返回词条

漂移波 · 历史版本

版本编辑者时间修改说明
第 4 版站方导入2026/11/9 06:56:42百科字段核查:fields.content
第 3 版站方导入2026/10/15 23:34:46按统一模板整理并保留字段来源
第 2 版站方导入2026/10/10 19:10:46根据原始来源更新词条
第 1 版站方导入2026/10/10 19:00:12根据原始来源更新词条

第 2 版

漂移波

Drift waves

The archetypal application of the drift model, the drift wave, does not involve EdE_{d} and propagates when Ti=0T_{i}=0 . The simplest drift-wave dispersion relation involves only ω∗e\omega_{*e} , which enters through E×BE\times B convection across the equilibrium density gradient.

The classical drift wave, sometimes called the “electron drift wave,” is an electrostatic perturbation,

E=−∇Φ\mathbf {E} = - \nabla \Phi

whose frequency satisfies

k∥vti≪ω≪k∥vte.(5.68)k _ {\parallel} v _ {t i} \ll \omega \ll k _ {\parallel} v _ {t e}.\tag{5.68}

We show that such a disturbance propagates, using the simplest version of the drift model, with vanishing ion temperature,

Ti=0.T _ {i} = 0.

The more general case is considered presently. We use Cartesian coordinates (x,y,z)(x,y,z) , with constant magnetic field in the zz -direction:

B=z^B=constant.\mathbf {B} = \hat {\mathbf {z}} B = \mathrm{constant}.

The equilibrium is assumed to vary only with x; in that case perturbations will depend on y and z through the familiar eikonal, so that we have, in particular.

n=n0(x)+n1(x)e−iωt+ik⊥y+ik∥z.n = n _ {0} (x) + n _ {1} (x) e ^ {- i \omega t + i k _ {\perp} y + i k _ {\parallel} z}.

where n1n_1 is the linear perturbation. We suppose that E∥\mathbf{E}_{\parallel} and EE\mathbf{E}_E vanish in equilibrium and for simplicity omit collisional effects.

Propagation of the electron drift wave depends upon the nature of the electron response to an electrostatic potential: its relaxation to a Maxwell-Boltzmann equilibrium. This so-called “adiabatic response,” resulting from fast electron motion through the nearly stationery potential ( ω≪k∥vte\omega \ll k_{\parallel}v_{te} ), is described by the parallel Olm's law, (5.64). In the absence of collisions the electron temperature is constant along field lines; then (5.64) can be seen to imply

n1n0=eΦTe.(5.69)\frac {n _ {1}}{n _ {0}} = \frac {e \Phi}{T _ {e}}.\tag{5.69}

Here and below we suppress the 1-subscript on Φ\Phi (since Φ0=0\Phi_0 = 0 ) and the 0-subscript on TeT_e (since Te1=0T_{e1} = 0 ). We emphasize that there is no counterpart to this relaxation—no allowance for adiabatic electron response—in MHD.

The ions, on the other hand, do not see a quasi-static potential and respond very differently. To understand their (quasineutral) density perturbation, note first that

∇⋅VE=0,\nabla \cdot \mathbf {V} _ {E} = 0,

for a uniform magnetic field. Hence the (nonlinear) particle conservation law becomes simply

∂n∂t+(VE+V∥)⋅∇n+n∇∥V∥=0,\frac {\partial n}{\partial t} + \left(\mathbf {V} _ {E} + \mathbf {V} _ {\parallel}\right) \cdot \nabla n + n \nabla_ {\parallel} V _ {\parallel} = 0,

or, after linearization,

−iωn1+ik∥n0V∥1+iω∗nen0eΦTe=0,(5.70)- i \omega n _ {1} + i k _ {\parallel} n _ {0} V _ {\parallel 1} + i \omega_ {* n e} n _ {0} \frac {e \Phi}{T _ {e}} = 0,\tag{5.70}

Here we noted that

VE⋅\n0=iω∗nen0eΦTe.\mathbf {V} _ {E} \cdot \backslash n _ {0} = i \omega_ {* n e} n _ {0} \frac {e \Phi}{T _ {e}}.

where

ω∗nϵ=−k⊥TeeBdlog⁡n0dx\omega_ {* n \epsilon} = - \frac {k _ {\perp} T _ {e}}{e B} \frac {d \log n _ {0}}{d x}

is the drift frequency defined with the density gradient, rather than the pressure gradient. Thus, as remarked previously, the drift frequency enters through E×BE \times B motion.

