Skip to content
AESTECHNO
AESTECHNO
Contact
Capacitors

Energy stored in a capacitor: U = ½CV²

A capacitor stores energy in the electric field between its plates. The amount depends only on its capacitance C and on the voltage V across it, not on how it was charged. With Q=CV, the same energy can be written Q22C or QV2.

That energy is what keeps a microcontroller alive for a last write when the supply drops, what a supercapacitor is built to hold, and why a large capacitor can still be dangerous after the power is switched off.

The formula

U=12CV2

Energy stored in an ideal capacitor of capacitance C charged to the voltage V.

Symbol Quantity Unit
U Stored energy joule (J)
C Capacitance farad (F)
V Voltage across the capacitor volt (V)
Closing S lets the capacitor charge from zero up to ℰ; the current falls to zero as it does.

Where the formula comes from

Charge the capacitor one small step at a time. When it already holds the charge q, its voltage is v=qC, and moving a further charge dq from the negative plate to the positive plate takes the work:

dW=vdq=qCdq

The voltage grows in proportion to the charge already there, so each new step costs a little more than the last. Integrate while the charge rises from 0 to its final value Q. Nothing in the calculation depends on the shape of the plates or on time, so the result holds for any capacitor whose capacitance does not change with the applied voltage:

U=∫0QqCdq=Q22C=12CV2=12QV

The factor ½ comes from the average voltage during the charge, V/2: the first charge moves against almost no voltage, the last against the full V.

This holds only while the capacitance stays the same at every voltage, as for the ideal capacitor of circuit theory. When the capacitance changes with the voltage, the simple formula no longer applies, and the energy has to be computed from the integral of vdq with the real relation between charge and voltage.

Where the energy actually sits: the electric field

The energy is not in the plates, it is in the electric field between them. For a parallel-plate capacitor with plate area A and gap d in vacuum, the capacitance and the field are:

C=ε0Ad,V=Ed

Putting them into 12CV2 and dividing by the volume Ad between the plates gives the electric energy density, in joules per cubic metre, for a field E in vacuum:

u=UAd=12ε0E2

With a dielectric of relative permittivity κ between the plates, replace ε0 by the permittivity ε=κε0 of the material. The same expression then holds wherever an electric field exists, even when the field and the permittivity vary from point to point:

u=12εE2

Why charging through a resistor wastes half the energy

Charge the capacitor in the circuit above from a fixed EMF ℰ. At the end it holds the charge Q=Cℰ, and the battery has pushed all of that charge through its full EMF, so it has supplied:

Wbat=Qℰ=Cℰ2

The capacitor keeps only 12Cℰ2. The other half is turned into heat in the resistor. With the charging current i=ℰRe−t/RC, the energy dissipated in the resistor is:

WR=∫0∞i2Rdt=12Cℰ2

The resistance cancels out of the result. A smaller resistor gives a larger current for a shorter time, and the heat is the same. Even the resistance of the wires alone takes that half; the only way to do better is to charge with a voltage that rises with the capacitor, as some switching converters do.

What it means in a real circuit

A capacitor on a supply rail rarely gives all its energy to the load. The circuit it feeds stops working below some minimum voltage, so only the energy between the starting voltage V1 and that minimum V2 is usable:

ΔU=12C(V12−V22)

This is how a hold-up capacitor is sized: it keeps a processor running long enough to save its state or finish a flash write after the supply fails. Because the energy grows with the square of the voltage, storing the energy at a higher voltage and regulating it down is more effective than adding capacitance at a low voltage. Supercapacitors use the same formula with a much larger C at a low voltage per cell, for longer interruptions.

The stored energy also stays after the supply is removed. A bleeder resistor R across a large capacitor drains it with the time constant RC:

v(t)=V0e−t/RC

Energy scales matter here: OpenStax gives the example of a defibrillator that delivers 400J from an 8µF capacitor charged to 10kV. A capacitor bank in a power supply holds far less, but enough to cause burns or damage tools, so check that it is discharged before touching it.

Worked example

Sizing a hold-up capacitor for a microcontroller

A regulator feeds a microcontroller from a 5 V rail. After a power loss the regulator keeps the microcontroller running as long as its input stays above 3 V, and it draws a nearly constant 100 mW from the hold-up capacitor, losses included. What capacitance keeps the microcontroller running for 4 ms, enough for a last write to flash?

  1. Write down the data: V1=5V,V2=3V,P=100mW,t=4ms.
  2. Energy the load needs during the hold-up time: U=Pt=100mW×4ms=400µJ.
  3. Solve the usable-energy formula for the capacitance: C=2UV12−V22=2×400µJ16V²=50µF.
  4. Check: at 5 V the capacitor holds 12×50µF×5V2=625µJ, and only 1−925=1625 of it, about 64%, is used before the input reaches 3 V.

Result: At least 50 µF, before margins. Choose a larger standard value to cover the capacitance tolerance and its loss over time.

Common mistakes

  • Forgetting the factor ½. The energy stored is 12CV2, not CV2; CV2 is what the source supplies when it charges the capacitor.
  • Counting the full 12CV2 as available. The load only gets the energy between the starting voltage and its minimum operating voltage.
  • Expecting a smaller series resistor to reduce the charging loss. From a fixed voltage the resistor always takes half the energy; it only sets how fast.
  • Applying the formula when the capacitance is not constant. It needs a capacitance that does not change with the voltage; otherwise go back to the integral of the work. Tolerance and ageing also move the real capacitance away from the nominal value, and the energy with it.
  • Assuming a capacitor is safe once the supply is unplugged. Without a bleeder resistor it can hold its charge for a long time.
A circuit to design?

Our design house in Montpellier, France designs and validates electronic boards, from schematic to production.

Our electronic design expertise

Frequently asked questions

Why is there a factor ½ in the capacitor energy formula?
Because the voltage rises from zero while the capacitor charges. Moving a small charge dq costs v dq, and v grows in proportion to the charge already stored. Integrating from 0 to the final charge Q gives Q squared over 2C, which is one half of C V squared. The ½ comes from the average voltage during the charge, which is V/2.
Where is the energy of a capacitor stored?
In the electric field between its plates, not in the plates themselves. The energy density of a field E in vacuum is one half of the vacuum permittivity times E squared, and in a dielectric the permittivity of the material replaces it. Integrated over the volume of the field, this gives back one half of C V squared.
Why is half the energy lost when charging a capacitor?
A fixed source of EMF E moves the whole charge Q through its full voltage and supplies Q times E, which is C E squared. The capacitor keeps only half of that. The rest is dissipated as heat in the resistance of the charging path, and the result does not depend on the value of that resistance.
How much energy can a hold-up capacitor actually deliver?
Only the energy between its starting voltage V1 and the minimum voltage V2 at which the load still works: one half of C times the difference between V1 squared and V2 squared. Below V2 the remaining charge is stored but useless, so the capacitor must be sized on that difference, not on its total energy.
What does a bleeder resistor do across a capacitor?
It gives the stored charge a permanent path to discharge once the supply is removed. The capacitor voltage then decays exponentially with the time constant R times C. A lower resistance discharges the capacitor faster but wastes more power while the circuit is running, so the value is a compromise between the two.

Sources

Formulas and values checked against these published references.