Skip to content
Projects

plasma_plots.theory.magnetron

Crossed-field (magnetron) discharges: Hull cutoff, Brillouin flow, drifts and gradient-drift modes.

SI units throughout (volts, tesla, metres, seconds, m⁻³); scalars or broadcasting numpy arrays. Electrons are the default species of the magnetic-confinement functions.

Examples

The Hull cutoff field of a planar magnetron with a 5 cm gap at 300 V:

>>> print(f"{hull_cutoff(voltage=300.0, gap=0.05) * 1e3:.4f} mT")
1.1681 mT

Functions

NameDescription
brillouin_flowCompute Brillouin flow: the rigidly rotating, uniform-density limit of a non-neutral cloud.
critical_ionisation_velocityCompute Alfvén's critical ionisation velocity v_c = sqrt(2 e U / m).
exb_drift_speedCompute the E×B drift speed E/B for perpendicular fields (via orbits.exb_drift()).
gradient_drift_frequencyCompute the gradient-drift dispersion relation with finite electron Larmor radius.
gradient_drift_growthCompute the gradient-drift growth rate γ = Im ω (zero if stable).
gradient_drift_phase_velocityCompute the phase velocity Re ω / k of the most unstable gradient-drift root.
hall_parameterCompute the Hall parameter ω_c / ν = |q| B / (m ν).
hull_cutoffCompute the Hull cutoff of a planar or coaxial magnetron (give all but one of voltage/field).
lower_hybrid_driftCompute the electrostatic lower-hybrid drift / modified two-stream root (Malkov and Sotnikov, Eq. 1).
simon_hoh_frequencyCompute the two roots of the cold-electron Simon–Hoh dispersion relation.
simon_hoh_growthCompute the Simon–Hoh growth rate γ = Im ω (zero if stable); see simon_hoh_frequency().

brillouin_flowfunction#

def brillouin_flow(field, radius=None, mass=electron_mass, charge=elementary_charge)

Compute Brillouin flow: the rigidly rotating, uniform-density limit of a non-neutral cloud.

Radial force balance of a cold non-neutral species rotating at ω_θ gives ω_p² = 2 ω_θ (Ω − ω_θ) with Ω = qB/m; the maximum density for which the rotation is rigid is ω_θ = Ω/2 and ω_p² = Ω²/2, i.e. n_B = ε₀ B² / (2 m).

Parameters

NameTypeDefaultDescription
fieldfloat or array_likerequiredMagnetic field B, in T.
radiusfloat or array_likeNoneIf given, the azimuthal speed v_θ = ω_θ r at this radius is returned as well.
massfloat or array_likeelectron_massSpecies mass (kg) and charge magnitude (C). Default: electron.
chargefloat or array_likeelectron_massSpecies mass (kg) and charge magnitude (C). Default: electron.

Returns

dict
"density" (m⁻³), "angular_frequency" (rad/s) and, if radius is given, "speed" (m/s).

Examples

>>> flow = brillouin_flow(0.05, radius=0.02)
>>> print(f"{flow['density']:.4g} m^-3, {flow['speed']:.4g} m/s")
1.215e+16 m^-3, 8.794e+07 m/s

critical_ionisation_velocityfunction#

def critical_ionisation_velocity(voltage, mass_number=1.0)

Compute Alfvén’s critical ionisation velocity v_c = sqrt(2 e U / m).

The speed at which a neutral of mass m has the kinetic energy e U. Taken with U the ionisation potential (Alfvén’s original hypothesis) or with U the discharge voltage when comparing to a spoke speed (Hecimovic and von Keudell 2018).

Parameters

NameTypeDefaultDescription
voltagefloat or array_likerequiredU, in V.
mass_numberfloat or array_like1.0Neutral mass in proton masses (as in alfven_speed()), e.g. 39.948 for argon. Default: 1.

Returns

float or numpy.ndarray
v_c, in m/s.

