The Rotating Ring-Disk Electrode: Catching Reaction Intermediates in Flight

A spinning electrode as a controlled hydrodynamic pump

The heart of the technique is not chemistry but fluid mechanics. When a disk embedded in an insulating shroud rotates about its axis in a viscous solution, it drags fluid around with it and, by centrifugal action, flings that fluid radially outward at the surface. Continuity demands replacement, so fresh solution is drawn up along the rotation axis toward the disk. This axial-in, radial-out flow is the classic von Kármán rotating-disk problem, solved by Theodore von Kármán in 1921 and refined by W. G. Cochran in 1934. Its great virtue is that the resulting mass transport is uniform across the entire disk face and, crucially, exactly calculable — a rarity in electrochemistry, where diffusion layers are usually ill-defined.

Because convection continuously renews the solution, a thin diffusion layer of well-defined thickness δ forms at the electrode. Benjamin Levich derived its thickness in the 1940s as δ ≈ 1.61 D^(1/3) ν^(1/6) ω^(−1/2), where D is the analyte diffusion coefficient, ν the kinematic viscosity, and ω the angular velocity (rad s⁻¹). For a typical aqueous system (D ≈ 1×10⁻⁵ cm² s⁻¹, ν ≈ 0.01 cm² s⁻¹) rotating at 1600 rpm (ω ≈ 168 rad s⁻¹), δ works out to roughly 12 μm. Spin faster and δ shrinks as ω^(−1/2), delivering more reactant per second.

The rotating ring-disk electrode exploits this radial flow deliberately. A concentric ring electrode, separated from the disk by a thin insulating gap, sits directly in the outward-flowing stream. Anything the disk generates — a dissolved intermediate, a partially reduced species, a metal ion — is carried across the gap and swept over the ring within a few to a few tens of milliseconds. Hold the ring at a potential where that species is electroactive, and the ring current becomes a direct, quantitative report on the disk's product flux.

The Levich and Koutecký-Levich equations

Under complete mass-transport control — every arriving molecule reacts instantly — the disk delivers the Levich limiting current: i_L = 0.620 n F A D^(2/3) ν^(−1/6) ω^(1/2) C*, where n is the number of electrons, F the Faraday constant (96485 C mol⁻¹), A the disk area, and C* the bulk concentration. The signature diagnostic is that i_L is linear in ω^(1/2) and passes through the origin. A straight Levich plot of i_L versus ω^(1/2) is the first proof that a system is behaving as a clean, diffusion-limited RDE; curvature or a nonzero intercept signals kinetic limitation, adsorption, or coupled chemistry.

When electron transfer is not infinitely fast, the measured current i is a series combination of a kinetic current i_K (independent of rotation) and the Levich current. Jaroslav Koutecký and Levich showed that 1/i = 1/i_K + 1/(0.620 n F A D^(2/3) ν^(−1/6) C* · ω^(1/2)). This is the workhorse Koutecký-Levich analysis: plot 1/i versus ω^(−1/2), and the slope gives n (if D and C* are known) while the intercept 1/i_K yields the intrinsic heterogeneous rate constant at that potential. The intercept is where the kinetics live — it is what survives when you extrapolate to infinite rotation speed, i.e., zero mass-transport resistance.

Two caveats keep this honest. First, the numerical prefactor is 0.620 when ω is in rad s⁻¹; using rpm or forgetting the 2π conversion is the single most common RRDE arithmetic error. Second, the slope of a K-L plot returns n only if you have the correct D and C* — a wrong diffusion coefficient masquerades as a wrong electron count. Practitioners therefore calibrate on a reversible outer-sphere couple such as [Fe(CN)₆]³⁻/⁴⁻ (n = 1, D ≈ 6.5×10⁻⁶ cm² s⁻¹) or ferrocenemethanol before trusting an n extracted for a mechanism they care about.

