \frac{1}{R^{2}}\right\rangle \frac{eT_{e}}{m_{e}}\nonumber\\ \displaystyle & & \displaystyle \times \,\left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\left\{\frac{\unicode[STIX]{x1D708}_{e}D^{(1)2}}{u^{2}}\right\}\right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\left\{\frac{\unicode[STIX]{x1D708}_{e}D^{(1)}D^{(2)}}{u^{2}}\right\}\right],\end{eqnarray}$$
leading to
$$\begin{eqnarray}\displaystyle & & \displaystyle (j_{\Vert }-j_{\Vert s})^{\text{Ban}}=\frac{f_{t}BIp_{0e}}{\langle B^{2}\rangle }\left[-1.67\frac{\tilde{p}^{\prime }}{p_{0e}}+1.19\frac{\tilde{T}_{e}^{\prime }}{T_{e}}+1.95\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,2.58\frac{f_{t}B\text{i}I\ell q^{\prime }xp_{0e}e\,\unicode[STIX]{x1D70F}_{ei}}{\langle B^{2}\rangle qm_{e}}\left\langle \frac{1}{R^{2}}\right\rangle \left[\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1\!+\!1.17\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}+0.17\frac{\tilde{T}_{e}}{T_{e}}\right].\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
Combining expressions (4.28), (4.32) and (4.37), we obtain
$$\begin{eqnarray}\displaystyle j_{\Vert } & = & \displaystyle -I\tilde{p}^{\prime }\left(\frac{1}{B}-\frac{B}{\langle B^{2}\rangle }\right)+1.97\unicode[STIX]{x1D70F}_{ei}\frac{\text{i}I\ell q^{\prime }xp_{0e}e}{m_{e}q}\frac{B}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \times \,\left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}(1-1.31f_{t})\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)+0.34(1-0.69f_{t})\right.\nonumber\\ \displaystyle & & \displaystyle \left.\times \,\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right]-\frac{f_{t}BIp_{0e}}{\langle B^{2}\rangle }\left[1.67\frac{\tilde{p}^{\prime }}{p_{0e}}-1.19\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-1.95\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right].\end{eqnarray}$$
The first term is the usual Pfirsch–Schlüter current while the second term represents the effects of the parallel electron pressure gradient, the parallel electric field and parallel thermal force, whereas the final term is the bootstrap current.
We can now calculate
$U_{\Vert e}$
from
with
$U_{\Vert i}$
given by (3.22). The contribution to the electron continuity equation from the divergence of the parallel flux (4.26) is then:
$$\begin{eqnarray}\displaystyle & & \displaystyle n_{0}\left\langle \frac{\text{i}\ell Iq^{\prime }x}{R^{2}Bq}U_{\Vert e}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\frac{\text{i}I^{2}\ell q^{\prime }x}{eq}\tilde{p}^{\prime }\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)+\frac{\text{i}I^{2}\ell q^{\prime }xp_{0e}f_{t}}{e\langle B^{2}\rangle q}\left\langle \frac{1}{R^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \qquad \times \,\left[1.67\frac{\tilde{p}^{\prime }}{p_{0e}}-1.19\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-3.12\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \qquad -\,\frac{\text{i}I^{2}\ell q^{\prime }xp_{0e}}{eq}\left[\left\langle \frac{1}{R^{2}B^{2}}\right\rangle \left(\frac{p_{i}^{\prime }}{p_{0e}}+\frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{e}}\right)-\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \qquad -\,\frac{1.97\,\unicode[STIX]{x1D70F}_{ei}(I\ell q^{\prime }x)^{2}p_{0e}}{m_{e}\langle B^{2}\rangle q^{2}}\left\langle \frac{1}{R^{2}}\right\rangle ^{2}\nonumber\\ \displaystyle & & \displaystyle \qquad \times \,\left[(1-1.31f_{t})\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)+0.34(1-0.69f_{t})\right.\nonumber\\ \displaystyle & & \displaystyle \qquad \left.\times \,\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right].\end{eqnarray}$$
The energy equation involves the parallel heat flux
We can decompose
$Q_{\Vert e}$
as
The Spitzer contribution is given as
$$\begin{eqnarray}\displaystyle Q_{\Vert se} & = & \displaystyle \frac{5}{2}n_{0}T_{e}U_{\Vert i}+\frac{4\unicode[STIX]{x1D70F}_{ei}}{3\sqrt{\unicode[STIX]{x03C0}}}n_{0}T_{e}\frac{\text{i}\ell Iq^{\prime }x}{R^{2}Bq}\frac{T_{e}}{m_{e}}\nonumber\\ \displaystyle & & \displaystyle \times \left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}\{uD^{(1)}(u)\}\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,\{uD^{(2)}(u)\}\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right],\end{eqnarray}$$
resulting in
$$\begin{eqnarray}\displaystyle Q_{\Vert se} & = & \displaystyle -\frac{5}{2}n_{0}T_{e}\frac{IT_{i}}{eB}\left[\left(\frac{p_{i}^{\prime }}{p_{0e}}+\frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{e}}\right)-\frac{1.17B^{2}}{\langle B^{2}\rangle }(1-f_{t})\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle -\,6.26n_{0}T_{e}\frac{\text{i}\ell Iq^{\prime }x\unicode[STIX]{x1D70F}_{ei}}{R^{2}Bq}\frac{T_{e}}{m_{e}}\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}+0.58\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+1.58\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right).\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
Introducing the Pfirsch–Schlüter part of
$h_{1e}$
, equation (4.5), we can calculate the corresponding Pfirsch–Schlüter contribution to
$Q_{\Vert e}$
directly, obtaining:
$$\begin{eqnarray}\displaystyle \frac{Q_{\Vert e}^{PS}}{p_{0e}} & = & \displaystyle \frac{5}{2}\frac{IT_{e}}{e}\left(\frac{1}{B}-\frac{B}{\langle B^{2}\rangle }\right)\left(\frac{\tilde{p}^{\prime }}{p_{0e}}+\frac{\tilde{T}_{e}^{\prime }}{T_{e}}\right)+6.26n_{0}T\unicode[STIX]{x1D70F}_{ei}\frac{\text{i}I\ell q^{\prime }xT_{e}}{m_{e}q}\left(\frac{1}{R^{2}B}-\frac{B}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\nonumber\\ \displaystyle & & \displaystyle \times \,\left[\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}+\frac{0.58\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+1.58\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right].\end{eqnarray}$$
We formulate the banana neoclassical contribution for
$q_{\Vert e}$
by analogy to the neoclassical current,
$j_{\Vert e}$
, as in (4.35), so that it can also be evaluated using the localised distribution function (4.6):
Using expression (4.6), we obtain the banana contribution to
$q_{\Vert e}$
:
$$\begin{eqnarray}\displaystyle & & \displaystyle (q_{\Vert e}-q_{\Vert se})^{\text{Ban}}=\frac{f_{t}BIT_{e}p_{0e}}{e\langle B^{2}\rangle }\left[0.04\frac{\tilde{p}^{\prime }}{p_{0e}}+0.64\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-0.05\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,0.46\frac{f_{t}\text{i}I\ell q^{\prime }xT_{e}p_{0e}\unicode[STIX]{x1D70F}_{ei}}{m_{e}q}\frac{B}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \left[\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+1.94\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.94\frac{\tilde{T}_{e}}{T_{e}}\right].\end{eqnarray}$$
Inserting the results (4.37), (4.44), (4.45) and (4.47) into (4.42), we finally obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1}{n_{0}T_{e}}\left\langle \frac{\text{i}\ell Iq^{\prime }x}{R^{2}Bq}Q_{\Vert e}\right\rangle =\frac{5}{2}\frac{\text{i}\ell I^{2}q^{\prime }xT_{e}}{eq}\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\left(\frac{\tilde{p}^{\prime }}{p_{0e}}+\frac{\tilde{T}_{e}^{\prime }}{T_{e}}\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \frac{\text{i}\ell I^{2}q^{\prime }xT_{e}}{eq}\left[4.18\frac{\tilde{p}^{\prime }}{p_{0e}}-2.36\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-7.86\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]-\frac{5}{2}\frac{\text{i}\ell I^{2}q^{\prime }xT_{i}}{eq}\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left[\left\langle \frac{1}{R^{2}B^{2}}\right\rangle \left(\frac{p_{i}^{\prime }}{p_{0e}}+\frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{e}}\right)-\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]+6.26\frac{(\ell Iq^{\prime }x)^{2}}{\langle B^{2}\rangle q^{2}}\left\langle \frac{1}{R^{2}}\right\rangle ^{2}\frac{T_{e}\unicode[STIX]{x1D70F}_{ei}}{m_{e}}\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D702}_{e}\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}(1-1.11f_{t})\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad +\,\left.0.58(1-0.31f_{t})\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D702}_{e}\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right].\end{eqnarray}$$
4.1.3 Azimuthal electron fluxes
Finally, we require the contributions from the azimuthal drift terms that appear in (4.2) and (4.3).
$$\begin{eqnarray}\displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\boldsymbol{v}_{de}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0e}\right\rangle & = & \displaystyle -\text{i}\left\langle \int \int \,\text{d}^{3}\boldsymbol{v}\boldsymbol{v}_{de}\boldsymbol{\cdot }(m\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}-\ell \unicode[STIX]{x1D735}\unicode[STIX]{x1D711})h_{0e}\right\rangle \nonumber\\ \displaystyle & = & \displaystyle -\text{i}\left\langle \frac{\boldsymbol{n}}{\unicode[STIX]{x1D6FA}_{e}}\times (\unicode[STIX]{x1D735}\ln B+\unicode[STIX]{x1D73F})\boldsymbol{\cdot }(m\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}-\ell \unicode[STIX]{x1D735}\unicode[STIX]{x1D711})\right\rangle \frac{T_{e}}{m_{e}}\left(\frac{\hat{n}_{e}}{n_{0}}+\frac{\hat{T}_{e}}{T_{e}}\right)n_{0}\nonumber\\ \displaystyle & = & \displaystyle -\text{i}\left\langle \frac{\boldsymbol{n}}{\unicode[STIX]{x1D6FA}_{e}}\times \frac{\unicode[STIX]{x1D735}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})}{B^{2}}\boldsymbol{\cdot }(m\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}-\ell \unicode[STIX]{x1D735}\unicode[STIX]{x1D711})\right\rangle \frac{T_{e}}{m_{e}}\left(\frac{\hat{n}_{e}}{n_{0}}+\frac{\hat{T}_{e}}{T_{e}}\right)n_{0}.\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
This can be expressed as
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\boldsymbol{v}_{de}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0e}\right\rangle =\text{i}\ell \frac{T_{e}}{e}\left\langle \left(\frac{1}{B^{2}}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})+\left(\frac{1}{RB^{2}}\right)^{2}\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}\frac{\unicode[STIX]{x2202}B^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\right)\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)n_{0}.\end{eqnarray}$$
Similarly, we also obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{m_{e}v^{2}}{2}\boldsymbol{v}_{de}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0e}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\frac{5}{2}\text{i}\ell \frac{T_{e}}{e}\left\langle \left(\frac{1}{B^{2}}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})+\left(\frac{1}{RB^{2}}\right)^{2}\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}\frac{\unicode[STIX]{x2202}B^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\right)\right\rangle \nonumber\\ \displaystyle & & \displaystyle \qquad \times \,\left(\frac{\tilde{p}_{e}}{p_{0e}}+\frac{\tilde{T}_{e}}{T_{e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)n_{0}T_{e}.\end{eqnarray}$$
4.2 The perturbed ion temperature
In the case of the ions there is no need to obtain the ion density equation for
$\hat{n}_{i}$
as it is determined by quasi-neutrality and given in (2.13). To obtain the equation for
$\hat{T}_{i}$
we apply the operator
$\int (\ldots )(m_{i}v^{2}/2)\,\text{d}^{3}\boldsymbol{v}$
to (3.23).
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{v_{\Vert }I}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{Iv_{\Vert }}{\unicode[STIX]{x1D6FA}_{e}}\right)\frac{m_{i}v^{2}}{2}\frac{\unicode[STIX]{x2202}h_{2i}}{\unicode[STIX]{x2202}x}\right\rangle +\left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{m_{i}v^{2}}{2}\boldsymbol{v}_{d\unicode[STIX]{x1D703}i}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0i}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad -\,\left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{m_{i}v^{2}}{2}\overline{L_{i}C_{ii}(g_{i}L_{i})}\right\rangle =\frac{3}{2}\text{i}\unicode[STIX]{x1D714}\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{e\unicode[STIX]{x1D6F7}}{T_{i}}-\frac{e\unicode[STIX]{x1D6F9}}{T_{i}}\left(1+\frac{T_{i}}{T_{e}}\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{i})\right)\right)\nonumber\\ \displaystyle & & \displaystyle \quad -\,\text{i}n_{0}\left\langle \frac{3}{2}\frac{e}{T_{i}}\left(\unicode[STIX]{x1D714}+\frac{T_{i}}{T_{e}}\unicode[STIX]{x1D714}_{\ast e}(1+\unicode[STIX]{x1D702}_{i})\right)(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})+\frac{5}{2}\frac{\unicode[STIX]{x1D6FF}B_{\Vert }}{B}\left(\unicode[STIX]{x1D714}+\frac{T_{i}}{T_{e}}\unicode[STIX]{x1D714}_{\ast e}(1+2\unicode[STIX]{x1D702}_{i})\right)\right\rangle .\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
Integrating the first term on the left-hand side by parts and substituting for
$h_{2i}$
from (3.11), it can be written as the divergence of the neoclassical ion heat flux
The second term is the ion azimuthal flux and the third is the ion classical heat flux:
This can be evaluated in a similar manner to the electron case, resulting in
where
$\unicode[STIX]{x1D70F}_{i}=(3\sqrt{\unicode[STIX]{x03C0}}/4)4\unicode[STIX]{x03C0}\unicode[STIX]{x1D700}_{0}^{2}m_{i}^{1/2}T_{i}^{3/2}/n_{0}\text{e}^{4}\ln \unicode[STIX]{x1D6EC}$
.