Examples

Argon at 100 V (equal to sqrt(2 e U / m) = 21.9 km/s):

>>> print(f"{critical_ionisation_velocity(100.0, 39.948):.1f} m/s")
21899.0 m/s

exb_drift_speedfunction#

def exb_drift_speed(electric_field, field)

Compute the E×B drift speed E/B for perpendicular fields (via orbits.exb_drift()).

Parameters

NameTypeDescription
electric_fieldfloat or array_likeComponent E⊥ of E perpendicular to B, in V/m (signed).
fieldfloat or array_likeMagnetic field B, in T.

Returns

float or numpy.ndarray
|E⊥| / B, in m/s.

Examples

>>> print(f"{exb_drift_speed(1e4, 0.05):.4g} m/s")
2e+05 m/s

gradient_drift_frequencyfunction#

def gradient_drift_frequency(wavenumber, electric_field, field, gradient_length, electron_temperature, mass_number=1.0, collision_frequency=0.0)

Compute the gradient-drift dispersion relation with finite electron Larmor radius.

Solves, for the root with the largest imaginary part, Eq. (1) of arXiv:2103.06559 (the form used for Hall-thruster and magnetron spokes: unmagnetized ions, magnetized electrons with Larmor radius ρ_e and electron–neutral collisions ν):

[ω_d + k²ρ_e² (ω − ω_E + iν)] ω² = k² c_s² [(ω − ω_E) + k²ρ_e² (ω − ω_E + iν)],

with ω_E = k E₀/B, ω_d = k T_e/(e B L_n), ρ_e = sqrt(T_e/m_e)/ω_ce and c_s = sqrt(T_e/m_i) (a cubic in ω, solved with numpy.roots()). Sign convention as in the module docstring. The ρ_e and ν terms are taken from that paper and not re-derived here; for ν = 0 and k²ρ_e² ω_E/ω_d → 0 (ω_ce → ∞, at fixed T_e: a 0.4 % effect in the doctest) the relation reduces to simon_hoh_frequency() (checked in the tests).

Parameters

NameTypeDefaultDescription
wavenumberfloat or array_likerequiredAzimuthal wavenumber k, in 1/m.
electric_fieldfloat or array_likerequiredE₀, in V/m (signed).
fieldfloat or array_likerequiredB, in T.
gradient_lengthfloat or array_likerequiredSigned density gradient length L_n, in m.
electron_temperaturefloat or array_likerequiredT_e, in eV (also the temperature in c_s; cold ions).
mass_numberfloat or array_like1.0Ion mass in proton masses. Default: 1.
collision_frequencyfloat or array_like0.0Electron–neutral collision frequency ν, in 1/s. Default: 0.

Returns

numpy.ndarray
Complex ω, in rad/s.

Examples

>>> k = 2 * np.pi / 0.05
>>> w = gradient_drift_frequency(k, 1e4, 0.05, 0.01, 3.0, 39.948)
>>> print(f"{w.real / k:.1f} m/s, {w.imag:.4g} 1/s")
603.8 m/s, 1.948e+06 1/s

gradient_drift_growthfunction#

def gradient_drift_growth(*args, **kwargs)

Compute the gradient-drift growth rate γ = Im ω (zero if stable).

Takes the arguments of gradient_drift_frequency().

Returns

float or numpy.ndarray
γ ≥ 0, in 1/s.

Examples

>>> k = 2 * np.pi / 0.05
>>> print(f"{gradient_drift_growth(k, 1e4, 0.05, 0.01, 3.0, 39.948):.4g} 1/s")
1.948e+06 1/s

gradient_drift_phase_velocityfunction#

def gradient_drift_phase_velocity(*args, **kwargs)

Compute the phase velocity Re ω / k of the most unstable gradient-drift root.

Takes the arguments of gradient_drift_frequency() (wavenumber first). In the cold-electron limit it is Ω_i L_n / 2, far below the E×B drift for typical magnetrons.

