Plasma Sheath

What a sheath actually is

Drop any solid object into a plasma — a probe, a reactor wall, a satellite — and within nanoseconds the plasma surrounds it with a thin charged layer. That layer is the plasma sheath. It exists because the two species in a plasma are wildly mismatched in mass: an electron is about 1836 times lighter than a proton (and far lighter than an argon or xenon ion). At a shared temperature the electron thermal speed beats the ion thermal speed by the square root of the mass ratio — a factor of roughly 43 for hydrogen, almost 270 for argon.

So when a fresh surface meets a plasma, electrons rain onto it far faster than ions. The wall charges negative. That negative charge does two things at once: it repels the very electrons that are arriving, and it pulls positive ions toward the surface. Equilibrium is reached when the slowed electron flux exactly matches the boosted ion flux. What's left over near the wall is a region with more ions than electrons — a net positive space charge — and a steep potential drop that confines the bulk plasma to nearly constant potential. That space-charge layer is the sheath.

The key idea: a plasma is a fantastic conductor, so it hates to support an electric field in its interior. When you force a boundary on it, it dumps the entire job of screening that boundary into a microscopically thin skin. Everything interesting at a plasma–wall interface happens inside that skin.

The Debye length sets the scale

The natural length over which a plasma screens out an electric perturbation is the Debye length:

λ_D = √( ε₀ k_B T_e / (n_e e²) )

Numerically, a convenient form is λ_D ≈ 7430 · √(T_e[eV] / n_e[m⁻³]) metres. A few representative values:

Plasma	T_e	n_e	λ_D
Capacitive etch / processing plasma	3 eV	10¹⁸ m⁻³	≈ 13 µm
Glow-discharge fluorescent tube	1 eV	10¹⁷ m⁻³	≈ 24 µm
Tokamak edge / scrape-off layer	20 eV	10¹⁹ m⁻³	≈ 11 µm
Solar wind at 1 AU	10 eV	5 × 10⁶ m⁻³	≈ 10 m
Ionosphere (F region)	0.1 eV	10¹² m⁻³	≈ 2.3 mm

A collisionless DC sheath is a few Debye lengths thick — so microns in a lab device, but metres around a spacecraft. The bulk plasma, which can be metres across, stays quasineutral; only this last sliver next to the wall carries net charge. The separation of scales (sheath ≪ bulk) is exactly what lets us treat the bulk as a neutral fluid and the sheath as a thin boundary condition.

Quasineutrality and where it breaks

In the bulk, quasineutrality holds: n_e ≈ n_i to within tiny fractions of a percent over any region larger than λ_D. Charge separation costs enormous energy — the Poisson equation, ∇·E = (e/ε₀)(n_i − n_e), means even a 1% imbalance over a centimetre would build fields of millions of volts per metre. The plasma simply won't tolerate it in the bulk.

The sheath is the deliberate exception. Inside it, ion density exceeds electron density because the negative wall potential exponentially suppresses electrons (Boltzmann factor) while ions keep streaming in. Poisson's equation then has a non-zero right-hand side, and a real electric field survives — typically 10⁴–10⁶ V/m in lab sheaths. This is the field that does the work: confining electrons and firing ions into the surface.

The potential structure: bulk, presheath, sheath

Reading from the plasma toward the wall, the potential falls through three zones:

• Bulk plasma — flat potential, perfectly quasineutral, no net field.
• Presheath — a gentle, quasineutral ramp many λ_D long (often set by collisions or geometry) that drops about 0.5·T_e/e of potential and accelerates ions up to the Bohm speed.
• Sheath proper — the thin non-neutral layer carrying the bulk of the potential drop, where the field is strong and electrons are nearly excluded.

The handoff between presheath and sheath is governed by the Bohm criterion.

The Bohm criterion

David Bohm showed in 1949 that a stable, monotonic sheath only exists if ions enter it fast enough. The threshold is the Bohm speed (the ion-acoustic speed):

u_B = √( k_B T_e / m_i )    →    ions must satisfy  u_ion ≥ u_B  at the sheath edge

The intuition: as the potential falls, ions accelerate and their density drops like 1/u. Electrons drop off exponentially (Boltzmann). For the net charge to stay positive everywhere — so the field never reverses and the sheath stays monotonic — the ions must already be supersonic relative to ion sound. The presheath is the plasma's mechanism for delivering ions at exactly u_B. For hydrogen at T_e = 5 eV, u_B ≈ 2.2 × 10⁴ m/s; for argon at the same T_e, u_B ≈ 3.5 × 10³ m/s.

Because the presheath accelerates ions and rarefies them, the density at the sheath edge is reduced from the bulk by a factor of about e^(−1/2) ≈ 0.61. The ion flux that reaches the wall is therefore the famous Bohm flux:

Γ_i ≈ 0.61 · n₀ · u_B

Floating, grounded, and biased walls

How much potential the sheath holds depends on what the wall is connected to.

Wall condition	Net current	Sheath potential
Floating (isolated)	Zero — ion flux = electron flux	V_f ≈ −(T_e/2e)·ln(m_i/2π m_e); ~ −2.8 T_e (H), ~ −4.7 T_e (Ar)
Grounded relative to plasma	Net electron current may flow	Plasma potential V_p sits ~ +(T_e/2e)·ln(...) above ground
Negatively biased (e.g. etch electrode)	Mostly ion current	Thick, high-voltage sheath; thickness ∝ V^(3/4) via Child–Langmuir
Positively biased (electron-collecting probe)	Large electron current	Sheath can vanish; electron-rich layer forms instead

For a floating wall the recipe is simple: set electron flux equal to ion flux and solve for the potential. The electron flux is the random thermal flux suppressed by the Boltzmann factor exp(eV/k_BT_e); the ion flux is the Bohm flux. Equating them gives the floating potential above. The lighter the ion, the smaller the logarithm — hydrogen walls float at about −2.8 T_e, heavy argon at about −4.7 T_e — because heavier ions arrive more slowly and the wall must charge more negative to throttle electrons down to match.