The calculation of the Pfirsch–Schlüter heat flux proceeds similarly to the electron one and yields
while the banana regime contribution is
Finally, the azimuthal drift contribution is
$$\begin{eqnarray}\displaystyle \left\langle \int \,\text{d}^{3}\boldsymbol{v}\frac{m_{i}v^{2}}{2}\boldsymbol{v}_{di}\boldsymbol{\cdot }\unicode[STIX]{x1D735}h_{0i}\right\rangle & = & \displaystyle -\frac{5}{2}i\ell \frac{T_{i}}{e}\left\langle \!\left(\frac{1}{B^{2}}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})+\left(\frac{1}{RB^{2}}\right)^{2}\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}\frac{\unicode[STIX]{x2202}B^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\!\right)\!\!\right\rangle \nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{\tilde{T}_{i}}{T_{i}}+\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{i}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{i})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)n_{0}T_{i}.\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
4.3 Summary
Here we collect together the above results to obtain the final form of the equations determining the electron density and temperature perturbations in terms of the perturbed fields
$\unicode[STIX]{x1D6F7}$
and
$\unicode[STIX]{x1D6F9}$
. The first, equation (4.2) can be written in the form
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{\text{d}}{\text{d}r}(\langle \unicode[STIX]{x1D6E4}_{r,e}^{C}\rangle +\langle \unicode[STIX]{x1D6E4}_{r,e}^{PS}\rangle +\langle \unicode[STIX]{x1D6E4}_{r,e}^{\text{Ban}}\rangle )+\text{i}k_{\Vert 0}\langle \unicode[STIX]{x1D6E4}_{\Vert ,e}\rangle +\text{i}k_{\unicode[STIX]{x1D703}}\langle \unicode[STIX]{x1D6E4}_{e}^{Az}\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\text{i}\unicode[STIX]{x1D714}\left[{\tilde{n}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F7}}{T_{e}}n_{0}+\left\langle \frac{\unicode[STIX]{x1D707}_{0}\tilde{p}}{B^{2}}\right\rangle \left(1-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\right)n_{0}\right],\end{eqnarray}$$
where
$k_{\Vert 0}=\ell Iq^{\prime }x\langle 1/R^{2}B\rangle /q$
,
$k_{\unicode[STIX]{x1D703}}=\ell q\langle |\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}|\rangle$
and we have substituted (4.1) for
$\unicode[STIX]{x1D6FF}B_{\Vert }$
. From (4.13) we have:
from (4.19):
$$\begin{eqnarray}\displaystyle \frac{1}{n_{0}}\frac{\text{d}}{\text{d}r}\langle \unicode[STIX]{x1D6E4}_{r,e}^{PS}\rangle & = & \displaystyle -\frac{m_{e}T_{e}}{\text{e}^{2}\unicode[STIX]{x1D70F}_{ei}}I^{2}\left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}}{p_{0e}}-\frac{3}{2}\frac{\tilde{T}_{e}}{T_{e}}\right)\nonumber\\ \displaystyle & & \displaystyle -\,\frac{\text{i}I^{2}\ell q^{\prime }T_{e}}{eq}\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\nonumber\\ \displaystyle & & \displaystyle \times \,\frac{\text{d}}{\text{d}x}\left[x\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e\unicode[STIX]{x1D6F7}}{T_{e}}+\left(1-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\right)\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right];\end{eqnarray}$$
from (4.23):
$$\begin{eqnarray}\displaystyle \frac{1}{n_{0}}\frac{\text{d}}{\text{d}r}\langle \unicode[STIX]{x1D6E4}_{r,e}^{\text{Ban}}\rangle & = & \displaystyle -1.53f_{t}\frac{I^{2}m_{e}T_{e}}{\text{e}^{2}\langle B^{2}\rangle \unicode[STIX]{x1D70F}_{ei}}\frac{\text{d}^{2}}{\text{d}x^{2}}\left[\frac{\tilde{p}}{p_{0e}}+1.39\frac{\tilde{T}_{e}}{T_{e}}-1.17\frac{\tilde{T}_{i}}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle -\,1.67\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{i}\ell q^{\prime }T_{e}}{eq}\frac{\text{d}}{\text{d}x}\left(x\left[\vphantom{\frac{\tilde{T}_{e}}{T_{e}}}\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}\right.\right.\nonumber\\ \displaystyle & & \displaystyle \left.\left.-\,\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+1.02\,\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}+0.02\frac{\tilde{T}_{e}}{T_{e}}\right]\right).\end{eqnarray}$$
From (4.40), we have defined
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \frac{1}{R^{2}B}\right\rangle \langle \unicode[STIX]{x1D6E4}_{\Vert ,e}\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\frac{I}{e}\tilde{p}^{\prime }\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)+\frac{Ip_{0e}f_{t}}{e\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \left[1.67\frac{\tilde{p}^{\prime }}{p_{0e}}-1.19\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-3.12\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \qquad -\,\frac{Ip_{0e}}{e}\left[\left\langle \frac{1}{R^{2}B^{2}}\right\rangle \left(\frac{p_{i}^{\prime }}{p_{0e}}+\frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{e}}\right)-\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]-\text{i}\frac{1.97\unicode[STIX]{x1D70F}_{ei}(I\ell q^{\prime }x)p_{0e}}{m_{e}\langle B^{2}\rangle q}\left\langle \frac{1}{R^{2}}\right\rangle ^{2}\nonumber\\ \displaystyle & & \displaystyle \qquad \times \,\left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}(1-1.31f_{t})\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \qquad \left.+\,0.34\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)(1-0.69f_{t})\right],\end{eqnarray}$$
while from (4.50) we have defined
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6E4}_{e}^{Az}\rangle & = & \displaystyle \frac{T_{e}}{eq\langle |\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}|\rangle }\left\langle \left(\frac{1}{B^{2}}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})+\left(\frac{1}{RB^{2}}\right)^{2}\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}\frac{\unicode[STIX]{x2202}B^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\right)\right\rangle \nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)n_{0}.\end{eqnarray}$$
Similarly, for (4.3) we obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{\text{d}}{\text{d}r}(\langle Q_{r,e}^{C}\rangle +\langle Q_{r,e}^{PS}\rangle +\langle Q_{r,e}^{\text{Ban}}\rangle )+\text{i}k_{\Vert 0}\langle Q_{\Vert e}\rangle +\text{i}k_{\unicode[STIX]{x1D703}}\langle Q_{e}^{Az}\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\frac{3}{2}\text{i}\unicode[STIX]{x1D714}\left[\tilde{p}_{e}-\frac{\unicode[STIX]{x1D714}_{\ast e}(1+\unicode[STIX]{x1D702}_{e})}{\unicode[STIX]{x1D714}}e\unicode[STIX]{x1D6F7}_{0}+\frac{5}{3}\left\langle \frac{\unicode[STIX]{x1D707}_{0}\tilde{p}}{B^{2}}\right\rangle \left(1-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{e})\right)p_{0e}\right],\end{eqnarray}$$
where, from (4.15),
from (4.20),
$$\begin{eqnarray}\displaystyle \frac{1}{n_{0}T_{e}}\frac{\text{d}}{\text{d}r}\langle Q_{r,e}^{PS}\rangle & = & \displaystyle -\frac{m_{e}T_{e}I^{2}}{\text{e}^{2}\unicode[STIX]{x1D70F}_{ei}}\left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}}{p_{0e}}+0.91\frac{\tilde{T}_{e}}{T_{e}}\right)\nonumber\\ \displaystyle & & \displaystyle -\,\frac{\text{i}I^{2}\ell q^{\prime }T_{e}}{eq}\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\nonumber\\ \displaystyle & & \displaystyle \times \,\frac{\text{d}}{\text{d}x}\left[\frac{5}{2}x\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e\unicode[STIX]{x1D6F7}}{T_{e}}+\frac{\tilde{T}_{e}}{T_{e}}+\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\left(1-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{e})\right)\right)\right]\qquad\end{eqnarray}$$
and from (4.25),
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1}{n_{0}T_{e}}\frac{\text{d}}{\text{d}r}\langle Q_{r,e}^{\text{Ban}}\rangle =-1.70f_{t}\frac{I^{2}m_{e}T_{e}}{\text{e}^{2}\langle B^{2}\rangle \unicode[STIX]{x1D70F}_{ei}}\frac{\text{d}^{2}}{\text{d}x^{2}}\left[\frac{\tilde{p}}{p_{0e}}-0.40\frac{\tilde{T}_{e}}{T_{e}}-1.17\frac{\tilde{T}_{i}}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,2.97\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{i}\ell q^{\prime }T_{e}}{eq}\frac{\text{d}}{\text{d}x}\left(x\left[\vphantom{\frac{\tilde{T}_{e}}{T_{e}}}\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+1.25\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right.\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.\left.+\,0.25\frac{\tilde{T}_{e}}{T_{e}}\right]\right),\end{eqnarray}$$
while, from (4.48),
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1}{n_{0}T_{e}}\left\langle \frac{1}{R^{2}B}\right\rangle \langle Q_{\Vert e}\rangle =\frac{5}{2}\frac{IT_{e}}{e}\left(\left\langle \frac{1}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \right)\left(\frac{\tilde{p}^{\prime }}{p_{0e}}+\frac{\tilde{T}_{e}^{\prime }}{T_{e}}\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{f_{t}\langle R^{2}B\rangle }{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \frac{IT_{e}}{e}\left[4.18\frac{\tilde{p}^{\prime }}{p_{0e}}-2.36\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-7.86\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{5}{2}\frac{\text{i}\ell I^{2}q^{\prime }xT_{i}}{eq}\left[\left\langle \frac{1}{R^{2}B^{2}}\right\rangle \left(\frac{p_{i}^{\prime }}{p_{0e}}+\frac{e\unicode[STIX]{x1D6F7}^{\prime }}{T_{e}}\right)-\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,6.26\frac{\text{i}\ell Iq^{\prime }x}{\langle B^{2}\rangle q}\left\langle \frac{1}{R^{2}}\right\rangle ^{2}\frac{T_{e}\unicode[STIX]{x1D70F}_{ei}}{m_{e}}\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left[(1-1.11f_{t})\left(\frac{\tilde{p}_{e}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.58(1-0.31f_{t})\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D702}_{e}\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right]\end{eqnarray}$$
and, from (4.51),
$$\begin{eqnarray}\displaystyle \langle Q_{e}^{Az}\rangle & = & \displaystyle \frac{5}{2}\frac{T_{e}}{qe\langle |\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}|\rangle }\left\langle \left(\frac{1}{B^{2}}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})+\left(\frac{1}{RB^{2}}\right)^{2}\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}\frac{\unicode[STIX]{x2202}B^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\right)\right\rangle \nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{\tilde{p}_{e}}{p_{0e}}+\frac{\tilde{T}_{e}}{T_{e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)n_{0}T_{e}.\end{eqnarray}$$
Summarising for the ions, we can rewrite (4.52) as:
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{\text{d}}{\text{d}r}(\langle Q_{r,i}^{C}\rangle +\langle Q_{r,i}^{PS}\rangle +\langle Q_{r,i}^{\text{Ban}}\rangle )+\text{i}k_{\unicode[STIX]{x1D703}}\langle Q_{i}^{Az}\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\frac{3}{2}\text{i}\unicode[STIX]{x1D714}\left[\tilde{p}_{i}-\frac{\unicode[STIX]{x1D714}_{\ast e}(1+\unicode[STIX]{x1D702}_{i})}{\unicode[STIX]{x1D714}}\frac{T_{i}}{T_{e}}e\unicode[STIX]{x1D6F7}n_{0}+\frac{5}{3}p_{0i}\left\langle \frac{\unicode[STIX]{x1D707}_{0}\tilde{p}}{B^{2}}\right\rangle \left(1+\frac{T_{i}}{T_{e}}\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{i})\right)\right],\qquad\end{eqnarray}$$
where, from (4.55),
from (4.56)
from (4.57)
and, finally, from (4.58):
$$\begin{eqnarray}\displaystyle \langle Q_{i}^{Az}\rangle & = & \displaystyle -\frac{5}{2}\frac{T_{i}}{eq\langle |\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}|\rangle }\left\langle \left(\frac{1}{B^{2}}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})+\left(\frac{1}{RB^{2}}\right)^{2}\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}\frac{\unicode[STIX]{x2202}B^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\right)\right\rangle \nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{\tilde{T}_{i}}{T_{i}}+\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{i}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+2\unicode[STIX]{x1D702}_{i})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)n_{0}T_{i}.\end{eqnarray}$$
5 The field equations
The set of equations is completed by using the quasi-neutrality condition (to higher order than introduced in the previous section) and the parallel Ampère’s law, as in Connor et al. (Reference Connor, Hastie and Helander2009). These will only be briefly discussed here, focussing on any differences from the work of Connor et al. (Reference Connor, Hastie and Helander2009) arising from the more complete description of the electron and ion continuity equations in the previous sections and, effectively, a novel calculation of the neoclassical angular toroidal viscosity.