Collection efficiency: the geometry that reads out flux

What makes the ring quantitative is a single dimensionless number, the collection efficiency N. It is the fraction of a stable species generated at the disk that is subsequently caught and reacted at the ring: N = −i_R/i_D (the minus sign because the ring current has opposite sign — the ring re-oxidizes what the disk reduced, or vice versa). Remarkably, N depends only on the electrode geometry — the disk radius r₁, the inner ring radius r₂, and the outer ring radius r₃ — and not at all on rotation speed, concentration, or diffusion coefficient. That independence is why N can be treated as an instrument constant.

The closed-form expression was derived by W. J. Albery and Stanley Bruckenstein in 1966. It is built from the geometric ratios and involves the function F(θ) = (√3/4π)·ln[(1+θ^(1/3))³/(1+θ)] + (3/2π)·arctan[(2θ^(1/3)−1)/√3] + 1/4, with N computed from combinations of (r₂/r₁)³ and (r₃/r₁)³. For commercial tips a typical value is N ≈ 0.2 to 0.4, with the exact figure set entirely by the specific tip geometry — some glassy-carbon disk / Pt ring designs give N ≈ 0.24–0.26, while other widely used tips (e.g. the Pine E7-series) run closer to N ≈ 0.37–0.42. Because a real N always falls slightly below the theoretical value (finite gaps, imperfect insulation, edge effects), careful workers measure N experimentally using a known reversible mediator — again [Fe(CN)₆]³⁻/⁴⁻ is the standard — rather than trusting the manufacturer's nominal figure.

• Larger gap (bigger r₂/r₁): fewer molecules survive the crossing → lower N.
• Wider ring (bigger r₃): a larger catchment area → higher N.
• Faster rotation: N is unchanged for a stable species, but the transit time shortens — which is exactly how N becomes a clock for unstable ones.

Worked example: how much peroxide does an ORR catalyst make?

The flagship application of the RRDE is dissecting the oxygen reduction reaction (ORR), the sluggish cathode reaction of fuel cells and metal-air batteries. O₂ can be reduced by a direct 4-electron pathway to water, O₂ + 4H⁺ + 4e⁻ → 2H₂O (E° = +1.23 V vs. RHE), or by a wasteful 2-electron pathway to hydrogen peroxide, O₂ + 2H⁺ + 2e⁻ → H₂O₂ (E° ≈ +0.695 V vs. RHE). A good fuel-cell catalyst like Pt should run almost entirely 4-electron; the corrosive, membrane-degrading H₂O₂ byproduct must be minimized. The RRDE quantifies exactly that branching.

Set the disk to sweep cathodically through the ORR while holding the Pt ring at ~+1.2 V vs. RHE, a potential where any H₂O₂ swept off the disk is immediately oxidized back to O₂ (H₂O₂ → O₂ + 2H⁺ + 2e⁻). The %H₂O₂ is then %H₂O₂ = 200·(i_R/N) / (i_D + i_R/N), and the average electron number is n = 4·i_D / (i_D + i_R/N). Suppose at a given potential a glassy-carbon-supported catalyst gives i_D = −1.00 mA and i_R = +0.030 mA on a tip with N = 0.25. Then i_R/N = 0.120 mA, so %H₂O₂ = 200·0.120/(1.00+0.120) = 21.4%, and n = 4·1.00/1.120 = 3.57. The catalyst is running a mixed pathway — meaningfully off the ideal n = 4 — a quantitative verdict no single-electrode voltammogram could deliver.

This same protocol underlies thousands of electrocatalysis papers benchmarking Pt/C, Fe-N-C single-atom catalysts, and non-precious alternatives. When a 2020s single-atom catalyst is reported to give '>95% H₂O₂ selectivity' for on-site peroxide synthesis, or '<2% H₂O₂ with n ≈ 3.98' for fuel-cell ORR, those numbers come almost universally from RRDE measurements analyzed exactly as above.