Returns

float or numpy.ndarray
Re ω / k, in m/s (signed with the convention of the module docstring).

Examples

>>> k = 2 * np.pi / 0.05
>>> print(
... f"{gradient_drift_phase_velocity(k, 1e4, 0.05, 0.01, 3.0, 39.948):.1f} m/s"
... )
603.8 m/s

hall_parameterfunction#

def hall_parameter(field, collision_frequency, mass=electron_mass, charge=elementary_charge)

Compute the Hall parameter ω_c / ν = |q| B / (m ν).

Parameters

NameTypeDefaultDescription
fieldfloat or array_likerequiredMagnetic field B, in T.
collision_frequencyfloat or array_likerequiredMomentum-transfer collision frequency ν, in 1/s.
massfloat or array_likeelectron_massSpecies mass (kg) and charge (C). Default: electron.
chargefloat or array_likeelectron_massSpecies mass (kg) and charge (C). Default: electron.

Returns

float or numpy.ndarray
ω_c τ, dimensionless (positive). Classical cross-field transport is reduced by 1/(1 + (ω_c/ν)²); Bohm diffusion corresponds to an effective value of 16.

Examples

>>> print(f"{hall_parameter(0.05, 1e8):.4g}")
87.94

hull_cutofffunction#

def hull_cutoff(voltage=None, field=None, *, gap=None, anode_radius=None, cathode_radius=None, mass=electron_mass, charge=elementary_charge)

Compute the Hull cutoff of a planar or coaxial magnetron (give all but one of voltage/field).

A particle starting at rest on the cathode in the crossed fields just grazes the anode when q B G = sqrt(2 m q V), i.e. B_c = sqrt(2 m V / q) / G and V_c = q B² G² / (2 m). The geometry factor is G = d for a planar gap d, and for coaxial cylinders (axial B) the conserved canonical angular momentum gives G = |r_a² − r_c²| / (2 r_a), with r_a the anode and r_c the cathode radius. For r_a/r_c → 1, G → d = |r_a − r_c| (the planar result). Non-relativistic.

Parameters

NameTypeDefaultDescription
voltagefloat or array_likeNoneAnode–cathode voltage V, in V. Omit it to compute the critical voltage.
fieldfloat or array_likeNoneMagnetic field B, in T. Omit it to compute the critical field.
gapfloat or array_likeNonePlanar gap d, in m. Give this or both radii.
anode_radiusfloat or array_likeNoneCoaxial geometry, in m.
cathode_radiusfloat or array_likeNoneCoaxial geometry, in m.
massfloat or array_likeelectron_massParticle mass (kg) and charge magnitude (C). Default: electron.
chargefloat or array_likeelectron_massParticle mass (kg) and charge magnitude (C). Default: electron.

Returns

float or numpy.ndarray
B_c in T if field is omitted, otherwise V_c in V. Above B_c (or below V_c) the particles cannot reach the anode.

Raises

ValueError
If not exactly one of voltage and field is omitted, or the geometry is not given.

Examples

>>> print(f"{hull_cutoff(voltage=300.0, gap=0.05) * 1e3:.4f} mT")
1.1681 mT
>>> print(
... f"{hull_cutoff(voltage=300.0, anode_radius=0.06, cathode_radius=0.01) * 1e3:.3f} mT"
... )
2.003 mT
>>> print(f"{hull_cutoff(field=1.16814e-3, gap=0.05):.1f} V")
300.0 V

lower_hybrid_driftfunction#

def lower_hybrid_drift(wavenumber, field, density, electron_temperature, ion_temperature, ion_drift, electron_drift, mass_number=1.0, parallel_ratio=0.0, diamagnetic_drift=0.0)

Compute the electrostatic lower-hybrid drift / modified two-stream root (Malkov and Sotnikov, Eq. 1).

