Use Ag/AgCl in 3 M KCl for most aqueous PEC work: +0.210 V vs SHE at 25 °C, converted with ERHE = EAg/AgCl + 0.210 + 0.0592 × pH. At pH 6.8 that offset is 0.612 V, so 0.40 V on the meter is 1.01 V vs RHE. SCE sits 31 mV away, which matters once currents exceed 5 mA/cm2.
The choice is set by the electrolyte, not by habit: chloride-containing references leak into the cell, changing the electrode surface as well as the junction potential.
| Reference | Filling solution | E vs SHE at 25 °C | E vs RHE at pH 7 | Notes |
|---|---|---|---|---|
| Ag/AgCl | 3 M KCl | +0.210 V | +0.624 V | General-purpose choice; about 0.2 mV/K |
| Ag/AgCl | Saturated KCl | +0.197 V | +0.611 V | Needs visible KCl crystals; drifts once diluted |
| SCE, Hg/Hg2Cl2 | Saturated KCl | +0.241 V | +0.655 V | Legacy comparison only; mercury waste stream |
| Hg/HgO | 1 M NaOH | +0.140 V | +0.554 V | Alkaline electrolytes, chloride-free |
| Hg/Hg2SO4 | Saturated K2SO4 | +0.640 V | +1.054 V | Sulfate electrolytes, no chloride at all |
At pH 14 every offset above rises by 0.414 V, so the 3 M Ag/AgCl electrode reads +1.039 V vs RHE. Quote the filling solution with the number: 3 M and saturated Ag/AgCl differ by 13 mV, more than most day-to-day drift.
The 0.0592 V per pH unit is the Nernst slope at 25 °C, and pH is where most conversions go wrong, because it is assumed rather than measured. A solution buffered at pH 6.8 instead of 7 costs 12 mV; an unbuffered electrolyte drifting from pH 6 to pH 4 moves the offset by 118 mV.
Measure pH in the same vessel, at the same temperature, before and after the experiment. Temperature enters twice: the Nernst slope rises to 0.0615 V per pH unit at 40 °C, and the reference itself shifts by roughly 0.2 mV/K for Ag/AgCl and 0.24 mV/K for SCE. A 5 K excursion therefore moves a converted potential by about 3 mV from the reference and 2 mV from the slope. Report potentials as V vs RHE with the pH and temperature stated, or report raw V vs the named reference and let the reader convert.
On a 60 mV/decade Tafel region a 10 mV offset is a factor of 1.47 in current density, a 47 % error in any rate comparison.
Every reference leaks filling solution into the cell. In a 100 mL cell a ceramic frit passing 1 µL/h adds chloride at roughly 30 µM/h from a 3 M KCl body, enough over eight hours to adsorb on a photoanode and shift the onset by tens of millivolts. Use a double-junction body with KNO3 or Na2SO4 in the outer chamber, or a chloride-free reference such as Hg/Hg2SO4.
The double junction is not free: each liquid junction adds 1–3 mV, and the outer chamber needs refilling weekly or the potential drifts as the bridge dilutes. Where 5 mV matters, calibrate the assembly as a unit against a fresh electrode.
Put the tip 1–2 mm from the working electrode, inside a Luggin capillary, and keep it out of the light path. An Ag/AgCl electrode exposed to the same beam develops a photopotential of 5–20 mV, which appears as a slow drift that looks like catalyst deactivation. Wrap the body in opaque tape or place it behind the electrode holder.
The reference must sit at the cell temperature: one at 22 °C in a cell held at 40 °C reads several millivolts off, and the error grows as the body equilibrates. Use a water-jacketed sleeve fed from the same bath.
Check the reference against a second one of the same type, both in the working electrolyte, and accept a difference below 5 mV. A healthy Ag/AgCl electrode drifts 1–2 mV per month; replace it when the offset exceeds 10 mV after cleaning, or when the internal silver wire turns patchy grey-black and the filling solution clouds. Keep it wet in 3 M KCl, never deionised water, where the filling solution leaches from the frit and the potential wanders for hours.
Saturated KCl bodies need solid crystals present at all times; a body stored cool can crystallise in the frit and block it, which shows up as tens of kilohms of resistance and a noisy trace.
Yes, if you name the filling solution and temperature, since 3 M and saturated bodies differ by 13 mV. Most journals prefer RHE for water splitting because it removes the pH dependence and lets curves at different pH be compared.
Add 0.241 V to the reading to reach the SHE scale, then add 0.0592 × pH. At pH 7 the total offset is 0.655 V, so −0.2 V vs SCE is 0.455 V vs RHE. The same arithmetic applies to any reference once its offset against SHE is known.
Light reaching the reference is the usual cause, producing a photopotential of 5–20 mV. Shielding the body and moving the tip out of the beam removes it. If the drift persists in the dark, suspect a blocked frit or a dilute filling solution.
Technically yes, but many laboratories have phased it out because of the mercury waste stream and disposal cost. Its 31 mV offset from 3 M Ag/AgCl is trivial to correct, so there is little reason to keep one for new work.
