On this page
Every student learns that a current heats a wire. In 1834 a self-taught physicist in Paris showed that this is not the whole story: at the joint between two different metals, a current can make things colder. Jean Peltier's result was so unexpected that other scientists first suspected a fault in his instruments. This post tells his story from his own paper and an 1847 memoir by his son, and shows how to calculate when a junction cools and when it heats.
Who was Jean Peltier?
The 1847 biographical notice written by his son, Ferdinand Peltier, records that Jean Charles Athanase Peltier was born at Ham, in the Somme region of France, on 22 February 1785, the son of a clog maker. At ten he took apart, cleaned and reassembled a clock. He was apprenticed to a watchmaker in Saint-Quentin and then in Paris, and was later hired by the celebrated watchmaker Breguet to work on chronometers, the hardest part of the trade.
He set up his own business in 1806. In 1815 he left trade, the notice says, and decided to re-educate himself, studying steadily for the first time. He worked on electricity and later on meteorology. He died on 27 October 1845.
What was he trying to measure?
His paper, "New experiments on the heating effect of electric currents", appeared in the Annales de chimie et de physique in 1834 (volume 56, from page 371). It opens with a doubt. Everyone had assumed that any current, through any conductor, raises its temperature. But earlier experiments had used strong currents. What happens with a very weak one?
He was trying to rank antimony and bismuth for electrical conductivity, using short thin bars and an extremely weak current from a zinc and copper thermocouple with one junction warmed by about 10 degrees. Every time he put a bismuth bar in the circuit, his galvanometer needle swung the wrong way. He covered the apparatus with a glass cage to rule out stray heat. The reversed reading stayed.
So he looked for a small temperature change at the junctions themselves. He built a sensitive thermoelectric clamp from bismuth and antimony couples to grip a conductor and read its temperature at chosen points.
What did he find?
At a joint between bismuth and copper, at the end where the current entered in one direction, the temperature fell. The other joint warmed. His table shows that as he raised the current, the cooling first grew, then shrank, passed through zero and became a rise. With antimony placed between two pieces of bismuth, he found alternating warm and cold points along the strip. He called it an unexpected result: cooling produced by an electric current. Some scientists, including Becquerel and de la Rive, at first suspected an induced current in his instruments.
The same physics, seen the other way round, is the effect Thomas Seebeck had found a decade earlier, where a temperature difference drives a current. Read our post on Thomas Seebeck and electricity from a temperature gap.
Worked example: when does a junction cool?
The NCERT Class 12 Physics chapter on current electricity gives the resistive heating as P = I²R. For the junction effect, use a simple model with sample numbers chosen to be easy to follow:
Heat removed at the junction by the Peltier effect: Π × I, with Π = 0.02 V.
Resistive heating in the nearby conductor: I²R, with R = 0.01 Ω, of which half reaches the junction.
Net cooling at the junction: Q = 0.02I − 0.005I² (in watts).
I = 1 A. Q = 0.02 − 0.005 = 0.015 W of cooling.
I = 2 A. Q = 0.04 − 0.020 = 0.020 W, the largest value.
I = 3 A. Q = 0.06 − 0.045 = 0.015 W. The cooling is shrinking.
I = 4 A. Q = 0.08 − 0.080 = 0. The two effects cancel.
I = 5 A. Q = 0.10 − 0.125 = −0.025 W. The junction now heats.
The pattern matches Peltier's table: cooling, a maximum, a zero, then heating. Setting dQ/dI = 0 gives 0.02 − 0.01I = 0, so the optimum current is I = 2 A. Reverse the current and the Peltier term changes sign, so the junction heats at every current.
Thermoelectric effects are not named in the JEE (Main) 2026 syllabus, but electrical energy and power are, and the maximum-finding step is ordinary calculus. Check the syllabus for your year.
Common mistakes with junction heating
Thinking a current always heats. Resistive heating does, but at a junction there is a second, reversible effect.
Treating both effects as I². The Peltier term is proportional to I, so it changes sign when the current reverses; I²R does not.
Assuming more current means more cooling. Past the optimum current, cooling falls and then becomes heating.
Ignoring the other junction. Heat taken from one joint appears at the other.
For a refresher on where resistive heat comes from, see our post on Joule and the mechanical equivalent of heat. If you get stuck setting up a maximum problem like this one, the Fermi Problem Library has a step-by-step solver that explains the reasoning rather than printing an answer.
What can a student take from how Peltier worked?
Distrust the rule everyone repeats. He asked whether "currents always heat" had really been tested with weak currents.
Chase the wrong-way reading. His discovery began as a needle swinging the wrong way.
Rule out the boring causes first. The glass cage came before any claim of a new effect.
It is never too late to study properly. He began systematic study in his thirties.
Practice routine: redo the worked example with Π = 0.03 V, find the optimum current and the current at which cooling stops, and sketch Q against I.
Next step: write down, in your own words, why reversing the current turns Peltier cooling into heating but leaves I²R unchanged.