A convenient approach to imposing quasi-neutrality in higher order is through the vorticity equation, obtained by taking the charge density moment of the gyro-kinetic equations for both species and adding them. Setting
${\tilde{n}}_{e}={\tilde{n}}_{i}$
then provides one relationship between
$\unicode[STIX]{x1D6F7}$
and
$\unicode[STIX]{x1D6F9}$
. The parallel Ampère’s law provides a second and hence these two equations lead to an eigenvalue condition on
$\unicode[STIX]{x1D714}$
, provided the solutions of the various continuity equations for
${\tilde{n}}$
,
$\tilde{T}_{e}$
and
$\tilde{T}_{i}$
are expressed in terms of
$\unicode[STIX]{x1D6F7}$
and
$\unicode[STIX]{x1D6F9}$
.
5.1 The parallel Ampère’s law
The parallel Ampère’s law states that
We expand (5.1) in the localisation,
$x$
, expressing
$A_{\Vert }$
in terms of
$\unicode[STIX]{x1D6F9}$
as defined above (2.9), which we expand in the form
$\unicode[STIX]{x1D6F9}=\unicode[STIX]{x1D6F9}^{(0)}+\unicode[STIX]{x1D6F9}^{(1)}+\cdots \,$
. In leading order we have
while in next order
leading to the solubility condition
where, from (4.38),
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \frac{j_{\Vert }B}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle =-I\tilde{p}^{\prime }\left(\left\langle \frac{1}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle -\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \frac{1}{\langle B^{2}\rangle }\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,1.97\unicode[STIX]{x1D70F}_{ei}\frac{\text{i}I\ell q^{\prime }xp_{0e}e}{m_{e}q}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left[\vphantom{\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)}(1-1.31f_{t})\left(\frac{\tilde{p}}{p_{0e}}-\frac{e(\unicode[STIX]{x1D6F7}-\unicode[STIX]{x1D6F9})}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{e})\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.34(1-0.69f_{t})\left(\frac{\tilde{T}_{e}}{T_{e}}-\frac{\unicode[STIX]{x1D714}_{\ast e}\unicode[STIX]{x1D702}_{e}}{\unicode[STIX]{x1D714}}\frac{e\unicode[STIX]{x1D6F9}}{T_{e}}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{f_{t}Ip_{0e}}{\langle B^{2}\rangle }\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left[1.67\frac{\tilde{p}^{\prime }}{p_{0e}}-1.19\frac{\tilde{T}_{e}^{\prime }}{T_{e}}-1.95\frac{\tilde{T}_{i}^{\prime }}{T_{e}}\right].\end{eqnarray}$$
5.2 The vorticity equation
The vorticity equation was obtained in the previous publication (Connor et al.
Reference Connor, Hastie and Helander2009). Here we list the key steps in its derivation and quote the result. The procedure was to add the velocity moments of the gyro-kinetic equations for the electrons and ions, take the long wavelength limit:
$k_{\bot }^{2}\unicode[STIX]{x1D70C}_{i}^{2}\ll 1$
and exploit the lowest-order quasi-neutrality condition:
${\tilde{n}}_{e}={\tilde{n}}_{i}$
, to obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{\boldsymbol{B}\boldsymbol{\cdot }\unicode[STIX]{x1D735}}{\unicode[STIX]{x1D714}^{2}}\left(\frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{\unicode[STIX]{x1D707}_{0}B^{2}}\frac{\text{d}^{2}}{\text{d}x^{2}}\boldsymbol{B}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D6F9}\right)\nonumber\\ \displaystyle & & \displaystyle \quad =\mathop{\sum }_{j}\frac{e_{j}^{2}}{T_{j}}\int \,\text{d}^{3}vf_{0j}\left\{\frac{\unicode[STIX]{x1D714}_{\ast j}^{T}}{\unicode[STIX]{x1D714}}\unicode[STIX]{x1D6F7}-\left(1-\frac{\unicode[STIX]{x1D714}_{\ast j}^{T}}{\unicode[STIX]{x1D714}}\right)\left[\frac{m_{j}^{2}v_{\bot }^{2}|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{2e_{j}^{2}B^{2}}\frac{\text{d}^{2}}{\text{d}x^{2}}\unicode[STIX]{x1D6F7}+\frac{m_{j}v_{\bot }^{2}}{2e_{j}B}\unicode[STIX]{x1D6FF}B_{\Vert }\right]\right\}\nonumber\\ \displaystyle & & \displaystyle \qquad +\,\text{i}\mathop{\sum }_{j}e_{j}\int \,\text{d}^{3}\boldsymbol{v}\frac{\boldsymbol{v}_{dj}\boldsymbol{\cdot }\unicode[STIX]{x1D735}g_{j}}{\unicode[STIX]{x1D714}}+\frac{1}{4\unicode[STIX]{x1D714}}\mathop{\sum }_{j}e_{j}\int \,\text{d}^{3}\boldsymbol{v}\overline{L_{j}^{2}C_{j}(g_{j}L_{j}^{2})}\nonumber\\ \displaystyle & & \displaystyle \qquad -\,\frac{\text{i}}{\unicode[STIX]{x1D714}}\mathop{\sum }_{j}\frac{m_{j}}{e_{j}}\int \,\text{d}^{3}\boldsymbol{v}\unicode[STIX]{x1D70E}|v_{\Vert }|\frac{\boldsymbol{B}\boldsymbol{\cdot }\unicode[STIX]{x1D735}}{B}\left(\frac{m_{j}v_{\bot }^{2}|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{2B^{2}}\right)\frac{\text{d}^{2}}{\text{d}x^{2}}g_{j},\end{eqnarray}$$
where the divergence of the parallel current has been expressed in terms of parallel gradients of
$A_{\Vert }$
, i.e.
$\unicode[STIX]{x1D6F9}$
, through Ampère’s law, equation (5.1). The distribution functions
$g_{j}$
in the velocity space integrations over the magnetic drift terms are expressed in terms of the quantities
$h_{j}$
and we note the final term vanishes for up–down symmetric equilibria. The expansions
$\unicode[STIX]{x1D6F9}=\unicode[STIX]{x1D6F9}^{(0)}+\unicode[STIX]{x1D6F9}^{(1)}+\unicode[STIX]{x1D6F9}^{(2)}\ldots$
;
$h_{j}=h_{0j}+h_{1j}+h_{2j}\ldots$
are introduced and the equation for
$\unicode[STIX]{x1D6F9}$
solved order by order.
In leading order one finds
$\unicode[STIX]{x1D6F9}^{(0)}$
is independent of
$\unicode[STIX]{x1D703}$
, while the equation for
$\unicode[STIX]{x1D6F9}^{(1)}$
:
can be integrated, introducing a constant of integration which can in turn be determined through a periodicity condition in
$\unicode[STIX]{x1D703}$
on
$\unicode[STIX]{x1D6F9}^{(1)}$
. Applying the same periodicity condition on
$\unicode[STIX]{x1D6F9}^{(2)}$
in second order provides the required equation for
$\unicode[STIX]{x1D6F9}^{(0)}(x)$
:
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{I^{2}}{\unicode[STIX]{x1D714}^{2}}\left[\frac{\ell ^{2}{q^{\prime }}^{2}}{\unicode[STIX]{x1D707}_{0}q^{2}}\left\langle \frac{1}{R^{2}}\right\rangle ^{2}x\frac{\text{d}^{2}(x\unicode[STIX]{x1D6F9}^{(0)})}{\text{d}x^{2}}+\frac{\ell q^{\prime }x\unicode[STIX]{x1D714}}{q}\tilde{p}^{\prime }\left(\left\langle \frac{1}{B^{2}R^{2}}\right\rangle \!\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle -\left\langle \frac{1}{R^{2}}\right\rangle \!\left\langle \frac{1}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad =m_{i}n_{0}\left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{B^{2}}\right\rangle \left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left[1-\frac{\unicode[STIX]{x1D714}_{\ast i}}{\unicode[STIX]{x1D714}}(1+\unicode[STIX]{x1D702}_{i})\right]\frac{\text{d}^{2}\unicode[STIX]{x1D6F7}}{\text{d}x^{2}}\nonumber\\ \displaystyle & & \displaystyle \qquad -\,2\frac{\ell }{\unicode[STIX]{x1D714}}\tilde{p}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{R^{2}B^{2}}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}\left(\unicode[STIX]{x1D707}_{0}p_{0}+\frac{B^{2}}{2}\right)+\frac{\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}}{R^{2}B^{2}}\frac{\unicode[STIX]{x2202}B}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \qquad -\,\text{i}\frac{I}{\unicode[STIX]{x1D714}}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left\langle \mathop{\sum }_{j}\int \,\text{d}^{3}\boldsymbol{v}\frac{m_{j}v_{\Vert }}{B}\frac{I}{R^{2}q}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{v_{\Vert }}{B}\right){h^{\prime }}_{j}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \qquad -\,\frac{\ell {p^{\prime }}_{0}I^{2}}{\unicode[STIX]{x1D714}}\tilde{p}\left[\left\langle \frac{1}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}B^{2}}\right\rangle \left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle -\left\langle \frac{1}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle ^{2}\right]\nonumber\\ \displaystyle & & \displaystyle \qquad -\,\frac{\ell ^{2}q^{\prime }{p^{\prime }}_{0}I^{2}}{\unicode[STIX]{x1D714}^{2}q}(x\,\unicode[STIX]{x1D6F9}^{(0)})^{\prime }\left[\left\langle \frac{1}{R^{2}}\right\rangle \left\langle \frac{1}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle -\left\langle \frac{1}{B^{2}R^{2}}\right\rangle \left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \right]\nonumber\\ \displaystyle & & \displaystyle \qquad +\,\frac{1}{4\unicode[STIX]{x1D714}}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left\langle \mathop{\sum }_{j}e_{j}\int \,\text{d}^{3}\boldsymbol{v}\overline{L_{j}^{2}C_{j}(g_{0j}L_{j}^{2})}\right\rangle ,\end{eqnarray}$$
where we have substituted for
$\unicode[STIX]{x1D6FF}B_{\Vert }$
from (4.1) (which has the effect of replacing the
$\unicode[STIX]{x1D735}B$
drift by the curvature drift) and substituted for
$\unicode[STIX]{x1D6F9}^{(1)}$
from the solution of (5.7). The term involving
$\langle \sum _{j}e_{j}\int \,\text{d}^{3}\boldsymbol{v}\overline{L_{j}^{2}C_{j}(g_{0j}L_{j}^{2})}\rangle$
represents the contribution from the classical radial transport of toroidal momentum, calculated in (B 2) of appendix B.