Our description is completed by the parallel equation of motion, whose linear version is

−iωmin0V∥1+ik∥pϵ1=0,(5.71)- i \omega m _ {i} n _ {0} V _ {\| 1} + i k _ {\|} p _ {\epsilon 1} = 0,\tag{5.71}

Equations (5.70) and (5.71) combine to yield

(ω−k∥2cs2ω)n1n0=ω∗neeΦTe,(5.72)\left(\omega - \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega}\right) \frac {n _ {1}}{n _ {0}} = \omega_ {* n e} \frac {e \Phi}{T _ {e}},\tag{5.72}

where

cs=Te/mi(5.73)c _ {s} = \sqrt {T _ {e} / m _ {i}}\tag{5.73}

is called the “ion sound speed.” Equation (5.72) is sometimes said to describe a hydrodynamic density response. The parenthesized factor on the left-hand side evidently describes sound waves propagating along the magnetic field.

Finally we combine (5.72) with electron adiabaticity, (5.69), to obtain the drift-wave dispersion relation

ω−ω∗nϵ=k∥2cs2ω,(5.74)\omega - \omega_ {* n \epsilon} = \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega},\tag{5.74}

which implies, in the regime of (5.68), that

ω≃ω∗ne+k∥2cs2ω∗ne.\omega \simeq \omega_ {* n e} + \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega_ {* n e}}.

Drift waves can be destabilized by a variety of effects. usually involving dissipation. The most common drift instabilities are characterized by a modified electron response, including a contribution out of phase with Φ\Phi . Thus (5.69) is replaced by

n1n0=eΦTϵ(1−iΔ)\frac {n _ {1}}{n _ {0}} = \frac {e \Phi}{T _ {\epsilon}} (1 - i \Delta)

where Δ\Delta reflects some dissipative process, such as resistivity, in Ohm's law. It is clear from (5.72) that positive Δ\Delta corresponds to instability.

Drift waves

The archetypal application of the drift model, the drift wave, does not involve EdE_{d} and propagates when Ti=0T_{i}=0 . The simplest drift-wave dispersion relation involves only ω∗e\omega_{*e} , which enters through E×BE\times B convection across the equilibrium density gradient.

The classical drift wave, sometimes called the “electron drift wave,” is an electrostatic perturbation,

E=−∇Φ\mathbf {E} = - \nabla \Phi

whose frequency satisfies

k∥vti≪ω≪k∥vte.(5.68)k _ {\parallel} v _ {t i} \ll \omega \ll k _ {\parallel} v _ {t e}.\tag{5.68}

We show that such a disturbance propagates, using the simplest version of the drift model, with vanishing ion temperature,

Ti=0.T _ {i} = 0.

The more general case is considered presently. We use Cartesian coordinates (x,y,z)(x,y,z) , with constant magnetic field in the zz -direction:

B=z^B=constant.\mathbf {B} = \hat {\mathbf {z}} B = \mathrm{constant}.

The equilibrium is assumed to vary only with x; in that case perturbations will depend on y and z through the familiar eikonal, so that we have, in particular.

n=n0(x)+n1(x)e−iωt+ik⊥y+ik∥z.n = n _ {0} (x) + n _ {1} (x) e ^ {- i \omega t + i k _ {\perp} y + i k _ {\parallel} z}.

where n1n_1 is the linear perturbation. We suppose that E∥\mathbf{E}_{\parallel} and EE\mathbf{E}_E vanish in equilibrium and for simplicity omit collisional effects.

Propagation of the electron drift wave depends upon the nature of the electron response to an electrostatic potential: its relaxation to a Maxwell-Boltzmann equilibrium. This so-called “adiabatic response,” resulting from fast electron motion through the nearly stationery potential ( ω≪k∥vte\omega \ll k_{\parallel}v_{te} ), is described by the parallel Olm's law, (5.64). In the absence of collisions the electron temperature is constant along field lines; then (5.64) can be seen to imply

n1n0=eΦTe.(5.69)\frac {n _ {1}}{n _ {0}} = \frac {e \Phi}{T _ {e}}.\tag{5.69}

Here and below we suppress the 1-subscript on Φ\Phi (since Φ0=0\Phi_0 = 0 ) and the 0-subscript on TeT_e (since Te1=0T_{e1} = 0 ). We emphasize that there is no counterpart to this relaxation—no allowance for adiabatic electron response—in MHD.

