I was explaining to my son that a capacitor is simply 2 sheets of foil separated by a dielectric and rolled up and he said, "can we make one then?"

I'm giving it a go. I used household aluminum foil and some parchment paper. I cut the foil into two strips, about 100mm by 200mm, placed parchment paper between them, inserted a paper clip into each, and rolled it up tightly. I'm pretty pleased that I made one that tests 9 nF consistently.

But if I want to have something big enough that we can continue to light an LED with it for a noticeable amount of time once we remove the battery, 9 nF isn't nearly good enough. What should we try next to hopefully get us a few orders of magnitude more capacitance?

  • 13
    \$\begingroup\$ A suggestion for making a smaller flash easier to see: first charge the capacitor, then disconnect the voltage source, then turn off all the lights, then connect the LED. According to this answer, about 23 nanocoulombs is enough of a discharge to produce a flash which is just barely visible if you’re watching closely. So, charging your 9 nF capacitor to about 5 V and then discharging it through a yellow LED might produce a visible flash in the dark. \$\endgroup\$ Commented Dec 29, 2021 at 2:37
  • 3
    \$\begingroup\$ @EdV That article says that the capacitance of a typical Leyden jar is about 1 nF, much less than the rolled-foil capacitor that the asker here made. \$\endgroup\$ Commented Dec 29, 2021 at 2:38
  • 10
    \$\begingroup\$ You can create a sapphire insulator (anodized aluminum) by using aluminum sheets with 1% glycerine, mostly water, and 1% fertilizers such as urea-ammonium phosphate or ammonium polyphosphates sandwiched in between, and applying a forming voltage continued until the current reaches about half of its initial current, then raising the voltage and doing the same, then raising the voltage and doing the same.. repeat and repeat until a desired end point is reached. Then test at lower voltage to check the leakage. You may be able to reach \$100\:\mu\text{F}\$ or better with your size given my notes. \$\endgroup\$
    – jonk
    Commented Dec 29, 2021 at 8:39
  • 4
    \$\begingroup\$ Notes say that I got somewhere around \$6\mu\:\frac{\text{F}\cdot\text{V}}{\text{cm}^2}\$ to about \$10\mu\:\frac{\text{F}\cdot\text{V}}{\text{cm}^2}\$, with some care and effort. \$\endgroup\$
    – jonk
    Commented Dec 29, 2021 at 8:43
  • 6
    \$\begingroup\$ Unfortunately for home construction, it's really easy to get a high voltage, 8 nF 10 kV is possible with a 2 litre PET bottle filled with salt water, with foil outside, but high capacitance needs huge area or very thin films. Not what you want if you are going to get 'hands on' with your child. Try a roll of plastic bags from the dollar store, easier to handle than Saran, and not much thicker. \$\endgroup\$
    – Neil_UK
    Commented Dec 29, 2021 at 10:06

11 Answers 11


I cut the foil into two strips, about 100mm by 200mm, placed parchment paper between them, inserted a paper clip into each, and rolled it up tightly. I'm pretty pleased that I made one that tests 9 nF consistently.

I made a capacitor the same size as yours, but used cling film (aka 'Saran wrap') as the insulator, and got 37 nF. Cling film is made from a plastic called polyvinylidene chloride (PVDC), which has a dielectric constant of 3.9 at 1 kHz. After charging it to 9 V I got a noticable flash from a 5mm superbright white LED.

  • 19
    \$\begingroup\$ It's probably good to keep in mind the other difference between cling-film and paper: the paper is thicker. That width is the charge separation in your capacitor, which you want to be small. \$\endgroup\$ Commented Dec 29, 2021 at 15:02
  • 3
    \$\begingroup\$ I don't have a citation, just a memory, but in a book I loved as a kid, plastic wrap (or in those days, probably cellophane) was the dielectric in the capacitor in a DIY crystal radio set. \$\endgroup\$ Commented Dec 29, 2021 at 16:47

hopefully get us a few orders of magnitude more capacitance?

The suggestion about Saran wrap is a good one, but the easiest way to get a few orders of magnitude is to remember the capacitor equation and increase the area. 100mm x 200mm is all very well, but what if you used a whole roll of foil + insulator? With a bit of care you could put the two source rolls next to one another and roll evenly onto a target cardboard tube or spindle.

The next source of improvement is to acknowledge that capacitors (huge voltage curve) and LEDs (constant voltage) are very badly matched. If you built a small buck converter you could charge the capacitor to a higher voltage, thereby storing more energy for the same number of Farads. Best to keep it below 50V, though.

