μblog: engineering from the trenches

The fine people over at FaradNet have put together an appreciable set of notes on the electrolytic capacitors that appear in almost all consumer electronics devices. Although this is a good read for those who are interested using the devices in a safe manner (and getting the most performance out of them), there is a lot of text, so I will try to summarize the two features of electrolytics that seem to be most important: polarization and frequency response.

To understand the principle behind the polarization of the device, we must think of a little bit of electrochemistry. The electrolytic capacitor can be thought of as two electrodes with an electrolyte (substance containing free ions that can pass charge) between. If we neglect the circuit outside of the capacitor, we have only the two electrode-electrolyte interfaces to consider. From a manufacturing stand point, we can make one of the interfaces matter more than the other, so we can essentially form a rectifying system. Just as the diode stores energy in the depletion region when it is reverse biased (below breakdown), the “important” electrode-electrolyte interface can store energy when the applied voltage is biased in such a way as it is unfavorable for charge to flow across. This ability to store energy due to electric is the source of the capacitance in a dielectric capacitor. Just as the diode, forward bias conditions can lead to a large current even at small applied voltages, this current damages the electrode-electrolyte interfaces and generates heat which can further damage the device. For this reason, the capacitor polarization should be closely followed to prevent damage to the device or to the device operator if the capacitor is really abused.

At this point, it would seem that using the device in an AC circuit would require additional design considerations to maintain proper polarization at all times. This is true, however, there are other issues with electrolytic capacitors that pertain to AC circuits. The ability to store charge at the electrode-electrolyte layer has some dependence on ion mobility in the electrolyte which can be substantially slower than deformation of electron orbits in a tantalum dielectric. This causes the capacitance to drop off, sometimes substantially, as the signal frequency increases. For example, this would cause a low-pass filter to roll off at a slower-than-expected rate. The exact frequency-dependent change in capacitance is not the same for all electrolytics and will depend both capacitor design and its state. This why electrolytic capacitors are often left to DC applications such as decoupling and power regulation.

Hope this was informative and remember, safety first, performance second.

[ Image is from Tim Reese's page. ]