High-voltage sheaths and the Child–Langmuir law

When you apply a large negative bias (hundreds of volts, as on a sputtering target or etch electrode), electrons are excluded almost entirely and the sheath becomes a space-charge-limited ion diode. Its thickness follows the matrix/Child–Langmuir scaling:

s ≈ (√2 / 3) · λ_D · ( 2 e |V| / (k_B T_e) )^(3/4)

So a 500 V bias on a 3 eV plasma can stretch a 13 µm Debye length into a sheath of order 0.5–1 mm — large enough to see and to shape with electrode geometry. The current that the sheath can pass scales as J ∝ V^(3/2)/s², the classic Child–Langmuir three-halves law for a space-charge-limited diode.

RF sheaths and self-bias

Most plasma processing runs on radio frequency (13.56 MHz is the industry standard). The ions, being heavy, respond only to the time-averaged field, while the light electrons follow the instantaneous RF voltage. The sheath oscillates: it collapses briefly each cycle to let a burst of electrons through, then re-expands. On an asymmetric, capacitively coupled electrode this rectifies into a large negative DC self-bias, which is exactly how engineers dial in the ion bombardment energy on a wafer independent of the plasma density.

Python — floating potential, Bohm speed, and sheath thickness

import math

e   = 1.602e-19      # C
me  = 9.109e-31      # kg
eps0 = 8.854e-12     # F/m
mp  = 1.673e-27      # kg

def debye_length(Te_eV, ne):
    """Electron Debye length [m]. Te in eV, ne in m^-3."""
    Te_J = Te_eV * e
    return math.sqrt(eps0 * Te_J / (ne * e**2))

def bohm_speed(Te_eV, mass_amu):
    """Ion-acoustic (Bohm) speed [m/s]."""
    mi = mass_amu * mp
    return math.sqrt(Te_eV * e / mi)

def floating_potential(Te_eV, mass_amu):
    """Floating potential relative to plasma [V] (negative)."""
    mi = mass_amu * mp
    return -(Te_eV / 2.0) * math.log(mi / (2 * math.pi * me))

def child_langmuir_thickness(Te_eV, ne, V_bias):
    """Approx high-voltage sheath thickness [m]."""
    lam = debye_length(Te_eV, ne)
    return (math.sqrt(2) / 3) * lam * (2 * abs(V_bias) / Te_eV)**0.75

# Argon processing plasma: Te = 3 eV, ne = 1e18 m^-3
print(f"Debye length:       {debye_length(3, 1e18)*1e6:.1f} um")     # ~12.9 um
print(f"Bohm speed (Ar):    {bohm_speed(3, 40):.0f} m/s")            # ~2680 m/s
print(f"Floating pot (Ar):  {floating_potential(3, 40):.1f} V")     # ~ -14 V  (= -4.7 Te)
print(f"Bohm flux factor:   {math.exp(-0.5):.3f} * n0 * u_B")       # 0.607
print(f"500 V sheath:       {child_langmuir_thickness(3, 1e18, 500)*1e3:.2f} mm")  # ~0.47 mm

# Hydrogen for comparison
print(f"Floating pot (H):   {floating_potential(3, 1):.1f} V  (= -2.8 Te)")

Where the sheath shows up

• Semiconductor manufacturing. The sheath field is the ion accelerator of plasma etching. Because it points straight into the wafer, ions arrive nearly vertically and carve anisotropic, high-aspect-ratio trenches. RF self-bias sets the ion energy; this is the heart of every reactive-ion etch tool.
• Magnetic fusion. In a tokamak divertor the sheath throttles the heat and particle flux onto the target plates. The sheath heat-transmission coefficient (≈ 7–8 times T_e per ion–electron pair) is a primary input to designing wall materials that survive megawatts per square metre.
• Spacecraft charging. A satellite in the magnetosphere floats inside its own sheath. Differential charging across a sheath can build kilovolts and trigger discharges that fry electronics — a real failure mode for GEO satellites.
• Langmuir probes. The workhorse plasma diagnostic measures current versus bias voltage; interpreting that I–V curve is sheath physics. The ion-saturation current gives density via the Bohm flux; the exponential electron region gives T_e.
• Electric propulsion. Hall thrusters and ion engines accelerate ions through sheath and quasineutral acceleration regions to produce thrust at high specific impulse.
• Plasma deposition and sputtering. Sheath-accelerated ions sputter target atoms (PVD) or drive surface chemistry (PECVD), depositing thin films for coatings and microelectronics.

Common misconceptions

• Thinking the sheath is neutral. The opposite is true — the sheath is the one non-neutral region. Net positive charge there is what holds up the field.
• Confusing presheath and sheath. The presheath is long, gentle, and quasineutral; the sheath is short, steep, and charged. Most of the potential drop lives in the sheath, but the Bohm-speed entry condition is set in the presheath.
• Forgetting the Bohm criterion. Ions don't drift lazily into the sheath; they must already be at the ion-acoustic speed, or no monotonic sheath solution exists. Plugging u = 0 into the equations gives nonsense.
• Assuming the wall sits at plasma potential. An isolated wall always floats negative of the plasma by several T_e. The plasma potential is the most positive potential in the system.
• Using DC sheath thickness for RF. RF sheaths breathe — their thickness oscillates each cycle and the time-averaged structure (and self-bias) is what matters for ion energy.
• Ignoring the mass ratio. Floating potential, Bohm speed, and ion energy all depend on which ion you have. Argon and hydrogen sheaths behave quite differently for the same T_e and n_e.