The Taylor Cone: How an Electric Field Sprays a Liquid

A drop that grows a sharp point

Put a droplet of salty water at the tip of a fine metal needle, place a grounded plate a few millimeters away, and slowly raise the voltage between them. At first nothing dramatic happens: the drop bulges slightly toward the plate but stays rounded. Then, at a well-defined critical voltage, the drop abruptly reorganizes into a crisp cone, and from the cone's tip a jet of liquid so thin you cannot see it carries a spray of charged droplets to the plate. That cone is the Taylor cone.

The effect was photographed in electrified menisci by John Zeleny around 1914–1917, and studied further by C.T.R. Wilson and G.I. Taylor in the 1920s. But the definitive theory came from Geoffrey Ingram Taylor in a 1964 paper, “Disintegration of water drops in an electric field” (Proc. Roy. Soc. A). Taylor predicted — and then confirmed with soap-film and water-drop experiments between electrodes — that the equilibrium cone has an exact half-angle of 49.3°, a number that depends on neither the liquid nor the size of the drop nor the strength of the field.

A tug of war between two pressures

The whole phenomenon is a contest between two surface stresses. The first is surface tension. A liquid interface behaves like a stretched skin of energy γ per unit area (for water γ ≈ 0.072 N/m), and it resists curvature with a Laplace pressure that pulls the surface inward: Δp = γ(1/R₁ + 1/R₂). Left alone, surface tension keeps a drop spherical, because a sphere minimizes area.

The second stress is electrical. A conductive liquid is an equipotential body, so the applied field drives free charge to its surface; the field just outside a conductor is purely normal, of magnitude E, and it exerts an outward Maxwell stress — a negative pressure, or tension — of magnitude σ_M = ½ ε₀E². This electrostatic pull tries to stretch the surface out along the field lines and sharpen it into a point, because charge crowds most densely where the surface is most curved, intensifying the field there in a self-reinforcing loop.

Below the threshold, surface tension dominates and the drop stays round. At the threshold the two stresses balance. Above it, the electric stress wins locally at the apex, the balance runs away, and the surface pulls out into a jet. For real liquids of finite conductivity, the full accounting is the Taylor–Melcher “leaky dielectric” model, which tracks how charge relaxes to the interface on a finite timescale rather than instantaneously.

Why the angle is 49.3 degrees, exactly

The precise cone angle falls out of a beautiful self-similarity argument. For a static cone to be an equilibrium shape, the electric stress and the surface-tension pressure must balance not just on average but at every distance r from the apex. A cone has one zero principal curvature (along its straight generators) and one that scales as 1/r, so its Laplace pressure falls off like γ/r. The Maxwell stress is ½ε₀E². These two can match at all r only if the field itself scales as

• E ∝ r^−1/2, so that ½ε₀E² ∝ 1/r, tracking γ/r term-for-term.

Now demand that the cone surface be an equipotential of a charge-free field (Laplace's equation). The only solution with the required V ∝ r^1/2 radial dependence involves the Legendre function of degree one-half, P_1/2(cosθ). For the cone at polar angle θ to be an equipotential, this function must vanish there. Its single relevant zero sits at θ ≈ 130.71°, which means the cone opens with a half-angle of 180° − 130.71° = 49.3° measured from the axis (a full apex angle of about 98.6°). Because the argument uses only the two pressure scalings and Laplace's equation, the answer is universal: any conducting liquid, at any scale, at the critical field, wants the same 49.3°.

From cone to jet to a mist of charged droplets

The perfect cone is a mathematical limit; a real apex cannot sustain the singular r^−1/2 field forever. Once the drop is fed with liquid, the tip emits a thin, highly charged filament — the celebrated cone-jet mode. The jet can be nanometers to microns across and moves fast, and it fragments a short distance downstream into small, remarkably uniform charged droplets by a charge-modified capillary (varicose) instability, a cousin of the Plateau–Rayleigh breakup that pinches an ordinary faucet stream into drops.

The steady cone-jet obeys clean scaling laws worked out by Fernández de la Mora & Loscertales (1994) and Gañán-Calvo: the emitted current scales roughly as I ∝ (γ K Q / ε)^1/2 and the droplet diameter as d ∝ (ε₀ Q / K)^1/3, where K is the liquid's conductivity, Q the flow rate, and γ its surface tension. Higher conductivity and lower flow rate give finer droplets — down to nanometers.

