Electrochemists like reference electrodes but potentiostats can have a different opinion. That lovingly crafted jelly filled bridge leading from a bubbly Luggin to a clogged up SCE placed next to the rotating-disc motor will give the instrument something to chew on. Most students have electrolysed cells dry as a result of a bubble in a Luggin. I know I did.
If the absolute potential is not so important but you are looking for changes or rate of change in the current with respect to voltage then consider another piece of the working electrode immersed into the electrolyte. Its potential will probably remain fairly constant during the test given a simple non-dynamic situation and its impedance is most likely very low. High impedance reference electrodes cause three problems:-
- DC offset - Slight DC offset due to the finite input current of the reference electrode terminal - say a 1 K Ohm SCE into a 20 pA input bias current RE terminal giving 20 nV offset - is no problem. The situation changes if a multimeter is connected between RE and WE which can typically have an input resistance of 10 M Ohm making the situation orders of magnitude worse. The motto is disconnect DVMs when doing tests.
- AC effects - When using reference electrodes for AC Impedance there are a couple of effects to watch out for. Firstly the interrelationship between the resistance of the RE and the input capacitance of the RE terminal gives an effect of low pass filtering. This is usually acceptable at the typical corrosion frequencies of 20KHz or less but gives perhaps a degree of phase shift at the top frequencies. The second problem is the fact that reference electrodes are themselves electrochemical cells possessing their own capacitance resistance and time constant. With an SCE above 17 kHz people generally end up measuring the reference electrode.
- EMI (Electromagnetic Interference) - A method of virtually eliminating electromagnetic interference has been invented by ACM. Using the ACM fourth electrode the noisiest of cells can be measured. An example of our new method is our Noise Reduction Probe. Please download the Noise Reduction Probe application note apn-Noise-ReducingElectrode-v1.0.pdf for more details
For additional information on noise pickup in your system, please view these application notes :-
1. Instrumentation Noise - Details on types and sources of noise and how limit noise. Please download apn-InstrumentNoise-v1.0.pdf
2. Noise Induction - Avoiding noise pickup in your test system. Please download apn-NoiseInduction-v1.0.pdf
3. Noise Reduction Probe - Removing noise from your test system using a Noise Reduction Probe. Please download apn-Noise-ReducingElectrode-v1.0.pdf
Stability and Time Constants
A perfect potentiostat is unconditionally stable with every cell. The way all potentiostats are constructed at present involves feeding an analogue signal back from the reference electrode to the virtual earth of an operational amplifier. The capacity for introducing phase shifts occurs on the WE; the AE and the RE these rendering current design potentiostats liable to oscillation. At ACM we tackle this using practical technology.
A potentiostat should only be as fast as it needs to be. This sounds simple but how fast does it need to be to perform a cyclic sweep at 1 mV/sec? You can buy potentiostats that let you or your technician perform this essentially DC test with a frequency response on the potentiostat of 1 MHz and the same on the ZRA (current measuring stage). This is a recipe for disaster. The software should look ahead to the forthcoming test and set the appropriate speed. All of ours do this slowing down the control amplifier when needed and damping the ZRA. At each current range the time constant of the ACM ZRA is the same allowing very smooth current range changes.
This paragraph is in the wrong location: it should be above the preceding one because (in)stability follows current interrupt IR compensation as Brazilians follow football. Current interrupt IR compensation looks good on paper but in practice it can be the biggest cause of unstable potentiostats there is. The best way to compensate for IR drop is to measure the solution resistance with an AC signal and then adjust the output voltage with a simple algorithm. We know that this is much more stable than current interrupt having made both. The reason people use current interruption is one of cost as it's much cheaper to put a little interrupter in the circuit than incorporate a full blown AC analyser. Using our own on-board DSP the solution resistance can be found in 1 second then used correctly. We found when we made our current interrupter prior to using the DSP we spent all our time answering the phone over sampling speeds cell oscillation cable capacitance and noise. That's the same noise that creeps into high electrolyte impedance cells - just the sort of cell that IR compensation is needed for. As for using current interrupt when measuring AC Impedance...!
During AC Impedance tests the potentiostat is put into a faster mode. The compromise between stability and speed is most apparent when measuring high impedance cells. A few pF of capacitance is needed across the ZRA; Potentiostat and RE to maintain stability. This then determines the maximum operating frequency for any current range. Our software understands this maximum and sets the appropriate range. This does not mean that a 100 MOhm cell can then be measured at 100 kHz. The practical limit for say a 100 MOhm cell is about 10 Hz. All manufacturers are constrained by the same laws of physics so only systems with impedances of around 100 -1000 Ohms can be measured with any accuracy using a top frequency of 100 kHz. Using our Femto Amp (was named Paint Buffer) and measuring a cell of 1 GOhm the top frequency for accurate measurement is about 1Hz the same as other honest manufacturers. We do however have an interesting device to squeeze more accuracy out of an instrument. Each of our instruments that incorporate AC (Gill 8, Field Machine, Gill AC and Gill 12) undergoes a very extensive calibration procedure over a series of 8 decades of cell load 10 amplitudes and the full range of frequencies to produce a calibration matrix we call the P.A.I.R. table (Phase and Impedance Reduction). This is then applied to the raw data to reduce cable and instrument induced errors. The instruments can easily be recalibrated using any length and type of cable. For example 60 feet of cable up a power station flue can easily effect the results when looking for dew point corrosion but not if calibrated with a P.A.I.R. table.
Some people need power and not just an Amp or two but tens even hundreds of Amps. We make big current systems. They are not cheap but they do work.
The last tip is just a note to scientists and technicians who come to testing for the first time. Don't be afraid to have a play with your system at first. Try simple jam jar tests with chunks of electrodes taped to the side. Waggle electrodes about when polarising. Try the effect of widely different sweep rates and polarising voltages. Shine lights on the cell, turn on motors, scratch electrodes and bubble the electrolyte. Try altering cell geometry and generally get a feel for the orders of magnitude of the system. Then tighten up your experimental design not the other way round.
When the design is tightened up watch out for those bubbles in Luggin probes; broken connections; crevices between sample and holder; dry SCEs; touching electrodes and inadvertent ground loops. If your computer is to be left unattended - remove all games and don't activate a screen saver (your test will cost more than a monitor) and make sure everything is hard to unplug. Let people know that a test is underway and sleep soundly.