It remains to evaluate the term involving the
$h_{j}$
, which we do by repeated application of the gyro-kinetic equations for
$h_{0j}$
,
$h_{1j}$
,
$h_{2j}$
and
$h_{3j}$
with integration by parts in
$\unicode[STIX]{x1D703}$
and noting momentum conservation in ion–ion collisions. The result can be expressed as
$$\begin{eqnarray}\displaystyle & & \displaystyle \text{i}\frac{I}{\unicode[STIX]{x1D714}}\left\langle \mathop{\sum }_{j}\int \,\text{d}^{3}\boldsymbol{v}\frac{m_{j}v_{\Vert }}{B}\frac{I}{R^{2}q}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{v_{\Vert }}{B}\right){h^{\prime }}_{j}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad =\text{i}\frac{I}{\unicode[STIX]{x1D714}}\left\langle \mathop{\sum }_{j}\int \,\text{d}^{3}\boldsymbol{v}\frac{Im_{j}v_{\Vert }}{B}\left\{-\frac{Iv_{\Vert }}{R^{2}Bq}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}\left(\frac{v_{\Vert }}{\unicode[STIX]{x1D6FA}_{i}}\right){h^{\prime \prime }}_{j}+C_{j}({h^{\prime }}_{j})\right.\right.\nonumber\\ \displaystyle & & \displaystyle \qquad \left.\left.+\,\text{i}{\unicode[STIX]{x1D714}h^{\prime }}_{j}-\text{i}\ell q^{\prime }\frac{Iv_{\Vert }}{R^{2}Bq}({xh^{\prime }}_{j})\vphantom{\mathop{\sum }_{j}}\right\}\right\rangle .\end{eqnarray}$$
The first term can be recognised as involving the neoclassical radial transport of toroidal angular momentum:
where the first non-vanishing contribution is from
$h_{3_{j}}$
. The right-hand side of expression (5.9) thus reduces to
Further use of the gyro-kinetic equation and integrations by parts in
$\unicode[STIX]{x1D703}$
implies
Assembling all the contributions to (5.8) and substituting for
$U_{\Vert i}$
from (3.22),
$$\begin{eqnarray}\displaystyle & & \displaystyle x\frac{\text{d}^{2}}{\text{d}x^{2}}(x\unicode[STIX]{x1D6F9}^{(0)})+\left(D_{I}+\frac{1}{4}\right)\frac{\unicode[STIX]{x1D714}}{\ell p_{0}^{\prime }}\tilde{p}-(L+H)x\frac{\text{d}}{\text{d}x}\left(\frac{\unicode[STIX]{x1D714}}{\ell p_{0}^{\prime }}\tilde{p}-\unicode[STIX]{x1D6F9}^{(0)}\right)\nonumber\\ \displaystyle & & \displaystyle \quad =\frac{\unicode[STIX]{x1D707}_{0}m_{i}n_{0}}{\ell ^{2}}\left(\frac{q}{Iq^{\prime }\langle 1/R^{2}\rangle }\right)^{2}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{B^{2}}\right\rangle \unicode[STIX]{x1D714}(\unicode[STIX]{x1D714}-\unicode[STIX]{x1D714}_{\ast i}(1+\unicode[STIX]{x1D702}_{i}))\frac{\text{d}^{2}}{\text{d}x^{2}}\unicode[STIX]{x1D6F7}\nonumber\\ \displaystyle & & \displaystyle \qquad +\,\frac{\unicode[STIX]{x1D707}_{0}m_{i}n_{0}}{\ell ^{2}}\left(\frac{q}{q^{\prime }\langle 1/R^{2}\rangle }\right)^{2}\frac{\unicode[STIX]{x1D714}^{2}T_{i}}{e}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \frac{\text{d}^{2}}{\text{d}x^{2}}\nonumber\\ \displaystyle & & \displaystyle \qquad \times \,\left\{\left\langle \frac{1}{B^{2}}\right\rangle \left(\frac{{\tilde{n}}}{n_{0}}+\frac{e\unicode[STIX]{x1D6F7}}{T_{i}}\right)+\left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1.17}{\langle B^{2}\rangle }(1-f_{t})\right)\frac{\tilde{T}_{i}}{T_{i}}\right\}\nonumber\\ \displaystyle & & \displaystyle \qquad -\,\frac{\text{i}\unicode[STIX]{x1D714}\,\unicode[STIX]{x1D707}_{0}}{\ell ^{2}}\left(\frac{q}{Iq^{\prime }\langle 1/R^{2}\rangle }\right)^{2}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \frac{\text{d}^{2}}{\text{d}x^{2}}\langle \unicode[STIX]{x1D6F1}^{C}+\unicode[STIX]{x1D6F1}^{NC}\rangle .\end{eqnarray}$$
Here we have introduced the quantity
$D_{I}$
in the notation of Glasser et al. (Reference Glasser, Greene and Johnson1975), where
$$\begin{eqnarray}\displaystyle D_{I} & = & \displaystyle -\frac{1}{4}+p^{\prime }\frac{q}{q^{\prime }}\left\langle \frac{R^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle -\left(p^{\prime }\frac{q}{q^{\prime }}\right)^{2}\left\langle \frac{R^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle ^{2}\nonumber\\ \displaystyle & & \displaystyle +\,\left(\frac{p^{\prime }}{I}\frac{q}{q^{\prime }}\right)^{2}\left\langle \frac{R^{4}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left\langle \frac{B^{2}R^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle -\,p^{\prime }\left(\frac{1}{I}\frac{q}{q^{\prime }}\right)^{2}\left\langle \frac{B^{2}R^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left\langle \frac{\unicode[STIX]{x2202}R^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}-\frac{R^{2}q}{I}\frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}\left(\frac{I}{q}\right)\right\rangle ,\end{eqnarray}$$
and
$L$
, appearing in the work of Hahm (Reference Hahm1988), where
The quantity
$D_{I}$
plays a role in the Mercier stability criterion:
$D_{I}+1/4>0$
(Mercier Reference Mercier1960) while the combination
$L+H$
in (5.13) is given by
An expression for
$\langle \unicode[STIX]{x1D6F1}^{NC}\rangle$
has been given by Wong & Chan (Reference Wong and Chan2005):
$\langle \unicode[STIX]{x1D6F1}^{NC}\rangle =0.19(n_{0}m_{i}^{2}T_{i}\,f_{t}^{3}/\unicode[STIX]{x1D70F}_{ii}e^{3})\,\text{d}^{2}\tilde{T}_{i}/\text{d}x^{2}$
, where
$\unicode[STIX]{x1D70F}_{ii}=\unicode[STIX]{x1D70F}_{i}/\sqrt{2}$
, but we generalise this in appendix B for arbitrary axisymmetric toroidal geometry, to obtain the total collisional toroidal angular viscosity:
where
and
$$\begin{eqnarray}\displaystyle & & \displaystyle \langle \unicode[STIX]{x1D6F1}^{NC}\rangle =0.80\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[\left\langle \frac{1}{B^{4}}\right\rangle +\frac{(G^{(3)}-G^{(2)})}{\langle B^{2}\rangle ^{2}}\right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{e\unicode[STIX]{x1D6F7}}{T_{i}}\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}\langle B^{2}\rangle }\left(\frac{I^{2}m_{i}T_{i}}{e^{2}}\right)^{2}\left[0.37\left(\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{B^{2}}\right\rangle -\left\langle \frac{1}{B}\right\rangle \left\langle \frac{1}{B^{3}}\right\rangle \right)+0.54\left\langle \frac{1}{B^{4}}\right\rangle \right]\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}\langle B^{2}\rangle ^{2}}\left(\frac{I^{2}m_{i}T_{i}}{e^{2}}\right)^{2}(0.37G^{(0)}-0.62G^{[1]}+0.94G^{(2)}-0.2G^{(3)}-0.74G^{(4)})\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right),\end{eqnarray}$$
with
$$\begin{eqnarray}\displaystyle \frac{G^{(1)}}{3\langle B^{2}\rangle } & = & \displaystyle \frac{15}{4}\int _{0}^{1/B_{\text{Min}}}\,\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}\frac{\left\langle B\sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\left\langle \frac{\sqrt{1-\unicode[STIX]{x1D706}B}}{B}\right\rangle \nonumber\\ \displaystyle & & \displaystyle +\,\frac{15}{16}\langle B^{2}\rangle \left\langle \frac{1}{B}\right\rangle \int _{0}^{1/B_{\text{Max}}}\frac{\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}^{2}}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\nonumber\\ \displaystyle & & \displaystyle -\,\frac{15}{8}\langle B^{2}\rangle \int _{0}^{1/B_{\text{Max}}}\,\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}\frac{1}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\left\langle \frac{\sqrt{1-\unicode[STIX]{x1D706}B}}{B}\right\rangle \int _{\unicode[STIX]{x1D706}}^{1/B_{\text{Max}}}\frac{\text{d}\unicode[STIX]{x1D706}^{\prime }}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}^{\prime }B}\right\rangle },\qquad\end{eqnarray}$$
and
$$\begin{eqnarray}\displaystyle G^{(4)} & = & \displaystyle \frac{15\langle B^{2}\rangle ^{2}}{4}\int _{0}^{1/B_{\text{Max}}}\,\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}\frac{1}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{1}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\left\langle \frac{\sqrt{1-\unicode[STIX]{x1D706}B}}{B}\right\rangle \left\langle \frac{1-\unicode[STIX]{x1D706}B}{B}\right\rangle -\left\langle \frac{1-\unicode[STIX]{x1D706}B}{B^{2}}\right\rangle \right).\end{eqnarray}$$
Insertion of the results (5.18) and (5.19) completes the form of the vorticity equation (5.13). Analytic evaluations of the coefficients in (5.19) in the large aspect ratio limit are presented in appendix B and are consistent with those given by Wong & Chan (Reference Wong and Chan2005).
6 Normalised equations
It is convenient to introduce a new radial coordinate normalised to the semi-collisional width,
$\unicode[STIX]{x1D6FF}_{e}$
(but now expressed in terms of flux coordinates) and a set of normalised parameters
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}s=\text{e}^{\text{i}\unicode[STIX]{x03C0}/4}{\displaystyle \frac{\ell q^{\prime }}{q}}\left(1.97{\displaystyle \frac{T_{e}}{m_{e}}}{\displaystyle \frac{\unicode[STIX]{x1D70F}_{ei}}{\unicode[STIX]{x1D714}}}\left\langle {\displaystyle \frac{1}{R^{2}}}\right\rangle ^{2}{\displaystyle \frac{I^{2}}{\langle B^{2}\rangle }}\right)^{1/2}x\equiv {\displaystyle \frac{x}{\unicode[STIX]{x1D6FF}_{e}}},\\ \hat{\unicode[STIX]{x1D714}}={\displaystyle \frac{en_{0}}{\ell T_{e}n_{0}^{\prime }}}\unicode[STIX]{x1D714},\quad \unicode[STIX]{x1D711}={\displaystyle \frac{e\unicode[STIX]{x1D6F7}}{T_{e}}},\quad \unicode[STIX]{x1D713}={\displaystyle \frac{e\unicode[STIX]{x1D6F9}}{T_{e}}},\quad n={\displaystyle \frac{{\tilde{n}}}{n_{0}}},\quad t_{e}={\displaystyle \frac{\tilde{T}_{e}}{T_{e}}},\quad t_{i}={\displaystyle \frac{\tilde{T}_{i}}{T_{e}}},\\ p_{j}={\displaystyle \frac{\tilde{p}_{j}}{p_{0e}}},\quad p={\displaystyle \frac{\tilde{p}}{p_{0e}}},\quad \unicode[STIX]{x1D705}={\displaystyle \frac{q^{\prime }}{q}}{\displaystyle \frac{n_{0}}{n_{0}^{\prime }}},\\ C={\displaystyle \frac{q}{q^{\prime }}}\left\langle \left({\displaystyle \frac{1}{B^{2}}}{\displaystyle \frac{\unicode[STIX]{x2202}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D712}}}(\unicode[STIX]{x1D707}_{0}p_{0}+B^{2})+\left({\displaystyle \frac{1}{RB^{2}}}\right)^{2}\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\boldsymbol{\cdot }\unicode[STIX]{x1D735}\unicode[STIX]{x1D703}{\displaystyle \frac{\unicode[STIX]{x2202}B^{2}}{\unicode[STIX]{x2202}\unicode[STIX]{x1D703}}}\right)\right\rangle ,\end{array}\right\}\end{eqnarray}$$
where
$\unicode[STIX]{x1D6FF}_{e}$
is the semi-collisional width in flux coordinates. Consequently, the normalised electron continuity equation (4.59) becomes:
$$\begin{eqnarray}\displaystyle & & \displaystyle n-\frac{\unicode[STIX]{x1D711}}{\hat{\unicode[STIX]{x1D714}}}+\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle p\left(1-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right)-\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}C\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,s^{2}\left[(1-1.31f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)+0.34(1-0.69f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{\unicode[STIX]{x1D705}s}{\hat{\unicode[STIX]{x1D714}}}\frac{\text{d}}{\text{d}s}\left[\left(\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right)p+\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle (1.67p-1.19t_{e}-3.12t_{i})\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.-\,\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle (p_{i}+\unicode[STIX]{x1D711})+\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle t_{i}\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\left(\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right)\frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,1.67f_{t}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+1.02\,\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}+0.02t_{e}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,1.97\left(\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\right)^{2}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad \times \,\frac{\text{d}^{2}}{\text{d}s^{2}}\left[\vphantom{\left\langle \frac{I^{2}}{R^{2}}\right\rangle }\left(\left\langle \frac{\left|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}\right|^{2}}{B^{2}}\right\rangle \left\langle \frac{1}{R^{2}}\right\rangle +\left\langle \frac{I^{2}}{R^{2}}\right\rangle \left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\right)\left(p-\frac{3}{2}t_{e}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,\frac{1.53f_{t}}{\left\langle B^{2}\right\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle (p+1.39t_{e}-1.17t_{i})\right]=0,\end{eqnarray}$$
the electron thermal equation (4.60):