Local, electrostatic, collisionless; unmagnetized Maxwellian ions of charge e (ω ≫ Ω_i and k ρ_i ≫ 1); magnetized Maxwellian electrons with finite Larmor radius; k ⊥ B up to a small k_z ≪ k. With exp(i(k y − ω t)) the relation is

1 + (2 ω_pi²/(k² v_Ti²)) [1 + ξ Z(ξ)] + (ω_pe²/Ω_ce²) (1 − Γ₀(b))/b
− (k_z²/k²) ω_pe²/(ω − k V_e)² + (2 ω_pe²/(k² v_Te²)) k V_Δ/(ω − k V_e) = 0,

with ξ = (ω − k V_i)/(|k| v_Ti), v_T = √(2T/m), Γ₀(b) = exp(−b) I₀(b) and b = k² T_e/(m_e Ω_ce²) = k² ρ_e². V_i and V_e are the ion and electron drifts along k (the instability depends on V_e − V_i, so V_i = 0 gives the ion rest frame) and V_Δ is the electron drift due to density, field and temperature gradients, taken as an input (the paper’s V_Δ = −(v_Te²/Ω_ce) Γ₀(b) {κ_n − …} includes the Larmor-radius factors). In the source the thermal speed v_T and b are printed ambiguously (b = |k| v_Te/Ω_ce); the values above are those for which the electron response reduces to the Debye one, (1 − Γ₀)/(k² λ_De²), and the V_Δ term to the standard Maxwellian drift term; the cold-electron limit is checked in the tests.

  • With k_z = 0 and V_Δ = 0 every root is real: a drive is needed.
  • With k_z ≠ 0, V_Δ = 0 and cold ions the relation is the modified two-stream relation ω_pi²/(ω − k V_i)² + (k_z/k)² ω_pe²/(ω − k V_e)² = R, R = 1 + (ω_pe²/Ω_ce²)(1 − Γ₀)/b, which has a complex pair iff R k² (V_e − V_i)² < (ω_pi^{2/3} + (k_z/k)^{2/3} ω_pe^{2/3})³ (Eq. (7) of the source, restated here and checked in the tests).
  • With k_z = 0 and V_Δ ≠ 0 the relation is the electrostatic lower-hybrid drift instability (LHDI): a lower-hybrid branch with ω of order ω_LH = ω_pi/√(1 + ω_pe²/Ω_ce²), driven by the gradient drift V_Δ through resonance with ions of relative drift V_i of a few v_Ti and T_i ~ T_e (the sign of V_Δ opposite to V_e − V_i, as in the example). This is the weak-β, electrostatic limit only: no magnetic fluctuations, no β-dependent stabilization, and no density-gradient (k_x) structure. The relation of this source was checked here against its cold-ion closed form and against scipy (see the tests), not against the original paper of Davidson and Gladd.

The root with the largest growth rate is found by Newton iteration from several seeds with solve_dispersion(); nan if none converged.

Parameters

NameTypeDefaultDescription
wavenumberfloat or array_likerequiredk along the drift, in 1/m.
fieldfloat or array_likerequiredB, in T.
densityfloat or array_likerequiredn_e = n_i, in m⁻³.
electron_temperaturefloat or array_likerequiredT_e and T_i, in eV.
ion_temperaturefloat or array_likerequiredT_e and T_i, in eV.
ion_driftfloat or array_likerequiredV_i and V_e along k, in m/s.
electron_driftfloat or array_likerequiredV_i and V_e along k, in m/s.
mass_numberfloat or array_like1.0Ion mass in proton masses. Default: 1.
parallel_ratiofloat or array_like0.0k_z/k. Default: 0.
diamagnetic_driftfloat or array_like0.0V_Δ in m/s along k. Default: 0.

Returns

numpy.ndarray
Complex ω in rad/s, in the frame in which V_i and V_e are given.