  • 8
    \$\begingroup\$ Or just discharge through a larger resistor, so you can still raise the voltage (more total charge) and hit the same current. Larger R means larger RC time constant. Not as good performance as a buck convert (total LED energy scaling with V, not V^2), but very simple and easy for a beginner to understand. \$\endgroup\$ Commented Dec 29, 2021 at 22:23
  • 2
    \$\begingroup\$ If you connect the capacitor to the LED via an inductor, you will deliver a much larger fraction of the capacitor energy to the LED (so it will be brighter). Wind 1 m of wire in a small loop and use that. There is a risk this would damage the LED (as the voltage reverses) -- so connect 2 LEDs in antiparallel. Both will light briefly. \$\endgroup\$
    – jp314
    Commented Dec 30, 2021 at 16:47
  • 1
    \$\begingroup\$ Certainly the resistor works, but ends up losing significant energy itself. What would be great is OP trying the resistor and inductor approaches and reporting their findings :) \$\endgroup\$
    – pjc50
    Commented Dec 30, 2021 at 16:51

The equation for the capacitance of a plate capacitor is...
C = E0 * Er * A / D

C = Capacitance in Farads
E0 = Permittivity of free space = 8.854 pF / m
Er = Relative permittivity of dielectric
A = Area of plates in square meters
D = The distance between the plates in meters

To increase capacitance, you can do any of the following...

  • Increase the area (use more foil)
  • Decrease the distance (use a thinner insulator)
  • Use a dielectric material with higher relative permittivity (Er)
  • Use a wet dielectric

Theoretical capacitance of a 100mm x 200mm capacitor made from various materials.

  • Seran Wrap (polyethylene): E0 x 2.3 100mm x 200mm / 12.7um = 32.1nF
  • Wax paper: E0 x 2.5 100mm x 200mm / 25um = 17.7nF
  • 20GSM Printer paper: E0 x 2.3 x 100mm x 200mm / 100um = 4.1nF

Its typically not possible (especially with a homemade capacitor) to make perfect contact between the foil and a solid dielectric material. This will have the effect of increasing D as well as reducing the effective Er. So, the real capacitance will probably be less than calculated above.

Using a wet dielectric allows the dielectric to make contact with the aluminum foil without any gaps due to surface roughness. Also, there are many readily available non-conductive liquids that have very high Er values.

Glycerin is a good choice. It has Er = 47-68, its non-toxic, it doesn't evaporate, and its readily available in places like Walmart in either the hand lotion or pharmacy sections. To use it, impregnate a thin piece of paper with the glycerin and then place the paper between your foil as usual.

  • 20GSM Printer paper + glycerine: E0 x 47 x 100mm x 200mm / 100um = 83nF
  • 10GSM Tissue paper + glycerine: E0 x 47 x 100mm x 200mm / 50um = 166nF
  • 10GSM Tissue paper + glycerine: E0 x 47 x 100mm x 1m / 50um = 832nF
  • 10GSM Tissue paper + glycerine: E0 x 47 x 100mm x 10m / 50um = 8.32uF

So, if you use glycerine impregnated tissue paper and make the foil longer (>1m) you can get values approaching 1uF or more.

  • 2
    \$\begingroup\$ Excellent suggestion to use glycerine, low cost, non-toxic, high Er. However, there's a potential problem which is its miscibility with water, a little contamination with which will increase the leakage significantly, and is worth a warning. \$\endgroup\$
    – Neil_UK
    Commented Dec 30, 2021 at 15:32
  • 1
    \$\begingroup\$ Yes, not only is glycerine miscible with water but it's actually hygroscopic, i.e. merely leaving it exposed on the paper will quickly attract moisture from the air. It's probably still ok as long as there are no ions in the water. \$\endgroup\$ Commented Dec 30, 2021 at 16:48
  • \$\begingroup\$ One could try coating the capacitor to prevent contaminants from getting into the glycerin. \$\endgroup\$
    – user4574
    Commented Jan 3, 2022 at 3:47
  • 1
    \$\begingroup\$ After posting my answer I tried making some of these glycerin capacitors. Two bottles of glycerin were used, some 99.5% pure glycerin from Walmart and some 99% pure industrial glycerin from McMaster. Two kinds of paper were used, regular printer paper and tissue paper. Measuring with a DMM the glycerin and paper provided many megaohms of isolation. The printer paper was much easier to work with in terms of not being easy to break... \$\endgroup\$
    – user4574
    Commented Jan 3, 2022 at 3:59
  • 1
    \$\begingroup\$ After applying 1V power, several mA of leakage was measured. The leakage increased over time, implying that a chemical reaction was taking place that was increasing the concentration of ion impurities in the glycerin. It is not clear where the original impurities came from, but it seems even the presence of small amounts of impurities either from the glycerin or the paper can destroy this type of capacitor. It might be better to use a fine nylon mesh or some other fairly inert insulator. \$\endgroup\$
    – user4574
    Commented Jan 3, 2022 at 4:00

You might find it better to make a battery, but that's not what you asked.