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{3}{2}\left[p_{e}-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D711}+\frac{5}{3}\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle p\,\left(1-\frac{1+2\,\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{5}{2}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}C\left(p_{e}+t_{e}-\left(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713}\right)-\frac{1+2\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,3.18s^{2}\!\left[(1-1.11f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)+0.58(1-0.31f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{5}{2}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\left(\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right)s\frac{\text{d}}{\text{d}s}(p+t_{e})\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle s\frac{\text{d}}{\text{d}s}(4.18p-2.36t_{e}-7.86t_{i})\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{5}{2}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}s\frac{\text{d}}{\text{d}s}\left[\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle (p_{i}+\unicode[STIX]{x1D711})-\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle t_{i}\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{5}{2}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\left(\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle -\frac{1}{\left\langle B^{2}\right\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right)\frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+2\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}+t_{e}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,2.97f_{t}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+1.25\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}+0.25t_{e}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,1.97\left(\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\right)^{2}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad \times \,\frac{\text{d}^{2}}{\text{d}s^{2}}\left[\left(\left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{B^{2}}\right\rangle \left\langle \frac{1}{R^{2}}\right\rangle +\left\langle \frac{I^{2}}{R^{2}}\right\rangle \left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\right)(p+0.91t_{e})\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,\frac{1.70f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle (p-0.4t_{e}-1.17t_{i})\right]=0,\end{eqnarray}$$
the ion thermal equation (4.61):
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{3}{2}\left(p_{i}-\frac{1+\unicode[STIX]{x1D702}_{i}}{\hat{\unicode[STIX]{x1D714}}}\frac{T_{i}}{T_{e}}\unicode[STIX]{x1D711}+\frac{5}{3}\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle \left(1+\frac{T_{i}}{T_{e}}\frac{1+2\unicode[STIX]{x1D702}_{i}}{\hat{\unicode[STIX]{x1D714}}}\right)\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{5}{2}\frac{T_{i}}{T_{e}}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}C\left(p_{i}+t_{i}+(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+2\unicode[STIX]{x1D702}_{i}}{\hat{\unicode[STIX]{x1D714}}}\frac{T_{i}}{T_{e}}\unicode[STIX]{x1D713}\right)+3.94\left(\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\right)^{2}\frac{m_{i}\unicode[STIX]{x1D70F}_{ei}}{m_{e}\unicode[STIX]{x1D70F}_{i}}\frac{T_{i}}{T_{e}}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \nonumber\\ \displaystyle & & \displaystyle \quad \times \,\frac{\text{d}^{2}}{\text{d}s^{2}}\left[\left(\left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{B^{2}}\right\rangle \left\langle \frac{1}{R^{2}}\right\rangle +\left\langle \frac{I^{2}}{R^{2}}\right\rangle \left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\right)+\frac{0.46f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right]t_{i}=0,\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
Ampère’s law (5.4), with result (5.5) for
$j_{\Vert }$
:
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{\text{d}^{2}}{\text{d}s^{2}}(s\unicode[STIX]{x1D713})=\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\langle R^{-2}\rangle }\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left[\left(\left\langle \frac{1}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \right)\frac{\text{d}p}{\text{d}s}+\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.\times \,\left(1.67\frac{\text{d}p}{\text{d}s}-1.19\frac{\text{d}t_{e}}{\text{d}s}-1.95\frac{\text{d}t_{i}}{\text{d}s}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\left(\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\right)^{2}\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{I^{2}\langle R^{-2}\rangle ^{2}}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle s\left[(1-1.31f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.34(1-0.66f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\vphantom{\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)}\right],\end{eqnarray}$$
and the vorticity equation (5.13):
$$\begin{eqnarray}\displaystyle & & \displaystyle s\frac{\text{d}^{2}}{\text{d}s^{2}}(s\unicode[STIX]{x1D713})+\left(D_{I}+\frac{1}{4}\right)\left(\frac{T_{e}n_{0}^{\prime }}{p_{0}^{\prime }}\right)\hat{\unicode[STIX]{x1D714}}p-(L+H)s\frac{\text{d}}{\text{d}s}\left(\frac{T_{e}n_{0}^{\prime }}{p_{0}^{\prime }}\hat{\unicode[STIX]{x1D714}}p-\unicode[STIX]{x1D713}\right)\nonumber\\ \displaystyle & & \displaystyle \quad =1.97\text{i}\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\langle B^{2}\rangle }\frac{m_{i}}{m_{e}}\left(\frac{\ell T_{e}n_{0}^{\prime }}{en_{0}}\unicode[STIX]{x1D70F}_{ei}\right)\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{B^{2}}\right\rangle \hat{\unicode[STIX]{x1D714}}\left(1+\frac{T_{i}}{\hat{\unicode[STIX]{x1D714}}T_{e}}\left(1+\unicode[STIX]{x1D702}_{i}\right)\right)\frac{\text{d}^{2}}{\text{d}s^{2}}\unicode[STIX]{x1D711}\nonumber\\ \displaystyle & & \displaystyle \qquad +\,1.97\text{i}\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\langle B^{2}\rangle }\frac{m_{i}}{m_{e}}\left(\frac{\ell T_{e}n_{0}^{\prime }}{en_{0}}\unicode[STIX]{x1D70F}_{ei}\right)\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \hat{\unicode[STIX]{x1D714}}\frac{\text{d}^{2}}{\text{d}s^{2}}\nonumber\\ \displaystyle & & \displaystyle \qquad \times \,\left[\left\langle \frac{I^{2}}{B^{2}}\right\rangle \left(\frac{T_{i}}{T_{e}}n+\unicode[STIX]{x1D711}\right)+\left(\left\langle \frac{I^{2}}{B^{2}}\right\rangle -\frac{1.17I^{2}}{\langle B^{2}\rangle }(1-f_{t})\right)t_{i}\right]\nonumber\\ \displaystyle & & \displaystyle \qquad -\,(1.97)^{2}\frac{\text{i}}{\hat{\unicode[STIX]{x1D714}}}\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\langle B^{2}\rangle }\frac{T_{i}}{T_{e}}\left(\frac{\unicode[STIX]{x1D70F}_{ei}^{2}}{\unicode[STIX]{x1D70F}_{i}}\frac{\ell T_{e}n_{0}^{\prime }}{en_{0}}\right)\left(\frac{m_{i}}{m_{e}\unicode[STIX]{x1D705}}\right)^{2}\frac{I^{2}}{\langle B^{2}\rangle }\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \frac{\text{d}^{2}}{\text{d}s^{2}}\hat{\unicode[STIX]{x1D6F1}},\end{eqnarray}$$
where
$$\begin{eqnarray}\displaystyle \frac{\text{d}^{2}}{\text{d}s^{2}}\hat{\unicode[STIX]{x1D6F1}} & = & \displaystyle -0.80\left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{4}}{B^{4}}\right\rangle \left\langle \frac{1}{R^{2}}\right\rangle ^{2}\frac{\text{d}^{4}}{\text{d}s^{4}}(p_{i}+\unicode[STIX]{x1D711})+0.80\left\langle \frac{I^{2}}{R^{2}}\right\rangle ^{2}\nonumber\\ \displaystyle & & \displaystyle \times \,\left[\left\langle \frac{1}{B^{4}}\right\rangle +\frac{(G^{(3)}-G^{(2)})}{\langle B^{2}\rangle ^{2}}\right]\frac{\text{d}^{4}}{\text{d}s^{4}}(p_{i}+\unicode[STIX]{x1D711})\nonumber\\ \displaystyle & & \displaystyle +\,\left\langle \frac{I^{2}}{R^{2}}\right\rangle ^{2}\left[0.37\left(\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{B^{2}}\right\rangle -\left\langle \frac{1}{B}\right\rangle \left\langle \frac{1}{B^{3}}\right\rangle \right)+0.54\left\langle \frac{1}{B^{4}}\right\rangle \right.\nonumber\\ \displaystyle & & \displaystyle \left.-\,2.4\left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{4}}{I^{4}B^{4}}\right\rangle \right]\frac{\text{d}^{4}t_{i}}{\text{d}s^{4}}\nonumber\\ \displaystyle & & \displaystyle +\,(0.37G^{(0)}-0.62G^{(1)}+0.94G^{(2)}-0.2G^{(3)}-0.74G^{(4)})\frac{\text{d}^{4}t_{i}}{\text{d}s^{4}},\end{eqnarray}$$
with
$G^{(0)}$
,
$G^{(1)}$
,
$G^{(2)}$
,
$G^{(3)}$
and
$G^{(4)}$
given in (5.20)–(5.24), respectively.
Because the ion thermal conductivity exceeds that of the electrons by
$0(\sqrt{m_{i}/m_{e}})$
, the ion temperature is flattened over the semi-collisional width and one can treat it as constant. In particular, the terms involving the ion temperature gradients in the classical and neoclassical toroidal viscosity (6.7) can be neglected, simplifying it considerably. It is only in the region
$s\sim (m_{i}/m_{e})^{1/4}$
that one needs to solve the ion thermal differential equation. As we shall see the electron equations simplify in this region of large
$s$
, thus simplifying the form of the ion thermal equation.
Equations (6.2)–(6.5) simplify if we ignore the classical and Pfirsch–Schlüter transport relative to the larger banana contribution. Furthermore, recognising that the azimuthal fluxes are small and that
$\unicode[STIX]{x1D6FD}_{e}=2\unicode[STIX]{x1D707}_{0}p_{0e}/B^{2}\ll 1$
, we obtain the following simpler set. The electron continuity equation (6.2) becomes:
$$\begin{eqnarray}\displaystyle & & \displaystyle n-\frac{\unicode[STIX]{x1D711}}{\hat{\unicode[STIX]{x1D714}}}+s^{2}\!\left[(1-1.31f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)+0.34(1-0.69f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{\unicode[STIX]{x1D705}s}{\hat{\unicode[STIX]{x1D714}}}\frac{\text{d}}{\text{d}s}\left[\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle (1.67p-1.19t_{e}-3.12t_{i})-\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle (p_{i}+\unicode[STIX]{x1D711})\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle t_{i}\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,1.67f_{t}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+1.02\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}+0.02t_{e}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,3.01\left(\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\right)^{2}\frac{f_{t}}{\langle B^{2}\rangle ^{2}}\left\langle \frac{I^{2}}{R^{2}}\right\rangle ^{2}\frac{\text{d}^{2}}{\text{d}s^{2}}(p+1.39t_{e}-1.17t_{i})=0,\end{eqnarray}$$
the electron thermal equation (6.3) becomes:
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{3}{2}\left[p_{e}-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D711}\right]+3.18s^{2}\left[(1-1.11f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.58(1-0.31f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\vphantom{\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{5}{2}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}s\frac{\text{d}}{\text{d}s}\left[\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle (4.18p-2.36t_{e}-7.86t_{i})\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.-\,\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle (p_{i}+\unicode[STIX]{x1D711})+\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle t_{i}\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,2.97f_{t}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+1.25\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}+0.25t_{e}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,3.35\left(\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\right)^{2}\frac{f_{t}}{\langle B^{2}\rangle ^{2}}\left\langle \frac{I^{2}}{R^{2}}\right\rangle ^{2}\frac{\text{d}^{2}}{\text{d}s^{2}}(p-0.4t_{e}-1.17t_{i})=0\end{eqnarray}$$
the ion thermal equation (6.4) becomes:
and Ampère’s law (6.5) becomes:
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{\text{d}^{2}}{\text{d}s^{2}}(s\unicode[STIX]{x1D713})=\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\langle R^{-2}\rangle }\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \left(1.67\frac{\text{d}p}{\text{d}s}-1.19\frac{\text{d}t_{e}}{\text{d}s}-1.95\frac{\text{d}t_{i}}{\text{d}s}\right)\nonumber\\ \displaystyle & & \displaystyle \quad -\,\left(\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\right)^{2}\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{I^{2}\langle R^{-2}\rangle ^{2}}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle s\left[(1-1.31f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.34(1-0.66f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\vphantom{\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)}\right].\end{eqnarray}$$
These approximations do not affect the vorticity equation (6.6).
7 Boundary conditions and an intermediate region
The purpose in solving the above layer equations is to match solutions of a given parity at
$s=0$
to the marginal ideal MHD solutions at large
$s$
, which involve the tearing mode parameter,
$\unicode[STIX]{x1D6E5}^{\prime }$
, to determine the eigenvalue
$\hat{\unicode[STIX]{x1D714}}$
in terms of
$\unicode[STIX]{x1D6E5}^{\prime }$
. In this limit, when
$\hat{\unicode[STIX]{x1D714}}\rightarrow 0$
,
$E_{\Vert }\propto (\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})\rightarrow 0$
,
$p\rightarrow (p_{0}^{\prime }/\hat{\unicode[STIX]{x1D714}}T_{e}n_{0}^{\prime })\unicode[STIX]{x1D713}$
and we can ignore the momentum flux, the solutions of (6.6) should behave as
$\unicode[STIX]{x1D713}\sim s^{\unicode[STIX]{x1D708}}$
,
$\unicode[STIX]{x1D708}=-(1/2)\pm \sqrt{-D_{I}}$
in the limit
$s\rightarrow \infty$
. However, as they stand, they do not lead to
$(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})\rightarrow 0$
. In fact, it is necessary to consider an intermediate region consisting of two sub-layers: (i) a transition layer around
$s=s_{1}\sim f_{t}^{1/2}(m_{i}/m_{e})^{1/4}$
, where an ion contribution to the electrical conductivity enters; and another (ii), around somewhat larger values of
$s$
,
$s=s_{2}\sim (\unicode[STIX]{x1D712}_{\Vert ,e}/\unicode[STIX]{x1D712}_{\Vert ,i})^{1/2}\sim (m_{i}/m_{e})^{1/4}$
, where parallel ion thermal transport forces
$t_{i}\rightarrow (T_{i}/T_{e})(\unicode[STIX]{x1D702}_{i}\unicode[STIX]{x1D713}/\hat{\unicode[STIX]{x1D714}})$
and ensures
$(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})\rightarrow 0$
. Clearly
$s_{1}$
and
$s_{2}$
are not very different and we can treat them together.
Let us first consider the simplification of the governing equations when
$s\gg 1$
.