The ions, on the other hand, do not see a quasi-static potential and respond very differently. To understand their (quasineutral) density perturbation, note first that

∇⋅VE=0,\nabla \cdot \mathbf {V} _ {E} = 0,

for a uniform magnetic field. Hence the (nonlinear) particle conservation law becomes simply

∂n∂t+(VE+V∥)⋅∇n+n∇∥V∥=0,\frac {\partial n}{\partial t} + \left(\mathbf {V} _ {E} + \mathbf {V} _ {\parallel}\right) \cdot \nabla n + n \nabla_ {\parallel} V _ {\parallel} = 0,

or, after linearization,

−iωn1+ik∥n0V∥1+iω∗nen0eΦTe=0,(5.70)- i \omega n _ {1} + i k _ {\parallel} n _ {0} V _ {\parallel 1} + i \omega_ {* n e} n _ {0} \frac {e \Phi}{T _ {e}} = 0,\tag{5.70}

Here we noted that

VE⋅\n0=iω∗nen0eΦTe.\mathbf {V} _ {E} \cdot \backslash n _ {0} = i \omega_ {* n e} n _ {0} \frac {e \Phi}{T _ {e}}.

where

ω∗nϵ=−k⊥TeeBdlog⁡n0dx\omega_ {* n \epsilon} = - \frac {k _ {\perp} T _ {e}}{e B} \frac {d \log n _ {0}}{d x}

is the drift frequency defined with the density gradient, rather than the pressure gradient. Thus, as remarked previously, the drift frequency enters through E×BE \times B motion.

Our description is completed by the parallel equation of motion, whose linear version is

−iωmin0V∥1+ik∥pϵ1=0,(5.71)- i \omega m _ {i} n _ {0} V _ {\| 1} + i k _ {\|} p _ {\epsilon 1} = 0,\tag{5.71}

Equations (5.70) and (5.71) combine to yield

(ω−k∥2cs2ω)n1n0=ω∗neeΦTe,(5.72)\left(\omega - \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega}\right) \frac {n _ {1}}{n _ {0}} = \omega_ {* n e} \frac {e \Phi}{T _ {e}},\tag{5.72}

where

cs=Te/mi(5.73)c _ {s} = \sqrt {T _ {e} / m _ {i}}\tag{5.73}

is called the “ion sound speed.” Equation (5.72) is sometimes said to describe a hydrodynamic density response. The parenthesized factor on the left-hand side evidently describes sound waves propagating along the magnetic field.

Finally we combine (5.72) with electron adiabaticity, (5.69), to obtain the drift-wave dispersion relation

ω−ω∗nϵ=k∥2cs2ω,(5.74)\omega - \omega_ {* n \epsilon} = \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega},\tag{5.74}

which implies, in the regime of (5.68), that

ω≃ω∗ne+k∥2cs2ω∗ne.\omega \simeq \omega_ {* n e} + \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega_ {* n e}}.

Drift waves can be destabilized by a variety of effects. usually involving dissipation. The most common drift instabilities are characterized by a modified electron response, including a contribution out of phase with Φ\Phi . Thus (5.69) is replaced by

n1n0=eΦTϵ(1−iΔ)\frac {n _ {1}}{n _ {0}} = \frac {e \Phi}{T _ {\epsilon}} (1 - i \Delta)

where Δ\Delta reflects some dissipative process, such as resistivity, in Ohm's law. It is clear from (5.72) that positive Δ\Delta corresponds to instability.

Drift waves

The archetypal application of the drift model, the drift wave, does not involve EdE_{d} and propagates when Ti=0T_{i}=0 . The simplest drift-wave dispersion relation involves only ω∗e\omega_{*e} , which enters through E×BE\times B convection across the equilibrium density gradient.

The classical drift wave, sometimes called the “electron drift wave,” is an electrostatic perturbation,

E=−∇Φ\mathbf {E} = - \nabla \Phi

whose frequency satisfies

k∥vti≪ω≪k∥vte.(5.68)k _ {\parallel} v _ {t i} \ll \omega \ll k _ {\parallel} v _ {t e}.\tag{5.68}

We show that such a disturbance propagates, using the simplest version of the drift model, with vanishing ion temperature,

Ti=0.T _ {i} = 0.

The more general case is considered presently. We use Cartesian coordinates (x,y,z)(x,y,z) , with constant magnetic field in the zz -direction:

B=z^B=constant.\mathbf {B} = \hat {\mathbf {z}} B = \mathrm{constant}.

The equilibrium is assumed to vary only with x; in that case perturbations will depend on y and z through the familiar eikonal, so that we have, in particular.

n=n0(x)+n1(x)e−iωt+ik⊥y+ik∥z.n = n _ {0} (x) + n _ {1} (x) e ^ {- i \omega t + i k _ {\perp} y + i k _ {\parallel} z}.

where n1n_1 is the linear perturbation. We suppose that E∥\mathbf{E}_{\parallel} and EE\mathbf{E}_E vanish in equilibrium and for simplicity omit collisional effects.

