# Highest energy density electrolytic capacitors

I'm building a project that involves discharging a fairly large capacitor through a pair of electromagnetic coils. It's electrically very similar to a coilgun, though for a different application. I can't use a supercapacitor because their power density is not high enough, and so I need to find the electrolytic capacitors with the highest energy density, because I'm limited to around 80g for the capacitor.

1. At approximately what maximum voltage do standard aluminum electrolytic capacitors have the highest energy density? I can design my coil and flyback transformer around the capacitor voltage if there is an advantage to higher voltages, but if they're all approximately the same, I think the electronics would be easier with a 35V cap than a 200V one.

2. Is there a site or a manufacturer who has a list of capacitors with the energy density for each one? It seems like a feature Parts.io or a similar site would have but I can't find it anywhere. The way I'm doing it now, I have to find the package size for each cap in the datasheet, and then find the weight of that package size and calculate energy density, which is getting rather tedious.

3. Should I be looking into another capacitor technology instead of electrolytic? I'm having a hard time finding tantalum capacitors this large, and the ones I do find have specs similar to the electrolytics.

• It is very uncommon to list "energy density" for a capacitor, you will have to make the calculation yourself for each capacitor. I doubt wether you will find a cap with a higher energy/volume than a supercap though. Even though supercaps can only handle very low voltages. – Bimpelrekkie Oct 22 '15 at 19:28
• Try running the numbers using (say) Digikey. I would expect around 10-50V for the capacitors, but that's just a guess. Volumetric energy density may be different from mass energy density. It's not normally specified, so you can download a spreadsheet of different capacitors and calculate it yourself then sort. Shouldn't take long. I doubt any capacitors approach the density of a rechargeable battery. – Spehro Pefhany Oct 22 '15 at 19:28
• The problem is that sites like Digikey, Mouser, and Parts.io don't list the weight as one of their standard items, so you have to search through the datasheet for every one. I can do that and put it into a spreadsheet, it just takes forever. And I know that no electrolytic capacitor can match the energy density of a rechargeable battery or supercap, but no 80g rechargeable battery or supercap will output the ~50KW peak power I need. – ahalekelly Oct 22 '15 at 19:37
• Interesting question, I have seen it one a few capacitor datasheets before measured in J/g. I guess for aerial systems it could be useful... My guess would be medium voltage as @SpehroPefhany mentioned. I would look at Electrolytic Polymer before anything else, it has a high density I believe and very low ESR. – MadHatter Oct 22 '15 at 19:47
• 80g and 50kW peak power means you will have only microseconds pulse duration with a capacitor. As @MadHatter says, check out conductive polymer but I don't see masses given even in the datasheets. It's not unusual in design to come up with a figure of merit that nobody else considers important. – Spehro Pefhany Oct 22 '15 at 20:05

Thanks for the help everyone, I ended up manually making the spreadsheet, with capacitors ranging from 7.5 to 82g and 16mm to 40mm diameter. Basically the higher voltage the better, within a capacitor family. Some capacitor families go up to 550V, but the most energy dense families only go up to 450V. The next largest factor is the connection style, in that snap-in capacitors are much lighter for a given capacity than either screw terminal or radial leaded capacitors. It also seems that smaller diameter capacitors had a higher energy density, but length had a fairly small effect. Manufacturer basically doesn't matter as their high end capacitors all have the same capacitance for a given case size and voltage.

Because of this, I ended up choosing the smallest diameter and longest snap-in capacitor available, the 22mm diameter, 50mm long, 400V 330uF United Chemi-Con ESMQ401VSN331MP50S, and am planning on putting 3 of them in parallel. They don't list the mass of their capacitors, but based on TDK's mass for the same size it is 2.66 J/g. The highest energy density I could find was 2.99 J/g, but that and several other capacitors at the highest energy density were not stocked anywhere.

• Nice work on the charts. Don't forget about your coil resistance and inductance. Lower-voltage caps may work fine for a flat or square-wire wound 0.01 Ohm coil (since the current will be higher into the low resistance) but if the coil is measured at 2 Ohms, then that will limit the current based on the available voltage. Inductance can also be a factor, since what you're really doing is connecting a capacitor in series with a time-variant inductor, which is very much an "LC" resonant tank circuit the moment it is energized. Current will surge quickly, then oscillate. – rdtsc Nov 5 '15 at 14:54

At first glance there are hints of fast discharge in your question.

The two types of capacitor that are rated best for this application of energy storage and fast discharge are flash lamp capacitors and electric fence energizer capacitors. Both should be close to the best you will find at reasonable prices. If you want better then something rated for aviation grade to keep the weight low might be better but will cost a lot more.

EDIT:
To cover your voltage question a bit more, The higher the better up to a point. Usually the point is dictated by your switching components or the device you plan to power. 350V DC is a popular limit for consumer switching semiconductors as it corresponds with peak line voltages for devices used at 230V AC.

Your choice of manufacturer is only relevant if you want lots, otherwise you are better served with searching retail or distributor sites.

Without a few more clues about your chosen project it would seem that aluminium electrolytics are the way to go. Tantalums have very few additional benefits here.

• This is a good answer. I would caution against tantalums, though. They are notoriously bad with both fast voltage spikes and/or high discharge currents. They can explode in a white flash :) – bitsmack Oct 23 '15 at 16:50

A very important specification for your capacitor is how long you want it to last.

Generally speaking, manufacturers get robust long lasting components by adding weight, and by derating performance. If you supply to the auto engine market for instance, you really do not want failures, you do not want to be responsible for upsetting millions of affluent consumers.

If you want to maximise performance, then limited life, and an unsophisticated non-complaining end user is required. You get very good energy density for the electrolytic flash energy storage caps used in disposable cameras. They are rated to discharge into a 1$\Omega$ flashlamp load. It's difficult to find a lifetime specs for these, I have seen a datasheet for something like them claiming 5000 shots, but I'm not certain whether it was for the disposable product, or a higher quality flash-gun end use. After all, the most a disposable camera will fire is 36 times.

I've just weighed out 80g of caps that I rescued from disposable cameras, comprising 11 off 120uF 330v (rated) 350v (surge) caps. At 330v, that will give you a total of 72J. Charge to the surge voltage for even lower lifetime, and you'll just scrape 1J/g specific energy. Is this remotely in your ballpark?

Given the rated load per capacitor, and you'd have 11, 50kW discharge power is comfortably within their capability.

• @adrian how much energy do you need? – Neil_UK Nov 4 '15 at 13:08
• I hadn't thought of the lifespan as a factor, and my requirements are fairly low. I would be fine with replacing the capacitor every 100 discharges. Because this is such a low lifespan, charging to a slightly higher voltage than it is rated for is a good idea. – ahalekelly Nov 4 '15 at 20:00
• As for energy storage, really the more energy and power the capacitor can output, the more effective this will be. I think the minimum energy for this to be useful will be somewhere in that order of magnitude, but that depends very heavily on the efficiency of my electromagnets. – ahalekelly Nov 4 '15 at 20:05