The electron continuity equation (6.2) becomes:
$$\begin{eqnarray}\displaystyle & & \displaystyle n-\frac{\unicode[STIX]{x1D711}}{\hat{\unicode[STIX]{x1D714}}}+\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle p\left(1-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right)-\unicode[STIX]{x1D705}C\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\nonumber\\ \displaystyle & & \displaystyle \quad +\,s^{2}\left[(1-1.31f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)+0.34(1-0.69f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{\unicode[STIX]{x1D705}s}{\hat{\unicode[STIX]{x1D714}}}\frac{\text{d}}{\text{d}s}\left[\left(\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right)p+\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle (1.67p-1.19t_{e}-3.12t_{i})\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.-\,\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle (p_{i}+\unicode[STIX]{x1D711})+\frac{1.17}{\langle B^{2}\rangle }(1-f_{t})\left\langle \frac{I^{2}}{R^{2}}\right\rangle t_{i}\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\left(\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right)\frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]\nonumber\\ \displaystyle & & \displaystyle \quad +\,1.67f_{t}\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{d}}{\text{d}s}\left[s\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]=0.\end{eqnarray}$$
However, this last equation is dominated by its second term, which requires:
A similar balance, but with different coefficients, appears in the electron thermal equation (6.3), so we can conclude that:
so that (7.1) simplifies further:
$$\begin{eqnarray}\displaystyle & & \displaystyle n-\frac{\unicode[STIX]{x1D711}}{\hat{\unicode[STIX]{x1D714}}}+\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle p\left(1-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right)+s^{2}\left[(1-1.31f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.34(1-0.69f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\vphantom{\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}s\frac{\text{d}}{\text{d}s}\left[\left(\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right)p+\frac{f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle (1.67p-1.19t_{e}-3.12t_{i})\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.-\,\left\langle \frac{I^{2}}{R^{2}B^{2}}\right\rangle (p_{i}+\unicode[STIX]{x1D711})+\frac{1.17}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle t_{i}\right]=0.\end{eqnarray}$$
With the results (7.3), the vorticity equation (6.6) reduces to
$$\begin{eqnarray}\displaystyle & & \displaystyle s\frac{\text{d}^{2}}{\text{d}s^{2}}(s\unicode[STIX]{x1D713})+\hat{\unicode[STIX]{x1D714}}\left(D_{I}+\frac{1}{4}\right)\left[\frac{\unicode[STIX]{x1D713}}{\hat{\unicode[STIX]{x1D714}}}+\frac{\left(1+{\displaystyle \frac{T_{e}}{T_{i}}}\right)(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})+\left(t_{i}-{\displaystyle \frac{T_{i}}{T_{e}}}{\displaystyle \frac{\unicode[STIX]{x1D702}_{i}\unicode[STIX]{x1D713}}{\hat{\unicode[STIX]{x1D714}}}}\right)}{(1+T_{e}/T_{i}+\unicode[STIX]{x1D702}_{e}+\unicode[STIX]{x1D702}_{i}T_{i}/T_{e})}\right]\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\unicode[STIX]{x1D705}}\frac{\langle 1/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle }{\langle R^{-2}\rangle }\hat{\unicode[STIX]{x1D714}}s\frac{\text{d}}{\text{d}s}\left(1+\frac{T_{e}}{T_{i}}\right)(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})+\left(t_{i}-\frac{T_{i}}{T_{e}}\frac{\unicode[STIX]{x1D702}_{i}\unicode[STIX]{x1D713}}{\hat{\unicode[STIX]{x1D714}}}\right)=0,\end{eqnarray}$$
where we have neglected the viscous term, which is valid for
$s>(m_{i}/m_{e})^{3/8}(\hat{\unicode[STIX]{x1D6FF}}_{e}/\unicode[STIX]{x1D708}_{\ast e})^{1/2}$
, with
$\hat{\unicode[STIX]{x1D6FF}}_{e}$
the semi-collisional width normalised to the plasma minor radius and
$\unicode[STIX]{x1D708}_{\ast e}$
the electron collisionality parameter. Ampère’s law (6.5) retains its form.
However, the simplified vorticity equation (7.5) does not reduce to the ideal form and we must consider the intermediate layer:
$s\sim s_{1}$
,
$s_{2}$
. To address this we first calculate the correction to the ion parallel velocity arising from the parallel ion pressure gradient
$-$
this is achieved by modifying (3.7) to give an equation for the correction to the ion distribution function
${\hat{h}}_{1i}$
Using the model ion–ion collision operator (3.13), we can calculate the resulting modification to the ion velocity arising from
${\hat{h}}_{1i}$
, obtaining
where we have used the results in (7.3), which are valid at large
$s$
.
We see that in the absence of the small ion–electron collision frequency, the friction with trapped ions determines the bulk ion velocity. This velocity changes the parallel electron velocity term proportional to
$s^{2}$
appearing in the electron continuity equation (6.2). This additional contribution modifies the behaviour (7.3) at large
$s$
:
$$\begin{eqnarray}\displaystyle & & \displaystyle (1-1.31f_{t})\left(p_{e}-(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)+0.34(1-0.69f_{t})\left(t_{e}-\frac{\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\nonumber\\ \displaystyle & & \displaystyle \quad -\,\unicode[STIX]{x1D6FF}_{i}\left[\left(1+\frac{T_{e}}{T_{i}}\right)(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})-0.17\left(t_{i}-\frac{T_{i}}{T_{e}}\frac{\unicode[STIX]{x1D702}_{i}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\right]=0,\end{eqnarray}$$
where, using results for the collision integrals from appendix A,
In this limit, equation (6.5) reduces to
We can also calculate the corresponding parallel ion heat flux, finding it is dominated by the convective component due to the inverse dependence on the trapped particle fraction of the parallel ion flow in (7.7):
$$\begin{eqnarray}\displaystyle & & \displaystyle \left\langle \frac{\text{i}\ell q^{\prime }x}{R^{2}Bq}Q_{\Vert i}\right\rangle =\frac{5}{2}\frac{f_{c}^{2}}{f_{t}}\left(\frac{\ell Iq^{\prime }x}{q}\right)^{2}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{R^{2}}\right\rangle ^{2}\frac{T_{i}}{m_{i}\{\unicode[STIX]{x1D708}_{ii}\}}\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left(\left(1+\frac{T_{e}}{T_{i}}\right)(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})+\left(t_{i}-\frac{T_{i}}{T_{e}}\frac{\unicode[STIX]{x1D702}_{i}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}\right)\left(\frac{\{\unicode[STIX]{x1D708}_{ii}u^{2}\}}{\{\unicode[STIX]{x1D708}_{ii}\}}-\frac{3}{2}\right)\right)n_{0}T_{i}.\end{eqnarray}$$
This modifies (6.4):
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{2.63}{D}\left(\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}\right)^{2}\frac{m_{i}\unicode[STIX]{x1D70F}_{ei}}{m_{e}\unicode[STIX]{x1D70F}_{i}}\frac{1}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \frac{\text{d}^{2}}{\text{d}s^{2}}\left[\left(\left\langle \frac{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}{B^{2}}\right\rangle \left\langle \frac{1}{R^{2}}\right\rangle +\left\langle \frac{I^{2}}{R^{2}}\right\rangle \left(\left\langle \frac{1}{B^{2}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\right)\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,\frac{0.46f_{t}}{\langle B^{2}\rangle }\left\langle \frac{I^{2}}{R^{2}}\right\rangle \right]t_{i}\nonumber\\ \displaystyle & & \displaystyle \quad +\,t_{i}-\frac{T_{i}}{T_{e}}\frac{\unicode[STIX]{x1D702}_{i}}{\hat{\unicode[STIX]{x1D714}}}\unicode[STIX]{x1D713}=-\frac{5}{3D}\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle \left(1+\frac{T_{i}}{T_{e}}\frac{(1+2\unicode[STIX]{x1D702}_{i})}{\hat{\unicode[STIX]{x1D714}}}\left(1+\frac{T_{i}}{T_{e}}+\unicode[STIX]{x1D702}_{e}+\unicode[STIX]{x1D702}_{i}\frac{T_{i}}{T_{e}}\right)\right)\frac{\unicode[STIX]{x1D713}}{\hat{\unicode[STIX]{x1D714}}}\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{T_{i}}{DT_{e}}\left(1-\frac{(1+\unicode[STIX]{x1D702}_{i})}{\hat{\unicode[STIX]{x1D714}}}+\frac{5}{3}\left(1+\frac{T_{i}}{T_{e}}\right)\left(\frac{\unicode[STIX]{x1D705}}{\hat{\unicode[STIX]{x1D714}}}C+\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle \left(\frac{T_{e}}{T_{i}}+\frac{(1+2\unicode[STIX]{x1D702}_{i})}{\hat{\unicode[STIX]{x1D714}}}\right)\right)\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.-\,1.59\unicode[STIX]{x1D6FF}_{i}s^{2}\vphantom{\left(\frac{T_{e}}{T_{i}}+\frac{(1+2\unicode[STIX]{x1D702}_{i})}{\hat{\unicode[STIX]{x1D714}}}\right)}\right)\left(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713}\right),\end{eqnarray}$$
where
and we have again used evaluations of collision integrals given in appendix A.
The new terms proportional to
$\unicode[STIX]{x1D6FF}_{i}$
dominate (7.10) and (7.12) when
$s\gg s_{1}$
,
$s_{2}$
and require
When all of conditions (7.3) and (7.14a,b ) are satisfied we see that (7.5) does indeed reduce to the marginal ideal MHD equation:
To discuss the transition though the intermediate region we combine (7.5) and (6.5), modifying the latter to take account of (7.8).
Using the expressions (7.3) for
$n$
and
$t_{e}$
to express
$p$
and
$p_{i}$
in terms of
$\unicode[STIX]{x1D711}$
,
$\unicode[STIX]{x1D713}$
and
$t_{i}$
, we obtain
$$\begin{eqnarray}\displaystyle & & \displaystyle \frac{1.17}{\langle B^{2}\rangle }\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\left\langle \frac{B^{2}}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\langle 1/R^{2}\rangle }(1-f_{t})s\frac{\text{d}t_{i}}{\text{d}s}\nonumber\\ \displaystyle & & \displaystyle \quad -\,\frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{\langle 1/R^{2}\rangle }\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\left[\frac{1}{\hat{\unicode[STIX]{x1D714}}}\left(1+\unicode[STIX]{x1D702}_{e}+\frac{T_{i}}{T_{e}}(1+\unicode[STIX]{x1D702}_{i})\right)\left\langle \frac{1}{|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}}\right\rangle +\left[1-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right]\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.\times \,\frac{\langle B^{2}/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle \langle 1/R^{2}B^{2}\rangle }{\langle 1/R^{2}\rangle }\right]\,s\frac{\text{d}\unicode[STIX]{x1D713}}{\text{d}s}\nonumber\\ \displaystyle & & \displaystyle \quad -\,\left[\vphantom{\left(\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\right)^{2}}\frac{(D_{I}+1/4)\hat{\unicode[STIX]{x1D714}}}{(1+\unicode[STIX]{x1D702}_{e}+T_{i}/T_{e}(1+\unicode[STIX]{x1D702}_{i}))}+\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}\langle B^{2}/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle }{I^{2}\langle 1/R^{2}\rangle ^{2}}\left(1-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right)\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}^{2}}\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.+\,0.17\frac{\unicode[STIX]{x1D707}_{0}p_{0e}\langle B^{2}/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle }{I^{2}\langle 1/R^{2}\rangle ^{2}}\left(\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\right)^{2}\unicode[STIX]{x1D6FF}_{i}s^{2}\right]t_{i}\nonumber\\ \displaystyle & & \displaystyle \quad -\,\left[\frac{(D_{I}+1/4)}{(1+\unicode[STIX]{x1D702}_{e}+T_{i}/T_{e}(1+\unicode[STIX]{x1D702}_{i}))}+\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}\langle B^{2}/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle }{I^{2}\langle 1/R^{2}\rangle ^{2}}\left(1-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right)\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}^{2}}\right]\nonumber\\ \displaystyle & & \displaystyle \quad \times \,\left(1+\unicode[STIX]{x1D702}_{e}+\frac{T_{i}}{T_{e}}\right)\unicode[STIX]{x1D713}\nonumber\\ \displaystyle & & \displaystyle \quad +\,0.17\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}^{2}}\frac{\unicode[STIX]{x1D707}_{0}p_{0e}\langle B^{2}/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle }{I^{2}\langle 1/R^{2}\rangle ^{2}}\unicode[STIX]{x1D702}_{i}\frac{T_{i}}{T_{e}}\unicode[STIX]{x1D6FF}_{i}s^{2}\unicode[STIX]{x1D713}-\frac{(D_{I}+1/4)(1+T_{i}/T_{e})\hat{\unicode[STIX]{x1D714}}}{(1+\unicode[STIX]{x1D702}_{e}+T_{i}/T_{e}(1+\unicode[STIX]{x1D702}_{i}))}(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713})\nonumber\\ \displaystyle & & \displaystyle \quad -\,\left[1-\frac{1}{\hat{\unicode[STIX]{x1D714}}}+\left\langle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}}{B^{2}}\right\rangle \frac{\unicode[STIX]{x1D707}_{0}p_{0e}\langle B^{2}/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle }{I^{2}\langle 1/R^{2}\rangle ^{2}}\left(1-\frac{1+\unicode[STIX]{x1D702}_{e}}{\hat{\unicode[STIX]{x1D714}}}\right)\left(\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\right)^{2}\right.\nonumber\\ \displaystyle & & \displaystyle \quad \left.-\,\frac{\unicode[STIX]{x1D707}_{0}p_{0e}\langle B^{2}/|\unicode[STIX]{x1D735}\unicode[STIX]{x1D712}|^{2}\rangle }{I^{2}\langle 1/R^{2}\rangle ^{2}}\left(\frac{\hat{\unicode[STIX]{x1D714}}}{\unicode[STIX]{x1D705}}\right)^{2}\left(1+\frac{T_{e}}{T_{i}}\right)\unicode[STIX]{x1D6FF}_{i}s^{2}\right]\left(\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713}\right)=0.\end{eqnarray}$$
Then (7.5) and (7.16), with
$t_{i}$
given by (7.12), provide a fourth-order system of equations for
$\unicode[STIX]{x1D711}$
and
$\unicode[STIX]{x1D713}$
to connect the semi-collisional electron layer solutions to the ideal MHD region through the region
$s\sim s_{1}$
,
$s_{2}$
. The boundary condition at large
$s$
is to match to the correct ratio of large and small solutions arising from the ideal region and ensure
$\unicode[STIX]{x1D711}-\unicode[STIX]{x1D713}\rightarrow 0$
. Thus, the jump in the ratio of large to small solutions through this layer, and hence the layer
$\unicode[STIX]{x1D6E5}^{\prime }$
, can be computed.
