A 59 mV peak-to-peak separation is essentially too good to be true! This is a theoretical value, and assumes an infinitely small step potential; in reality, the potentiostat applies a series of steps so different potentiostats can have different peak-to-peak separations. The peak separation (in volts) is only theoretically equal to 59 mV ideal at exactly 25*C.

Thiago mentioned the uncompensated resistance (iR) drop. A Luggin capillary is typically required to minimise this effect. Often people simply apply an arbitrary resistance compensation to their models or even via the potentiostat's software, to 'correct' the peak separation to get their expected 59 mV, but I would not recommend this.

Hedieh mentioned quasi-reversibility, and indeed the ferri/ferrocyanide redox couple is only a quasi-ideal outer sphere redox couple. It's kinetics are sensitive to the oxygen content on the electrode surface (some gold oxide = better). Finally, it is extremely sensitive to the alkali metal content in solution (more = better), and the nature of the alkali metal (Cs best, Li worst, from memory). If you are using something like a large excess of KNO3 as supporting electrolyte in water it should not be significant, but if you do not have a large excess of alkali metal present then the rate of electron transfer could be slow enough to become quasi-reversible (on the timescale of your scans).

Hello,

Generally, (E2-E1) is greater than 59 mV for quasi-reversible systems and increases with increasing sweep scan rate. So, maybe it is better to conduct a diagnostic test and evaluate the Cyclic voltammetry diagram for different scan rates.

It is not so easy to reach a peak separation of 59 mV. The distance between WE and RE should be very small. Beyond this, the sotution resistence needs to be very small too. To reach the value of 59 mV of peak separation, you will need to reapeat ideal conditions.

A 59 mV peak-to-peak separation is essentially too good to be true! This is a theoretical value, and assumes an infinitely small step potential; in reality, the potentiostat applies a series of steps so different potentiostats can have different peak-to-peak separations. The peak separation (in volts) is only theoretically equal to 59 mV ideal at exactly 25*C.

Thiago mentioned the uncompensated resistance (iR) drop. A Luggin capillary is typically required to minimise this effect. Often people simply apply an arbitrary resistance compensation to their models or even via the potentiostat's software, to 'correct' the peak separation to get their expected 59 mV, but I would not recommend this.

Hedieh mentioned quasi-reversibility, and indeed the ferri/ferrocyanide redox couple is only a quasi-ideal outer sphere redox couple. It's kinetics are sensitive to the oxygen content on the electrode surface (some gold oxide = better). Finally, it is extremely sensitive to the alkali metal content in solution (more = better), and the nature of the alkali metal (Cs best, Li worst, from memory). If you are using something like a large excess of KNO3 as supporting electrolyte in water it should not be significant, but if you do not have a large excess of alkali metal present then the rate of electron transfer could be slow enough to become quasi-reversible (on the timescale of your scans).

Since, all electrodes are gold, maybe there is an issue[1] of a (slow) gold surface modification, in alkali metal solutions (as supporting electrolyte).

1. Ionic-metallic transition in gold-based liquid alkali metal alloys  http://iopscience.iop.org/article/10.1088/0305-4608/11/5/009/pdf

Have a look at

http://demonstrations.wolfram.com/CyclicVoltammetryForARedoxReactionWithDiffusion/

Enjoy !

Might still be helpful now in 2020: two papers regarding this point.

First, the classic paper of Nicholson and Shain (1965) 'Theory of stationary electrode polarography' first developed the diagnostic criteria by simulation, which gives, for a one-electron-transfer reversible electrode reaction, the peak potential of 28.5 mV away from the formation potential. Therefore, the peak-to-peak separation should ideally be 57 mV, rather than 59 mV. This is an easy mistake as the latter is also a charateristic number in voltammetry - the full width at half maximum (FWHM) for the reversible peak, or the formal peak potential shift per proton/hydroxide ion per pH unit for a different pH from the standard condition (pH=0 or 14) in aqeuous solutions.

Second, J. Phys. Chem. C (2009) 113, 11157–11171 by Compton group comments that a few millivolts' deviation from the 57 mV may mostly be due to slightly incomplete electrolytic support (to compensate the iR drop of the solution between your working and referece electrodes). A sufficient concentration of supporting electrolyte is generally required to be greater than 100 times of that of your analyte. The other more possible reason I tend to believe for your '80 mV' is what's given by Leigh Aldous upstairs: the actual quasi-reversibility of ferri/ferrocyanide redox couple (which I assume your experiment used). Using the other common redox couple of ReHux will probably give you an almost ideal value as I've tried (often