Turning N into a clock: kinetics of the intercepted intermediate

For a perfectly stable intermediate, N is fixed. But if the species decays during its flight across the gap — disproportionating, reacting with solvent, or being consumed on the electrode's own surface — then fewer molecules reach the ring and the apparent collection efficiency N' falls below the geometric N. Because faster rotation shortens the transit time, N' rises toward N as ω increases. Analyzing N'/N (or N/N') versus ω^(1/2) therefore extracts the homogeneous rate constant of the intermediate's decay — the RRDE becomes a millisecond-resolution chemical kinetics instrument for species too reactive to isolate.

Albery and Bruckenstein systematized these EC, ECE, and DISP mechanism diagnostics in the late 1960s. The transit time across a typical gap at 1600 rpm is on the order of 5–20 ms, so the accessible rate-constant window is roughly 10⁰–10³ s⁻¹ for first-order homogeneous chemistry — bracketing many real intermediate lifetimes. A textbook case is the electrogeneration of an aryl radical anion at the disk whose protonation kinetics in the following chemical step are read out from the ring's shortfall; another is superoxide (O₂•⁻) generated in aprotic media, whose stability toward disproportionation and comproportionation is probed by comparing collected flux against the geometric N.

The technique also does reverse detection: hold the disk to generate a species and the ring to detect it, or generate at the ring and collect at the disk (the 'shielding' experiment, where an active ring depletes reactant reaching the disk). Shielding factors, like N, are pure geometry and provide an independent cross-check that the hydrodynamics are behaving as the von Kármán solution predicts.

Limits, artifacts, and good practice

The RRDE's beautiful math holds only under strict conditions, and violating them silently corrupts the numbers. The flow must be laminar: above a rotation-rate-dependent Reynolds number (roughly Re ≈ ωr²/ν ~ 2×10⁵, in practice above ~5000–10000 rpm for typical tips) the boundary layer becomes turbulent, mass transport stops scaling as ω^(1/2), and Levich linearity collapses. At the low end, below ~100 rpm, natural convection and vibration compete with forced convection and the diffusion layer is no longer clean. The reliable window is therefore roughly 100–4000 rpm for most work.

Practical pitfalls compound the physics. The insulating gap and ring must be perfectly concentric and coplanar; a wobbling or recessed tip destroys the uniform flow and drops N. Bubble formation — inevitable in vigorous ORR, HER, or OER measurements — pins to the surface and blocks flux, producing spiky, irreproducible currents. Ohmic (iR) drop is severe at the high currents RRDE can pass, so iR compensation and a low-resistance electrolyte are essential. And a bipotentiostat is mandatory: disk and ring must be controlled at independent potentials against a shared reference and counter electrode, or the collection experiment is meaningless.

Finally, interpretation demands care. A ring set to a diffusion-limited detection potential assumes it catches all arriving intermediate; if the ring reaction is itself kinetically sluggish, the measured N' underestimates the true flux. Background subtraction (capacitive and O₂-reduction backgrounds on the ring) and rigorous deaeration matter enormously when the intermediate flux is a few percent of the disk current. Done carelessly, an RRDE will confidently report a %H₂O₂ that is off by a factor of two; done rigorously — calibrated N, verified Levich linearity, iR-corrected, background-subtracted — it remains, six decades after Frumkin and Nekrasov, the gold-standard tool for reading a reaction mechanism straight off the current axis.

Rotating disk electrode (RDE) vs. rotating ring-disk electrode (RRDE)

Feature	RDE	RRDE
Electrodes	Single disk	Central disk + concentric ring (insulating gap)
Primary information	Total current, overall electron count n	Yield/flux of soluble intermediates and products
Key analysis	Levich & Koutecký-Levich plots	Collection efficiency N, ring/disk current ratio
ORR selectivity	Infers %H₂O₂ only indirectly from n	Directly quantifies %H₂O₂ from i_R/(N·i_D)
Detects unstable species	No — only sees net disk reaction	Yes — intercepts intermediate before it decays
Extra requirement	Bipotentiostat not needed	Bipotentiostat (independent disk & ring control)