8 Conclusions
We have derived a set of equations to describe the linear stability of semi-collisional, neoclassical toroidal plasma in general axisymmetric geometry, albeit provided that: (i) the fraction of trapped particles is small to justify the use of a model, pitch-angle scattering, collision operator; and (ii) that the ions are magnetised. The assumption that the ions are magnetised may require low magnetic shear or
$T_{i}\ll T_{e}$
, otherwise a non-local model for the ion response will be needed (Fitzpatrick Reference Fitzpatrick1989). A related constraint on the ion parameters follows from the ion drift orbit expansion of the ion gyro-kinetic equation that we invoked in § 3.1. This requires that
$k_{r}\unicode[STIX]{x1D70C}_{i}q/f_{t}<1$
, compared to the somewhat lesser constraint,
$k_{r}\unicode[STIX]{x1D70C}_{i}<1$
, from the condition on the ion Larmor orbits and implies the following condition on the tokamak parameters:
The set of equations comprises a pair of second-order radial differential equations for the electron density,
${\tilde{n}}$
(or pressure,
$\tilde{p}_{e}$
) and temperature,
$\tilde{T}_{e}$
perturbations (4.59) and (4.60) and one for the ion temperature perturbation,
$\tilde{T}_{i}$
, equation (4.61), in terms of the perturbed electrostatic potential,
$\unicode[STIX]{x1D6F7}$
, and parallel vector potential,
$\tilde{A}_{\Vert }$
, described by the potential
$\unicode[STIX]{x1D6F9}$
, with these two quantities in turn being determined by Ampère’s law, equation (5.4), with (5.5) for the parallel current, and the vorticity equation, equation (5.13). The perturbed parallel magnetic field is given simply by (4.1).
The analysis of neoclassical electron physics utilised the Spitzer functions (Spitzer & Harm Reference Spitzer and Harm1953), where we remark that parallel gradients in perturbed electron temperature that have no counterpart in standard neoclassical theory (Helander & Sigmar Reference Helander and Sigmar2002), necessitate the introduction of the second Spitzer function corresponding to parallel electron thermal conduction. The vorticity equation requires a calculation of radial collisional transport of toroidal angular momentum and we have needed to generalise the treatment by Wong & Chan (Reference Wong and Chan2005) to arbitrary axisymmetric geometry, as in appendix B, although still assuming a small number of trapped particles. This calculation employed the model ion collision operator. Analytical evaluations of the coefficients in the large aspect ratio limit are presented there; these are consistent with the results of Wong & Chan (Reference Wong and Chan2005).
The introduction of general toroidal equilibria and the use of the Spitzer functions extend the treatment of electron neoclassical physics given by Fitzpatrick (Reference Fitzpatrick1989), as well as providing a more consistent treatment of the neoclassical radial transport than in the work of Connor et al. (Reference Connor, Hastie and Helander2009). The general geometry aspect also means we have needed to include classical collisional transport.
The resultant equations, summarised in normalised form in (6.2)–(6.7), are equivalent to a twelfth-order system of radial differential equations if all effects are retained – it reduces to tenth order if we neglect the radial angular momentum transport. However, the relatively large ion thermal diffusivity means one can treat the ion temperature as a constant over the semi-collisional layer. (This also greatly simplifies the expression for the radial transport of toroidal angular momentum given in (5.19).) The system of equations then reduces to tenth order and one only needs to solve a simplified version of the ion thermal equation in the intermediate region,
$s\sim (m_{i}/m_{e})^{1/4}$
, where the system reduces further to a fourth-order set. A simpler version of (6.2)–(6.5) in which we ignore the subdominant classical and Pfirsch–Schlüter fluxes relative to the banana contributions, the smaller azimuthal fluxes and effects proportional to
$\unicode[STIX]{x1D6FD}_{e}=2\unicode[STIX]{x1D707}_{0}p_{0e}/B^{2}\ll 1$
, is presented in (6.8)–(6.11). The vorticity equation (6.6) is unaffected by these approximations.
The solution of these equations in the narrow radial layer around a low-order resonant surface needs to be matched to ‘external’ solutions of the marginal ideal MHD equations. However, to achieve this matching, as pointed out by Fitzpatrick (Reference Fitzpatrick1989), we need to consider the intermediate layer where small corrections arising from the ion contribution to the electrical conductivity and the parallel ion thermal diffusivity enter and ensure that the perturbed parallel electric field vanishes. Since this correction is determined by the friction of the passing ion population with the trapped ones, rather than the schematic ion sound model suggested by Fitzpatrick (Reference Fitzpatrick1989), we were able to provide an explicit form for it. This intermediate region is described by a fourth-order set of (7.5) and (7.12) together with (7.16), that allow a proper matching to the ideal MHD region. Continuing the set of equations through this layer allows us to obtain a dispersion relation
where
$\unicode[STIX]{x1D6E5}^{\prime }$
is the toroidal tearing mode stability parameter (Glasser et al.
Reference Glasser, Greene and Johnson1975) and
$\hat{\unicode[STIX]{x1D6E5}}(\unicode[STIX]{x1D714}+\text{i}\unicode[STIX]{x1D6FE})$
is obtained from the solution of the layer equations, allowing for the effect of the intermediate layer.
Glasser, Wang & Park (Reference Glasser, Wang and Park2016) have pointed out that, in the case of the simple resistive MHD layer model, the requirement that there be a sufficient overlap region to allow the successful asymptotic matching of the solution of the inner layer equations to the outer ideal MHD solution, imposes rather a stringent requirement on the Lundquist number,
$S$
, namely
$S\sim 10^{9}$
. In the present case, one needs to ensure that the width of the ‘ion layer’,
$\unicode[STIX]{x1D6FF}_{i}$
, which exceeds the semi-collisional width
$\unicode[STIX]{x1D6FF}_{e}$
by a factor
$\sqrt{m_{i}/m_{e}}$
, is much less than the plasma minor radius,
$a$
. Using the definition (6.1) for
$\unicode[STIX]{x1D6FF}_{e}$
, this can be cast in the form:
where
$\unicode[STIX]{x1D708}_{\ast e}$
is the electron collisionality parameter and
$\unicode[STIX]{x1D70C}_{e}$
the electron Larmor radius. Typical parameters for an ITER-like plasma,
$\unicode[STIX]{x1D708}_{\ast e}\sim 10^{-2}$
,
$\unicode[STIX]{x1D70C}_{e}/a\sim 10^{-4}$
, yield
$\unicode[STIX]{x1D6FF}_{i}/a\sim 10^{-3}$
, which should provide sufficient overlap for the matching. However, this conflicts with the requirements for magnetised ions regarding the scalings with magnetic shear, ion–electron temperature ratio and collisionality: satisfying both requires sufficiently small normalised electron Larmor radius.
Although our analysis is linear, the treatment of the semi-collisional, neoclassical electron physics could be generalised to describe the evolution of nonlinear neoclassical tearing mode islands, extending the analysis given by Wilson et al. (Reference Wilson, Connor, Hastie and Hegna1996) for magnetised ions to this more collisional regime, or be used in conjunction with a numerical treatment for ions when the width of the poloidal ion Larmor orbit is comparable to the island width (Imada, Connor & Wilson Reference Imada, Connor and Wilson2016). Alternatively, it could serve to incorporate neoclassical physics in the collisional model of Smolyakov (Reference Smolyakov1993).
Appendix A. Some collisional integrals
Here we list numerical evaluations of the various integrals, including those involving the Spitzer functions
$D^{(1)}$
and
$D^{(2)}$
tabulated in Spitzer & Harm (Reference Spitzer and Harm1953) (to be precise,
$D^{(1)}=-D/A$
and
$D^{(2)}=2D/B$
as given in (Spitzer & Harm Reference Spitzer and Harm1953, Tables I and II)):
$$\begin{eqnarray}\left.\begin{array}{@{}c@{}}\{\unicode[STIX]{x1D708}_{e}\}=1.153\unicode[STIX]{x1D708}_{0e},\quad \{u^{2}\unicode[STIX]{x1D708}_{e}\}=1.284\unicode[STIX]{x1D708}_{0e},\quad \{u^{4}\unicode[STIX]{x1D708}_{e}\}=2.701\unicode[STIX]{x1D708}_{0e},\\ \left\{{\displaystyle \frac{D^{(1)}(u)}{u}}\right\}=-2.616,\quad \left\{{\displaystyle \frac{D^{(2)}(u)}{u}}\right\}=-0.898,\quad \{uD^{(1)}(u)\}=-8.297,\\ \{uD^{(2)}(u)\}=-4.782,\quad \left\{\unicode[STIX]{x1D708}_{e}{\displaystyle \frac{D^{(1)}(u)}{u}}\right\}=-1.666\unicode[STIX]{x1D708}_{0e},\quad \left\{\unicode[STIX]{x1D708}_{e}{\displaystyle \frac{D^{(2)}(u)}{u}}\right\}=-0.040\unicode[STIX]{x1D708}_{0e},\\ \{\unicode[STIX]{x1D708}_{e}uD^{(1)}(u)\}=-2.972\unicode[STIX]{x1D708}_{0e},\quad \{\unicode[STIX]{x1D708}_{e}uD^{(2)}(u)\}=-0.738\unicode[STIX]{x1D708}_{0e},\\ \left\{\unicode[STIX]{x1D708}_{e}\left({\displaystyle \frac{D^{(1)}(u)}{u}}\right)^{2}\right\}=3.432\unicode[STIX]{x1D708}_{0e},\quad \left\{\unicode[STIX]{x1D708}_{e}{\displaystyle \frac{D^{(1)}(u)D^{(2)}(u)}{u^{2}}}\right\}=0.605\unicode[STIX]{x1D708}_{0e},\\ \left\{\unicode[STIX]{x1D708}_{e}\left({\displaystyle \frac{D^{(2)}(u)}{u}}\right)^{2}\right\}=0.566\unicode[STIX]{x1D708}_{0e},\\ \{\unicode[STIX]{x1D708}_{ii}\}=0.401\unicode[STIX]{x1D708}_{0i},\quad \{u^{2}\unicode[STIX]{x1D708}_{ii}\}=0.532\unicode[STIX]{x1D708}_{0i},\quad \{u^{4}\unicode[STIX]{x1D708}_{ii}\}=1.19\unicode[STIX]{x1D708}_{0i},\\ \{u^{2}/\unicode[STIX]{x1D708}_{ii}\}=20.62/\unicode[STIX]{x1D708}_{0i},\quad \{u^{4}/\unicode[STIX]{x1D708}_{ii}\}=100.37/\unicode[STIX]{x1D708}_{0i}.\end{array}\right\}\end{eqnarray}$$
Appendix B. Radial toroidal angular momentum transport
There are two contributions from the radial transport of momentum appearing in the vorticity equation (5.13): the classical one,
$\langle \unicode[STIX]{x1D6F1}^{C}\rangle$
, and the neoclassical one,
$\langle \unicode[STIX]{x1D6F1}^{NC}\rangle$
. We evaluate these in this appendix.