Examples

Xenon, n = 1e17 m⁻³, B = 20 mT, T_e = T_i = 10 eV, k ρ_e = 0.5 (k = 1326 1/m), ions drifting at −10 km/s (2.6 v_Ti) through electrons with V_Δ = 3 km/s and no E×B drift:

>>> w = lower_hybrid_drift(
... 1326.0, 0.02, 1e17, 10.0, 10.0, -1e4, 0.0, 131.293, 0.0, 3e3
... )
>>> print(f"{w.real:.4g} rad/s, {w.imag:.4g} 1/s")
-8.989e+06 rad/s, 4.691e+06 1/s

simon_hoh_frequencyfunction#

def simon_hoh_frequency(wavenumber, electric_field, field, gradient_length, mass_number=1.0, charge_number=1.0)

Compute the two roots of the cold-electron Simon–Hoh dispersion relation.

Assumptions: collisionless, k along E₀×B (k_∥ = 0), electrostatic, quasi-neutral; electrons magnetized with E×B and diamagnetic drifts and no Larmor-radius correction; ions unmagnetized (ω ≫ Ω_i) and cold, with inertia, so the ion density response is n_i1/n₀ = k² c_s² φ/ω² in units of e/T_e. The electron continuity n₁/n₀ = −(k φ / B L_n) / (ω − ω_E) closes to

ω² = a (ω − ω_E), a = Ω_i k L_n, ω_E = k E₀ / B, Ω_i = Z e B / m_i,

independent of T_e (the sound-speed and diamagnetic terms cancel). The roots are ω = [a ± sqrt(a² − 4 a ω_E)] / 2: unstable for ω_E / a > 1/4, i.e. E₀/(B Ω_i L_n) > 1/4, with real part a/2 (phase velocity Ω_i L_n / 2, independent of E₀ and k) and growth rate sqrt(a ω_E − a²/4), which is k sqrt(Ω_i L_n v_E) for v_E ≫ Ω_i L_n. L_n and E₀ are signed as in the module docstring (E₀ L_n > 0 unstable). This is the ω_ce → ∞, ν → 0 limit of gradient_drift_frequency().

Parameters

NameTypeDefaultDescription
wavenumberfloat or array_likerequiredAzimuthal wavenumber k, in 1/m.
electric_fieldfloat or array_likerequiredE₀, in V/m (signed).
fieldfloat or array_likerequiredB, in T.
gradient_lengthfloat or array_likerequiredSigned density gradient length L_n = n/(dn/dx), in m.
mass_numberfloat or array_like1.0Ion mass (proton masses) and charge number. Default: 1.
charge_numberfloat or array_like1.0Ion mass (proton masses) and charge number. Default: 1.

Returns

numpy.ndarray
Complex ω (rad/s) of the root with the larger imaginary part. The phase velocity is .real / k.

Examples

Argon, E₀ = 10 kV/m, B = 50 mT, L_n = 1 cm, k = 2π/(5 cm):

>>> k = 2 * np.pi / 0.05
>>> w = simon_hoh_frequency(k, 1e4, 0.05, 0.01, 39.948)
>>> print(f"{w.real / k:.1f} m/s, {w.imag:.4g} 1/s")
599.5 m/s, 1.944e+06 1/s

simon_hoh_growthfunction#

def simon_hoh_growth(wavenumber, electric_field, field, gradient_length, mass_number=1.0, charge_number=1.0)

Compute the Simon–Hoh growth rate γ = Im ω (zero if stable); see simon_hoh_frequency().

Returns

float or numpy.ndarray
γ ≥ 0, in 1/s.

Examples

>>> print(
... f"{simon_hoh_growth(2 * np.pi / 0.05, 1e4, 0.05, 0.01, 39.948):.4g} 1/s"
... )
1.944e+06 1/s
>>> print(simon_hoh_growth(2 * np.pi / 0.05, 1e4, 0.05, -0.01, 39.948))
0.0