The droplets do not stop shrinking there. As solvent evaporates, a droplet's radius R falls while its charge stays roughly fixed, until it hits the Rayleigh limit — the maximum charge a drop can hold, Q_R = 8π√(ε₀γR³), predicted by Lord Rayleigh in 1882. At that point Coulomb repulsion overwhelms surface tension and the drop undergoes Coulomb fission, spitting out a jet of tiny progeny droplets. The cascade repeats, ultimately delivering nanodroplets or bare ions (the charge residue model of Dole and the ion evaporation model of Iribarne & Thomson).

Viscosity decides: a spray or a fiber

The same Taylor cone can end in a spray of droplets or an unbroken thread, and the deciding factor is the liquid's viscosity and viscoelasticity. A low-viscosity Newtonian liquid lets the jet neck down and pinch off into droplets — this is electrospray. But a polymer solution with long, entangled chains resists capillary pinch-off; the jet survives, thins, and instead undergoes a violent, electrically driven whipping (bending) instability that stretches it enormously while the solvent evaporates, leaving a solid nanofiber tens to hundreds of nanometers across. This is electrospinning, patented by Anton Formhals in the 1930s.

So the choice between mist and fiber is set at the fluid level even though the cone at the source looks the same. Electrospun mats — with their huge surface-area-to-volume ratio — are now used for high-efficiency filtration media (electrospun nanofiber filters, a technology distinct from the melt-blown electret media used in standard N95 respirators), tissue-engineering scaffolds, battery separators, and drug-delivery membranes.

Where Taylor cones do real work

The Taylor cone is not a laboratory curiosity; it underpins several important technologies.

• Electrospray ionization mass spectrometry (ESI-MS). By spraying a solution of proteins or other large molecules from a Taylor cone, John Fenn (in the late 1980s) showed you could lift intact, multiply charged biomolecules into the gas phase without shattering them — a “soft” ionization that opened up proteomics. Fenn shared the 2002 Nobel Prize in Chemistry (with Koichi Tanaka and Kurt Wüthrich) for this work.
• Electrospray and colloid thrusters. Fire charged droplets or ions of an ionic liquid (e.g., EMI-BF₄) out of a Taylor cone and accelerate them through a kilovolt field, and you get an extremely efficient micro-thruster with specific impulse ranging from hundreds of seconds (droplet/colloid mode) up to several thousand seconds (pure ionic-liquid ion mode). ESA's LISA Pathfinder used colloid micro-Newton thrusters for precise, drag-free spacecraft control, and such thrusters are attractive for CubeSat propulsion.
• Liquid-metal ion sources. A Taylor cone of molten gallium is the emission tip in the focused ion beam (FIB) instruments used to mill and image at the nanoscale.
• Other uses span fine-particle synthesis, precision printing, agricultural electrostatic spraying, and fuel atomization.

How we know it — and how it fails

The cone, jet, and droplet cascade are studied with high-speed and stroboscopic imaging (which resolve the 49.3° angle and the emitting jet), with current probes on the collector, and with phase-Doppler sizing of the droplet spray. A stable single cone-jet exists only inside a finite window of voltage and flow rate: too low and the meniscus drips or pulsates; too high and it breaks into multiple jets or, in air, the intense field triggers corona discharge — the surrounding gas ionizes and breaks down before the liquid can spray cleanly. Operating in vacuum, in an electronegative gas such as SF₆, or with volatile solvents suppresses that failure mode.

Two subtleties are worth keeping straight. First, 49.3° is the ideal static equilibrium angle; a real flowing cone-jet usually sits at a somewhat smaller dynamic angle because the momentum of the emitted liquid perturbs the pure pressure balance. Second, the crossover from droplet emission to direct ion emission at the nanoscale tip — and the precise thresholds separating dripping, cone-jet, and multi-jet modes — remain active research questions in electrohydrodynamics.

Meniscus behavior of a conductive liquid as the applied voltage climbs past the electrospray threshold

Regime	Applied voltage	Meniscus shape	Outcome
Subcritical	Below onset (< V_on)	Rounded, near-hemispherical drop	Surface tension wins; a pendant drop hangs or drips
Critical (Taylor cone)	≈ V_on	Sharp cone, 49.3° half-angle	Electric stress balances surface tension in a static cone
Cone-jet	Slightly above V_on	Cone with a thin jet at the apex	Steady jet emits fine, nearly monodisperse charged droplets
Multi-jet / overspray	Well above V_on	Rim sprouts several emission points	Multiple unstable jets; in air, risk of corona breakdown