B.1 Classical radial angular momentum transport
The expression for
$\text{d}^{2}\langle \unicode[STIX]{x1D6F1}^{C}\rangle /\text{d}x^{2}$
involves
$\langle \sum _{j}e\int \,\text{d}^{3}\boldsymbol{v}L_{j}^{2}C_{j}(g_{0j}L_{j}^{2})\rangle$
, where the sum is dominated by the ions. Thus, substituting for
$g_{0i}$
from (2.8) and (3.8) and recalling the definition of
$L_{i}$
below (2.2), we obtain
The collision integrals in (B 1) can be evaluated using the entropy functional
$S(\hat{f}(\boldsymbol{v}^{\prime }),{\hat{g}}(\boldsymbol{v}))$
(Helander & Sigmar Reference Helander and Sigmar2002), with
$\hat{f}=v_{\Vert }^{\prime 3}$
and
${\hat{g}}=v_{\Vert }(m_{i}v^{2}/2T_{i}-5/2)$
. The quantity
$S$
can be easily calculated using Cartesian coordinates in velocity space, labelling the direction parallel to the magnetic field as the
$x$
-direction, and introducing
$\boldsymbol{u}=(\boldsymbol{v}^{\prime }-\boldsymbol{v})/2$
and
$\boldsymbol{w}=(\boldsymbol{v}^{\prime }+\boldsymbol{v})/2$
to execute the velocity space integrations. The result is
B.2 Neoclassical radial angular momentum transport
Equation (5.10) requires the evaluation of
representing the radial transport of the toroidal angular momentum arising from the parallel flow:
$Rv_{\unicode[STIX]{x1D711}}=Iv_{\Vert }/B$
. Noting that it is dominated by the ion contribution, we express this quantity as
To avoid the need to calculate
$h_{2i}$
, we follow Wong & Chan (Reference Wong and Chan2005) in defining the adjoint function,
$g$
:
so that, using the self-adjointness of the collision operator, integrations by parts and the gyro-kinetic equation for
$h_{2i}$
, we obtain
We use the gyro-kinetic equation for
$h_{2i}$
and (B 5) for
$g$
, which we expand for weak collisions, so that we can integrate it to give
where
$G$
is given by the collisional constraint that follows from (B 5):
Thus
$$\begin{eqnarray}\displaystyle g & = & \displaystyle \frac{Im_{i}^{2}v^{2}}{2e}\left(\frac{\unicode[STIX]{x1D706}}{B}-\int _{0}^{\unicode[STIX]{x1D706}}\,\text{d}\unicode[STIX]{x1D706}\frac{\langle v_{\Vert }/B\rangle }{\langle v_{\Vert }\rangle }\right)f_{0i}\nonumber\\ \displaystyle & = & \displaystyle \frac{Im_{i}^{2}v^{2}}{2e}\left(\unicode[STIX]{x1D706}\left(\frac{1}{B}-\left\langle \frac{1}{B}\right\rangle \right)-\int _{0}^{\unicode[STIX]{x1D706}}\,\text{d}\unicode[STIX]{x1D706}\frac{\langle v_{\Vert }/B-v_{\Vert }\langle 1/B\rangle \rangle }{\langle v_{\Vert }\rangle }\right)f_{0i},\end{eqnarray}$$
where the final integral is small in the trapped particle fraction.
Performing further integrations by parts on (B 6), we finally obtain
where
and
where,
$h_{1i}^{\prime }$
is given by (3.14).
It is helpful to re-write (B 11) as
The first term can be evaluated by using the model collision operator (3.13) acting on
$h_{1i}$
given in (4.7). Since the ion–ion collision operator vanishes when acting on a displaced Maxwellian distribution, only the perturbed temperature gradient terms in expressions (4.9) and (4.10) contribute. First, we consider the banana contribution (4.10). Integrating by parts in
$\unicode[STIX]{x1D706}$
, using the results (3.16), (3.18) and the evaluations in appendix A, we obtain
where
For the Pfirsch–Schlüter contribution we apply the collision operator directly to (4.9). The result is
Thus, the total contribution to
$\langle \unicode[STIX]{x1D6F1}^{NC}\rangle ^{A1}$
is given by the sum of results (B 14) and (B 16):
Equation (B 13) also has a contribution from the function
$g$
, which is labelled
$\langle \unicode[STIX]{x1D6F1}^{NC}\rangle ^{A2}$
. The contribution to this from the banana term in
$h_{1i}$
, equation (4.10), can be evaluated using (B 9) for
$g$
. We employ the model collision operator, recalling the banana contribution to
$U_{\Vert i}^{\ast }=0$
, and integrate by parts in
$\unicode[STIX]{x1D706}$
twice. Since
$h_{1i}^{\text{Ban}}$
is localised, we can approximate
$\unicode[STIX]{x2202}(\unicode[STIX]{x1D706}v_{\Vert }^{2}\unicode[STIX]{x2202}g/\unicode[STIX]{x2202}\unicode[STIX]{x1D706})/\unicode[STIX]{x2202}\unicode[STIX]{x1D706}\cong -Bv^{2}\unicode[STIX]{x1D706}\unicode[STIX]{x2202}g/\unicode[STIX]{x2202}\unicode[STIX]{x1D706}$
. The integral term in expression (B 9) for
$g$
only contributes significantly as
$\unicode[STIX]{x1D706}$
enters the trapping region, but remains small (i.e.
$0(\unicode[STIX]{x1D6FF}\unicode[STIX]{x1D706})$
compared to the first term, where
$\unicode[STIX]{x1D6FF}\unicode[STIX]{x1D706}$
is the trapped width in
$\unicode[STIX]{x1D706}$
) and can be ignored. Finally, inserting expression (4.10) for
$h_{1i}^{\text{Ban}}$
, performing the velocity space integrals recalling the results in appendix A, and changing the order of the
$\unicode[STIX]{x1D706}$
integration and the
$\langle \,\rangle$
averaging operation, we obtain
where
$$\begin{eqnarray}\displaystyle \frac{G^{(1)}}{3\langle B^{2}\rangle } & = & \displaystyle \frac{15}{4}\int _{0}^{1/B_{\text{Min}}}\,\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}\frac{\left\langle B\sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\left\langle \frac{\sqrt{1-\unicode[STIX]{x1D706}B}}{B}\right\rangle \nonumber\\ \displaystyle & & \displaystyle +\,\frac{15}{16}\langle B^{2}\rangle \left\langle \frac{1}{B}\right\rangle \int _{0}^{1/B_{\text{Max}}}\frac{\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}^{2}}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\nonumber\\ \displaystyle & & \displaystyle -\,\frac{15}{8}\langle B^{2}\rangle \int _{0}^{1/B_{\text{Max}}}\,\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}\frac{1}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\left\langle \frac{\sqrt{1-\unicode[STIX]{x1D706}B}}{B}\right\rangle \int _{\unicode[STIX]{x1D706}}^{1/B_{\text{Max}}}\frac{\text{d}\unicode[STIX]{x1D706}^{\prime }}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}^{\prime }B}\right\rangle }.\nonumber\\ \displaystyle & & \displaystyle\end{eqnarray}$$
The calculation of the Pfirsch–Schlüter contribution from (4.9) is more straightforward. We evaluate the collision operator acting on
$h_{1i}^{PS}$
, again neglect the small integral tern in
$g$
and evaluate the velocity space integrals using the results in appendix A. The result is:
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6F1}^{PS}\rangle ^{A2} & = & \displaystyle 0.37\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[\left\langle \frac{1}{B^{4}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{B^{2}}\right\rangle \right.\nonumber\\ \displaystyle & & \displaystyle \left.-\,\left\langle \frac{1}{B}\right\rangle \left(\left\langle \frac{1}{B^{3}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{B}\right\rangle \right)\right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right).\end{eqnarray}$$
Thus, the total contribution to
$\langle \unicode[STIX]{x1D6F1}^{NC}\rangle ^{A2}$
from results (B 18) and (B 20) is
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6F1}^{NC}\rangle ^{A2} & = & \displaystyle 0.37\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[\left\langle \frac{1}{B^{4}}\right\rangle +\frac{3}{\langle B^{2}\rangle }\left\langle \frac{1}{B^{2}}\right\rangle \right.\nonumber\\ \displaystyle & & \displaystyle \left.-\left\langle \frac{1}{B^{3}}\right\rangle \left\langle \frac{1}{B}\right\rangle -\frac{G^{(1)}}{\langle B^{2}\rangle ^{2}}\right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right).\end{eqnarray}$$
Combining the results from (B 17) and (B 21), we finally obtain
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6F1}^{NC}\rangle ^{A} & = & \displaystyle 0.37\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[2\left\langle \frac{1}{B^{4}}\right\rangle -\left\langle \frac{1}{B^{3}}\right\rangle \left\langle \frac{1}{B}\right\rangle +\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{B^{2}}\right\rangle \right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,\frac{(G^{(0)}-G^{(1)})}{\langle B^{2}\rangle ^{2}}\right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right).\end{eqnarray}$$
Turning to the quantity
$\langle \unicode[STIX]{x1D6F1}^{NC}\rangle ^{B}$
, we first observe that the periodicity constraint (B 8) allows us to rewrite (B 12) as
We introduce the three contributions to
$h_{1_{i}}$
from (4.7), use the model ion–ion collision operator (3.13), substitute
as follows from (B 9), and integrate by parts in
$\unicode[STIX]{x1D706}$
. The contribution from
$h_{1}^{\text{Flow}}$
is:
where
that from
$h_{1}^{PS}$
is:
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6F1}^{PS}\rangle ^{B} & = & \displaystyle 0.80\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left(\left\langle \frac{1}{B^{4}}\right\rangle -\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{B^{2}}\right\rangle +\frac{G^{(3)}}{\langle B^{2}\rangle ^{2}}\right)\frac{\text{d}^{2}}{\text{d}x^{2}}\nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{e\unicode[STIX]{x1D6F7}}{T_{i}}-0.25\frac{\tilde{T}_{i}}{T_{i}}\right),\end{eqnarray}$$
where
and from
$h_{1}^{\text{Ban}}$
is:
where
$$\begin{eqnarray}\displaystyle G^{(4)} & = & \displaystyle \frac{15\langle B^{2}\rangle ^{2}}{4}\int _{0}^{1/B_{\text{Max}}}\,\text{d}\unicode[STIX]{x1D706}\unicode[STIX]{x1D706}\frac{1}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\nonumber\\ \displaystyle & & \displaystyle \times \,\left(\frac{1}{\left\langle \sqrt{1-\unicode[STIX]{x1D706}B}\right\rangle }\left\langle \frac{\sqrt{1-\unicode[STIX]{x1D706}B}}{B}\right\rangle \left\langle \frac{1-\unicode[STIX]{x1D706}B}{B}\right\rangle -\left\langle \frac{1-\unicode[STIX]{x1D706}B}{B^{2}}\right\rangle \right).\end{eqnarray}$$
Consequently, combining results (B 25), (B 27) and (B 29), we obtain the result
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6F1}^{NC}\rangle ^{B} & = & \displaystyle 0.80\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[\left\langle \frac{1}{B^{4}}\right\rangle +\frac{\left(G^{(3)}-G^{(2)}\right)}{\left\langle B^{2}\right\rangle ^{2}}\right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{e\unicode[STIX]{x1D6F7}}{T_{i}}\right)\nonumber\\ \displaystyle & & \displaystyle -\,\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[0.2\left\langle \frac{1}{B^{4}}\right\rangle \right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,\frac{(0.74(G^{(1)}/3+G^{(4)})-0.94G^{(2)}+0.2G^{(3)})}{\langle B^{2}\rangle ^{2}}\right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right).\end{eqnarray}$$
Finally, combining results (B 22) and (B 31), we obtain the expression for the neoclassical toroidal angular viscosity:
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6F1}^{NC}\rangle & = & \displaystyle 0.80\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[\left\langle \frac{1}{B^{4}}\right\rangle +\frac{(G^{(3)}-G^{(2)})}{\langle B^{2}\rangle ^{2}}\right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{e\unicode[STIX]{x1D6F7}}{T_{i}}\right)\nonumber\\ \displaystyle & & \displaystyle +\,\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}\langle B^{2}\rangle }\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2}\left[0.37\left(\frac{1}{\langle B^{2}\rangle }\left\langle \frac{1}{B^{2}}\right\rangle -\left\langle \frac{1}{B}\right\rangle \left\langle \frac{1}{B^{3}}\right\rangle \right)\right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,0.54\left\langle \frac{1}{B^{4}}\right\rangle \right]\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right)\nonumber\\ \displaystyle & & \displaystyle +\,\frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}\langle B^{2}\rangle ^{2}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}}\right)^{2} (0.37G^{(0)}-0.62G^{(1)}+0.94G^{(2)}-0.2G^{(3)}\nonumber\\ \displaystyle & & \displaystyle -\,0.74G^{(4)} )\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right).\end{eqnarray}$$
Result (B 32) is the expression appearing in (5.19).
It is interesting to consider the large aspect ratio limit with
$B=B_{0}(1-\unicode[STIX]{x1D700}\cos \unicode[STIX]{x1D703})$
,
$\unicode[STIX]{x1D700}=r/R_{0}\ll 1$
. Analytical evaluation of the coefficients in (B 32) in this limit yields
$$\begin{eqnarray}\displaystyle \langle \unicode[STIX]{x1D6F1}^{NC}\rangle & = & \displaystyle \frac{n_{0}e}{\unicode[STIX]{x1D70F}_{i}}\left(\frac{I^{2}m_{i}T_{i}}{\text{e}^{2}B_{0}^{2}}\right)^{2}\left[0.8\unicode[STIX]{x1D700}^{2}\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{p}_{i}}{p_{0i}}+\frac{e\unicode[STIX]{x1D6F7}}{T_{i}}\right)-0.74\unicode[STIX]{x1D700}^{3/2}\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right)\right.\nonumber\\ \displaystyle & & \displaystyle \left.+\,0(\unicode[STIX]{x1D700}^{2})\frac{\text{d}^{2}}{\text{d}x^{2}}\left(\frac{\tilde{T}_{i}}{T_{i}}\right)\right],\end{eqnarray}$$
in agreement with Wong & Chan (Reference Wong and Chan2005).