You could try making an electrolytic capacitor using aluminum foil and a paper separator. Maybe some engine antifreeze and borax for the electrolyte. Optimum mixture unknown (by me) but borax is fairly soluble.

You can roll it up, with the electrolyte saturating the separator, and immerse it in a cylindrical container.

You can apply current-limited voltage for a period of time to 'form' the oxide layer if it is not adequate. Store aluminum foil does not have the etching that commercial electrolytic caps have (which increases the effective oxide surface area immensely) so I wouldn't expect much but you may be able to get a noticeable flash out of a modern LED.

I have not tried this, but it should work.

The polarity will be that of the 'forming' voltage, but up to a volt or two it should be bipolar.


A brief try yielded a very conductive electrolyte (great) but way too much leakage (>1mA) which is a show-stopper here. I think it is necessary to use aluminum wire to get out of the electrolyte (which I did not do). That way an oxide layer can be anodized onto all the materials connected to the electrode.


You can use a thin layer of insulating varnish as the dielectric, use a sheet with increased area and roll it up tightly.

  • Take 2 foils of large area, without ANY insulating coating.
  • Dip one foil in insulating varnish, then let it dry. (this encapsulates that foil with a thin layer of varnish).
  • Press the 2 foils together.
  • Solder the leads and take them out.
  • Roll it up very tightly over a thin insulating rod.
  • Dip the whole thing in glue / varnish , without loosening it (this is just to maintain its shape and tightness).
  • 1
    \$\begingroup\$ I have my doubt that you'll manage to produce a reliably insulating varnish film that's actually thinner than cling-wrap foil, without precision techniques for applying it. \$\endgroup\$ Commented Dec 30, 2021 at 16:45

Make it bigger and use a thinner foil. 2 m² polyethylene foil with 20 μm thickness should give you something just below 1 μF, probably enough to see a flash.


The old, soviet era capacitors I disassembled as a child were constructed similarly to yours, but with one very important difference. They were separated not with parchment, but rather with some very thin porous paper that was soaked in oil. There was no air in the package. They were also rolled up very, very tightly, after unrolling them and rolling them back I was in no way able to fit them back into the aluminum case.

Dimensions were also different, you have 10 x 20cm, ones I played with were more like 2 x 100 cm, rolled into a package 2cm high and significantly under 2cm in diameter.

If my memory serves me, strips with similar area as yours were in the order of tens or low hundreds nF.


Simply put: the thinner the dielectric, and the larger the area of the plates, the larger the capacitance.

Historically, a class of capacitors known as polythene used the approach described by the person suggesting Saran Wrap (UK: cling film); capacitances of a few microfarads were achieved.

The Wikipedia page has a good description of the physics and scaling with area and thickness.


The capacitance depends on three things, plate surface area, separation between plates, and the dielectric constant of the insulating material.
The easiest way to increase the capacitance is to increase the plate area, so you'd want longer, wider foil.
Having the plates closer together increases the capacitance, this also tends to lower the breakdown voltage but with low voltages such as a battery this shouldn't be a problem. You could try gluing the foil to the paper with a very thin layer of glue, that would keep it together tighter and might increase the capacitance. Alternatively you could just try rolling it tighter.
As for the dielectric constant, for an experimental cap you're kind of stuck with whatever you get with paper.


Great answers here on making a nice capacitor, but remember that big capacitors in reality are made up of a set of capacitors connected in parallel. In the image below, each rectangular box contains several cylindrical capacitors, and each four boxes are connected in parallel.

image source: https://commons.wikimedia.org/wiki/File:Condensor_bank_150kV_-_75MVAR.jpg

enter image description here


Kitchen aluminum foil is OK, but you need a thin insulator layer. Anodizing works well. For capacitor use, a thin oxide insulator is desired and thus you'd normally want Type 1 anodizing with chromic acid. That's a severely nasty thing to work with and requires special handling. Best forget about I even mentioned it. I don't even suggest the typical sulphuric acid, since that's not really fit for ad-hoc occasional use without a good safety protocol and PPE in place. Instead, use a 20-30% solution of NaHSO4 (sodium hydrogen sulfate), commonly used for lowering pH in pools. You can get it anywhere pool supplies are sold. It's less nasty than H2SO4, but still an irritant and you want goggles and gloves when working with it.

Dunk two well degreased pieces of aluminum foil into such solution, attached to a current limiting power supply, and run some DC current through it - a couple amps/sq.ft should be more than enough. The positive anode will get coated with oxide. You'll need some experimentation to get the thinnest possible oxide layer that doesn't short the two foils once you press them together. The oxide layer can be sealed by putting the anodized piece into boiling tap water for a couple of minutes. This may help make the capacitor more durable perhaps.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.