Propagation of the electron drift wave depends upon the nature of the electron response to an electrostatic potential: its relaxation to a Maxwell-Boltzmann equilibrium. This so-called “adiabatic response,” resulting from fast electron motion through the nearly stationery potential ( ω≪k∥vte\omega \ll k_{\parallel}v_{te} ), is described by the parallel Olm's law, (5.64). In the absence of collisions the electron temperature is constant along field lines; then (5.64) can be seen to imply

n1n0=eΦTe.(5.69)\frac {n _ {1}}{n _ {0}} = \frac {e \Phi}{T _ {e}}.\tag{5.69}

Here and below we suppress the 1-subscript on Φ\Phi (since Φ0=0\Phi_0 = 0 ) and the 0-subscript on TeT_e (since Te1=0T_{e1} = 0 ). We emphasize that there is no counterpart to this relaxation—no allowance for adiabatic electron response—in MHD.

The ions, on the other hand, do not see a quasi-static potential and respond very differently. To understand their (quasineutral) density perturbation, note first that

∇⋅VE=0,\nabla \cdot \mathbf {V} _ {E} = 0,

for a uniform magnetic field. Hence the (nonlinear) particle conservation law becomes simply

∂n∂t+(VE+V∥)⋅∇n+n∇∥V∥=0,\frac {\partial n}{\partial t} + \left(\mathbf {V} _ {E} + \mathbf {V} _ {\parallel}\right) \cdot \nabla n + n \nabla_ {\parallel} V _ {\parallel} = 0,

or, after linearization,

−iωn1+ik∥n0V∥1+iω∗nen0eΦTe=0,(5.70)- i \omega n _ {1} + i k _ {\parallel} n _ {0} V _ {\parallel 1} + i \omega_ {* n e} n _ {0} \frac {e \Phi}{T _ {e}} = 0,\tag{5.70}

Here we noted that

VE⋅\n0=iω∗nen0eΦTe.\mathbf {V} _ {E} \cdot \backslash n _ {0} = i \omega_ {* n e} n _ {0} \frac {e \Phi}{T _ {e}}.

where

ω∗nϵ=−k⊥TeeBdlog⁡n0dx\omega_ {* n \epsilon} = - \frac {k _ {\perp} T _ {e}}{e B} \frac {d \log n _ {0}}{d x}

is the drift frequency defined with the density gradient, rather than the pressure gradient. Thus, as remarked previously, the drift frequency enters through E×BE \times B motion.

Our description is completed by the parallel equation of motion, whose linear version is

−iωmin0V∥1+ik∥pϵ1=0,(5.71)- i \omega m _ {i} n _ {0} V _ {\| 1} + i k _ {\|} p _ {\epsilon 1} = 0,\tag{5.71}

Equations (5.70) and (5.71) combine to yield

(ω−k∥2cs2ω)n1n0=ω∗neeΦTe,(5.72)\left(\omega - \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega}\right) \frac {n _ {1}}{n _ {0}} = \omega_ {* n e} \frac {e \Phi}{T _ {e}},\tag{5.72}

where

cs=Te/mi(5.73)c _ {s} = \sqrt {T _ {e} / m _ {i}}\tag{5.73}

is called the “ion sound speed.” Equation (5.72) is sometimes said to describe a hydrodynamic density response. The parenthesized factor on the left-hand side evidently describes sound waves propagating along the magnetic field.

Finally we combine (5.72) with electron adiabaticity, (5.69), to obtain the drift-wave dispersion relation

ω−ω∗nϵ=k∥2cs2ω,(5.74)\omega - \omega_ {* n \epsilon} = \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega},\tag{5.74}

which implies, in the regime of (5.68), that

ω≃ω∗ne+k∥2cs2ω∗ne.\omega \simeq \omega_ {* n e} + \frac {k _ {\parallel} ^ {2} c _ {s} ^ {2}}{\omega_ {* n e}}.

Drift waves can be destabilized by a variety of effects. usually involving dissipation. The most common drift instabilities are characterized by a modified electron response, including a contribution out of phase with Φ\Phi . Thus (5.69) is replaced by

n1n0=eΦTϵ(1−iΔ)\frac {n _ {1}}{n _ {0}} = \frac {e \Phi}{T _ {\epsilon}} (1 - i \Delta)

where Δ\Delta reflects some dissipative process, such as resistivity, in Ohm's law. It is clear from (5.72) that positive Δ\Delta corresponds to instability.

参考资料