Department of Energy Argonne National Laboratory Office of Science NEWTON's Homepage NEWTON's Homepage
NEWTON, Ask A Scientist!
NEWTON Home Page NEWTON Teachers Visit Our Archives Ask A Question How To Ask A Question Question of the Week Our Expert Scientists Volunteer at NEWTON! Frequently Asked Questions Referencing NEWTON About NEWTON About Ask A Scientist Education At Argonne Capacitor Charge Holding Capability
Name: Jason
Status: other
Grade: other
Location: AP
Country: N/A
Date: 12/31/2004

I understand that in the electronics field capacitors cannot hold their charge for very long because they will discharge even when not connected to a circuit. Why is this, and what could be done about it? Is the charge leaking through the air or the insulators, or elsewhere? What types of capacitors hold their charge better, and why do we not just use those?

Jason -

Let me answer some of your questions. Capacitors loose their charge both through the insulation between the plates and through the air surrounding the capacitor. The charge is a surplus of electrons on one plate and a rarefaction of electrons on the other. Where the electrons are compacted (the negative plate) the electrons tend to push each other off. Where there is a deficit of electrons (the positive plate) electrons are attracted from other sources - air, the positive plate. Both of these actions tend to decrease the potential difference between the plates... to discharge the capacitor.

How can this be discouraged? There are a number of possibilities, but they are selectively employed due to practical and economic reasons. Two possible methods - Increase the distance between the plates or change the material separating the plates. For instance glass insulators are sometimes used on very large (tall as a house) capacitors or the capacitor may be packed in oil.

The loss of charge to the air can be controlled by increasing the distance between the contacts, by placing them in a vacuum, by placing caps on terminals, by insulating wiring, or by keeping the capacitor in a less humid environment.

In searching for more information on your subject you may wish to explore electrolytic capacitors. Be careful in handling large capacitors. The charge can be harmful.

Larry Krengel


It is impossible to make a capacitor that holds a charge indefinitely, though they can be engineered to hold a charge for a longer or shorter time. Remember that a capacitor consists of two conducting plates separated by an insulator, which could be solid, liquid, air or even a vacuum. A vacuum would make a capacitor hold its charge the longest. It is not used because it is difficult to make and would have a relatively low capacitance. Often one wants a capacitor to have the largest possible capacitance. This is accomplished by making the plates large in area and close together and filling the space between the plates with an insulator which has a large dielectric constant. A parallel plate capacitor has a capacitance given by C = eA/d, where e is the dielectric constant, A is the area of the plates and d is the separation between the plates. Notice that making the area large and the separation small makes it easier for a current to flow between the plates, thereby discharging them. For many purposes the small leakage current is not a serious problem. Often a large value of capacitance is much more important than a slow discharge. Notice that even with air between the plates, cosmic rays will occasionally pass through the capacitor, ionizing the air and thereby discharging the capacitor slightly.

Best, Dick Plano...

It is true, most capacitors tend to self-discharge about 50% in something like 15 minutes. You can think of your body storing static electricity- no matter how rubbery the soles of your shows, it usually fades a factor of two every half-minute or so. Sure, the voltage on most capacitors is far lower the 10kV you might walk around with, but the insulating layers in capacitors are also far thinner. They must be to get a competitive capacitance per unit volume. Nobody buys a capacitor the size of a 12 x 2 inch wax-paper roll to get 1micro-Farad capacitance for 100v use.

Some of the leakage is through the insulating layers. It is difficult to find a plastic substance with no electron migration or ion migration or charge absorption, a substance that has no low-level conduction mechanisms even though you are stressing it to 50% of its catastrophic breakdown field.

This implies that a capacitor will hold charge much longer if you are charging it only up to half or less of its rated voltage. I think this is often true. Using less than the rated maximum voltage is called "de-rating".

Electrolytic capacitors are not the best contenders for duration. There is always some ionic conduction occurring at field-strengths well below the breakdown field of the aluminum oxide skin on the aluminum foil inside. After all, the skin was deliberately _grown_ to the proper thickness by passing breakdown current through it, until the breakdown voltage was as high as desired.

Mineral-insulated capacitors would be the best (dry, hard SiO2 or Al2O3 or similar), but they are not doable in significantly large values with present marketable technologies. Silver-Mica uses Mica, which is a good mineral, but they only reach about 0.01 microfarads max. And they are not usually made with regard for the enhanced leakage that can occur at the sharp edges of a metal foil. It can be difficult to design away this problem. That leakage is similar to the humidity-and-air leakage that does most of the discharging of static electricity. The TV tube that holds a lethal charge for months is plausible, and it is a mineral-insulated capacitor. If most of the glass was to have a breakdown rating it would be about 100kV, and the charge that makes the tube dangerous is lower, perhaps 15kV. Possibly another example of de-rating, as well as of mineral-insulation.

Totally non-polar plastics are next best. Poly-propylene and polystyrene. Non-polar so they will have minimum absorbed water vapor "dissolved" within their solid structure, just waiting for the call to ionize and conduct some leakage current. Sometimes you can find some polypropylene up to 10 microfarads.

I suppose this kind of capacitor might be optimized by:

- never being touched by salty human fingerprints

- drying out the plastic at warm temperatures in vacuum or very dry gas for a long time, then

- holding at cool temperatures in glass-sealed metal canisters of perfectly dry air, and using only at reduced voltages.

Then the discharge time might be very long indeed.

Polar plastics like Mylar and polyesters are probably a good third.

Something I do not know how to rank yet are the "gold/double-layer" capacitors and the related "hybrid" capacitors. They are even more "wet-electrochemical" than electrolytic capacitors, but they have huge capacitance density (~1 Farad per cubic inch...) and the molecule-thick dipole layer between the water and the gold may actually self-refine, producing an interface with an extremely low density of current-conducting defect sites. I have heard they are used as memory-retention batteries for some electronics, but I do not know yet how long they might hold their charge. Bear in mind their voltage ratings only go up to a few volts.

There are a large number of types of capacitors. No one type or manufacturing method is best for everything yet. So all of them stay on the market. If I am making a pen-cam, I want the tiniest possible 100-microfarad capacitor for damping 100-100,000 cycle/second impulses, and the leakage of being electrolytic is a small price to pay. If I am a Chinese maker of $5 toy electric organs, cost matters more than leakage or size. Cheap, low-density electrolytics exist. If I'm making an integrating control circuit or a filter, I need slower self-discharge than electrolytic, but I cannot find any 30-microfarad polypropylene, and probably cannot afford the cost or volume of 6 pieces of 5-microfarad polypropylene. Then I would use polyester or Mylar. If I am making a 20kV capacitor, I cannot find a way to make thick enough plastic layers for 20kV, with no random pinholes, at a reasonable price, in an assembly process compatible with the metal sheets that must interleave with the plastic. Then I might actually go back to old-fashioned oil-soaked paper. Or maybe now there is something better, just as cheap....

Jim Swenson

Click here to return to the Physics Archives

NEWTON is an electronic community for Science, Math, and Computer Science K-12 Educators, sponsored and operated by Argonne National Laboratory's Educational Programs, Andrew Skipor, Ph.D., Head of Educational Programs.

For assistance with NEWTON contact a System Operator (, or at Argonne's Educational Programs

Educational Programs
Building 360
9700 S. Cass Ave.
Argonne, Illinois
60439-4845, USA
Update: June 2012
Weclome To Newton

Argonne National Laboratory