loading...
running...
How to use
The platinum disk at the bottom of the shaft spins at the chosen rate and drags the electrolyte with it. The flow pulls fresh solution up to the disk and flings it outwards, so reagent reaches the whole surface at a steady, well-defined rate. The potential is swept in a triangle (or held); the current is whatever the reaction rate and the supply allow.
Try: pick Ferricyanide and watch the S-shaped wave and the flat plateau appear on the wall screen; the dashed grey curve is the closed-form steady state. Slide the rotation rate: the plateau grows with the square root of the rate and the yellow concentration slice under the disk gets thinner. Make the electrode sluggish (log k0 = -4) and the wave shifts to larger overpotential. Press Run Levich + Koutecky-Levich experiment to sweep seven rotation rates and fit the two textbook plots; the checks compare the fitted slopes and intercept with closed forms. At 200 mV/s the forward and backward scans no longer overlap.
3D rig: Blender 5.2 built the glass cell, PTFE lid, retort stand and rotator, the Pt disk in its PTFE shroud, the Ag/AgCl reference electrode with Luggin capillary, the Pt counter coil, the potentiostat, cables and bench props, all with Cycles-baked textures. The flow tracers follow the exact von Karman solution; their vertical scale is magnified x15 (the real hydrodynamic layer is a fraction of a millimetre thick) and the concentration slice is magnified by the slider. A real 1600 rpm spin would alias on a screen, so the drawn rotation is slowed (the HUD says by how much). Drag to orbit, wheel to zoom, Space pauses, R resets.
The science
A rotating disk electrode (RDE) is a flat disk in an insulating shroud spun about its normal at angular speed w = 2 pi (rpm)/60. It is a standard tool for measuring electrode kinetics, because the flow is laminar and the supply of reagent to the whole disk is uniform and known in closed form.
1. Flow. Von Karman (1921) and Cochran (1934) found a similarity solution of the Navier-Stokes equations: with zeta = z sqrt(w/nu), the velocity is vr = r w F(zeta), vtheta = r w G(zeta), vz = sqrt(nu w) H(zeta) and
2F + H' = 0, F2 − G2 + H F' − F'' = 0, 2FG + H G' − G'' = 0; F=0, G=1, H=0 at zeta=0; F, G → 0 as zeta → ∞
The page solves this boundary-value problem by shooting with a Newton iteration on F'(0) and G'(0). Near the wall H = −0.51023 zeta2 + zeta3/3 − ..., far away H → −0.8845: the disk pumps fluid axially toward itself at speed 0.8845 sqrt(nu w).
2. Convective diffusion. The reactant obeys the advection-diffusion equation; only the axial velocity matters because the concentration does not depend on r (uniform accessibility):
∂c/∂t + vz(z) ∂c/∂z = D ∂2c/∂z2, c(∞) = C, D ∂c/∂z|0 = kf c(0) − kb (C − c(0))
The product has the same D, so cR = C − cO everywhere and one field is solved by an implicit finite-volume scheme on a stretched grid with exponential-fitting (Scharfetter-Gummel) fluxes; the grid is scaled by the Nernst thickness so one solver serves all rotation rates. Rates follow Butler-Volmer: kf = k0 exp(−alpha n f eta), kb = k0 exp((1−alpha) n f eta), f = F/RT, eta = E − E0'.
3. Levich. With vz = −0.51023 w3/2 nu−1/2 z2 the steady limiting flux integrates exactly; the Nernst layer thickness is
δ = Γ(4/3) (3 D nu1/2 / (0.51023 w3/2))1/3 = 1.61 D1/3 nu1/6 w−1/2, iL = nFA D C/δ = 0.620 nFA D2/3 nu−1/6 w1/2 C
The Levich plot of iL against w1/2 is a straight line through the origin whose slope gives D (or n, or C). Using the full velocity profile instead of the leading term lowers the coefficient by the Schmidt-number factor 1/(1 + 0.298 Sc−1/3 + 0.145 Sc−2/3) (Newman), about 3 % for water.
4. Koutecky-Levich. At steady state cO(0) = C − J δ/D and J = kf cO(0) − kb cR(0), so
J = kf C / (1 + (kf + kb) δ/D) ⇒ 1/i = 1/ik + (1 + kb/kf) / (0.620 nFA D2/3 nu−1/6 C w1/2), ik = nFA kf C
Plotting 1/i against w−1/2 at fixed potential gives a line whose intercept is the kinetic current, free of mass transport.
Numerical checks run in the page. (1) The shot von Karman constants against Cochran's values 0.51023, 0.61592, −0.88447; (2) continuity, the integral of F equals −H(∞)/2 (all fluid pulled in must leave radially); (3) the finite-volume plateau against the textbook Levich coefficient 0.620; (4) the full-profile plateau against Newman's Sc correction; (5) the Cottrell current nFAC sqrt(D/(pi t)) just after a potential step; (6) a Monte Carlo of random walkers in the same drift field (escape probability against the scale-function integral, limit three standard errors); (7) after the experiment, the Levich slope and the Koutecky-Levich intercept and slope against the closed forms.
Simplifications. Infinite-disk hydrodynamics (real disks deviate by a few per cent from edge effects, and flow turns turbulent near Re = w r2/nu of about 2 x 105); equal diffusion coefficients for both partners; a one-step n-electron Butler-Volmer rate, which is only a phenomenological description of multi-step reactions such as the 4-electron oxygen reduction; no double-layer charging current, no ohmic drop (an ideal potentiostat with a Luggin capillary), no migration (excess supporting electrolyte), constant 25 C. The preset values (ferricyanide D about 7.6 x 10−6 cm2/s in 1 M KCl, O2 in 0.1 M KOH D about 1.9 x 10−5 cm2/s) are typical figures quoted from memory, not fits to measured data.