|
<< Click to Display Table of Contents >> Navigation: Experiments > H - Electrochemical Noise |
Noise in electrochemical corrosion measurements is common, showing up in both the current and voltage signals. It originates from several different sources, including:
•Electronic and thermal noise in the instrument
•Antenna pickup in the cell and cell leads
•Digitization of analog signals
•Turbulent mass-transport in the chemical cell
•Stochastic processes at the electrode surface
Noise is often a nuisance degrading the signals you are trying to measure, but in some situations, it can be a benefit to the corrosion engineer. In pitting corrosion, noise measurements can detect the onset of pitting. There is a background noise in uniform corrosion processes caused by the first four mechanisms in the above list. At the onset of pitting, however, the noise from electrode surface reactions can increase dramatically. This noise can show up in either the current signal or voltage signal, or both.
There are several ways to measure current noise and voltage noise. You can apply a constant I signal and measure E noise, or vice versa. You can also measure both current and voltage simultaneously on an electrochemical cell under control of a ZRA (zero-resistance ammeter).
The Electrochemical Noise system measures noise using potentiostatic, galvanostatic, or ZRA control modes:
•In potentiostatic tests, a corrosion specimen is immersed in a solution and its potential is controlled using a Gamry Instruments, Inc. potentiostat/galvanostat/ZRA in its three-electrode potentiostat mode. The noise in the cell current is measured as a function of time.
•Galvanostatic tests are also three-electrode experiments where a constant current is applied to the cell, and the noise in the cell voltage is measured as a function of time.
•In ZRA tests, two identical corrosion specimens are immersed in the same solution. The same potentiostat/galvanostat/ZRA is used, but this time it is connected as a ZRA. Current is measured between the two working electrodes which are held at the same potential. Voltage is measured independently between the two specimens and a reference electrode.
The Electrochemical Noise software uses a quite simple method to measure noise. Periodically, short blocks of data are acquired using the potentiostat's internal A/D converter. Current is measured for potentiostatic mode, potential for galvanostatic mode, and both for ZRA mode. You control both the length of the data block and the point density of the block. Typically, each block is 20 seconds long and contains 100 data points.
You also control the time-spacing between blocks, and the total test time. Typical values are one block every 10 minutes for 24 hours.
There are also a number of methods for ECN data-treatment. The Electrochemical Noise system tests use a simple detrended RMS calculation on each data block to get an estimate of the noise-level during the block. This block/mean calculation is repeated periodically to build up the noise-versus-time curve.
Data-treatment for ECN measurements in potentiostatic or galvanostatic modes do not require a separate discussion. Those modes are simpler than ZRA mode in which only current or potential noise, not both, is measured and analyzed, but they are otherwise similar.
The noise calculation is straightforward. Each block of current-and-potential-versus-time (I vs. E vs. t) data is described by three arrays:
A DC trend is calculated from the data arrays using a standard linear best fit:
The noise of the block is estimated by subtracting the DC trend and calculating the RMS of the residual AC signal:
These values are a measure of current noise and potential noise at the time the block was acquired. These values, along with the time at the start of the block, are appended to the output curve. The output curve is therefore current noise and potential noise versus time.