The Laws of Thermodynamics for Beginners
Thermodynamics has a reputation for being abstract, and part of the reason is that it is taught as four numbered laws that sound like legal clauses. They are not. Each one is a short, concrete statement about how energy behaves, and together they explain why engines waste fuel, why coffee cools, and why you cannot build a machine that runs forever.
The other reason students struggle is bookkeeping. Thermodynamics uses signed quantities, and two different sign conventions are in common use - one in physics, one in chemistry. Mixing them halfway through a problem produces answers that are wrong by exactly twice the work term. This guide states its convention up front, keeps it, and shows what each law lets you calculate.
The zeroth law: what temperature means
The zeroth law says that if two systems are each in thermal equilibrium with a third system, they are in thermal equilibrium with each other. It sounds like a triviality, and it was named zeroth because it was recognised as necessary only after the other laws had been numbered. But it is the law that makes thermometers possible.
Without it, 'temperature' would not be a well-defined property. Because equilibrium is transitive, every object carries a single number that predicts which way heat will flow when it touches another object, and a thermometer that has equilibrated with one body can be trusted to compare it with any other. Heat always flows from higher temperature to lower, and never the reverse without outside help - which is the second law arriving early.
The first law: energy accounting with signs
The first law is conservation of energy written for systems that exchange heat and do work: ΔU = Q − W. Here ΔU is the change in the system's internal energy, Q is the heat added to the system, and W is the work done by the system. That is the physics convention, and it is the one PhysRef uses. Its logic is that heat coming in raises the internal energy, while work going out lowers it.
Chemistry textbooks commonly write ΔU = Q + W instead, where W now means the work done on the system. Both are correct and describe identical physics; they differ only in who the work is credited to. The failure mode is switching between them mid-problem. Decide which convention you are using, write it at the top of your working, and check every sign against it.
A quick example in the physics convention: a gas absorbs 500 J of heat and expands, doing 200 J of work on its surroundings. Then ΔU = 500 − 200 = +300 J, so its internal energy rises by 300 J and, for an ideal gas, so does its temperature.
- Q is positive when heat enters the system, negative when it leaves.
- W is positive when the system expands and does work on its surroundings.
- In an adiabatic process Q = 0, so ΔU = −W: all the work comes out of internal energy.
- In an isothermal ideal-gas process ΔU = 0, so Q = W: every joule of heat in leaves as work.
- Around any complete cycle ΔU = 0, because internal energy is a state function.
The second law: why processes have a direction
The first law permits plenty of things that never happen. It has no objection to a cold room spontaneously warming your coffee, or to a car engine converting its fuel entirely into motion, because energy would be conserved either way. The second law supplies the missing direction: the total entropy of an isolated system never decreases. Real processes increase it, and only idealised reversible ones leave it unchanged.
Entropy is best read as a measure of how many microscopic arrangements match the same macroscopic state. For a reversible transfer of heat at absolute temperature T, the entropy change is ΔS = Q_rev/T. Note the division by temperature: the same joule of heat carries far more entropy when delivered to a cold body than a hot one, which is exactly why heat flowing from hot to cold increases the total.
Put numbers on it. Let 1000 J move directly from a reservoir at 600 K to one at 300 K. The hot reservoir loses 1000/600 ≈ 1.67 J/K while the cold one gains 1000/300 ≈ 3.33 J/K, a net increase of about 1.67 J/K. Run it the other way and the total would fall, which is why it never happens on its own.
Heat engines and the efficiency ceiling
A heat engine takes in heat Q_h from a hot reservoir, converts part of it into work W, and dumps the remainder Q_c into a cold reservoir. Over a full cycle the internal energy returns to its starting value, so W = Q_h − Q_c, and the efficiency is η = W/Q_h = 1 − Q_c/Q_h. The second law says Q_c can never be zero: some heat must always be rejected.
The best possible engine operating between two fixed temperatures is the Carnot engine, and its efficiency is η = 1 − T_c/T_h. Those temperatures must be absolute, in kelvin. Using Celsius here is the single most common error in the topic, and it produces answers that can exceed 100 percent or go negative - always convert first by adding 273.15.
- Identify the reservoirs. A steam plant runs its boiler at 500 °C and its condenser at 30 °C.
- Convert to kelvin: T_h = 773.15 K and T_c = 303.15 K.
- Apply the Carnot limit: η = 1 − 303.15/773.15 ≈ 1 − 0.392 = 0.608, so about 60.8 percent.
- Interpret it honestly. This is a ceiling no real plant reaches; friction, turbulence and finite-rate heat transfer push actual figures nearer 40 percent.
- Check a real engine against the ceiling: one taking in 1000 J and rejecting 600 J does 400 J of work, so η = 1 − 600/1000 = 0.40.
- Confirm the second law is satisfied. If that engine ran between 600 K and 300 K, the Carnot limit would be 1 − 300/600 = 0.50, and 0.40 sits safely below it.
The third law and absolute zero
The third law states that as the absolute temperature of a system approaches zero, its entropy approaches a constant minimum - zero for a perfect crystal, because there is only one way to arrange it. The practical consequence is that absolute zero cannot be reached in a finite number of steps: each stage of cooling extracts less than the last, so you can approach 0 K asymptotically but never arrive.
This law also does the quiet work of anchoring the Kelvin scale. Because entropy has a genuine zero point, absolute temperatures are ratios that mean something, which is what allows the Carnot expression to compare T_c with T_h directly. Laboratories now reach billionths of a kelvin, and the trend holds: each further step costs more.
Making the laws usable
When a thermodynamics problem stalls, it is almost always one of four things: an unstated sign convention, a temperature left in Celsius, a state function confused with a path function, or a system boundary that was never drawn. Internal energy, temperature and entropy are properties of a state, so their changes depend only on the endpoints. Heat and work are not - they depend on the route taken, which is why you cannot speak of the heat 'in' a gas.
A workable habit is to draw the boundary first, list what crosses it and in which direction, convert every temperature to kelvin, then apply the first law for the energy total and the second for whether the process is allowed at all. The thermodynamics category on PhysRef sets out the first law, entropy change, the two efficiency relations and the ideal gas law with their variables defined, and the constants page carries the molar gas constant and the Boltzmann constant at full precision when a calculation needs them.
Frequently asked questions
Is ΔU = Q − W or ΔU = Q + W?
Both appear in textbooks and both are right. In the physics convention, ΔU = Q − W with W the work done by the system. In the chemistry convention, ΔU = Q + W with W the work done on the system. Pick one, write it down, and never switch inside a single problem.
Why can no engine be 100 percent efficient?
Completing a cycle requires returning the working substance to its starting state, and that means rejecting heat to a cold reservoir. The second law forbids converting heat entirely into work in a cycle, so efficiency is capped at 1 − T_c/T_h even for a perfect engine.
Do I have to use kelvin in efficiency calculations?
For the Carnot expression, yes, always. It is a ratio of absolute temperatures, and Celsius values give meaningless results - sometimes above one, sometimes negative. Convert with T(K) = T(°C) + 273.15 before dividing.
What is entropy in simple terms?
A count of how many microscopic arrangements produce the same overall state. High entropy means many equivalent arrangements, so a system is overwhelmingly likely to drift toward those states. For reversible heat transfer, ΔS = Q_rev/T.
Why is the zeroth law called zeroth?
Because it was recognised as fundamental after the first, second and third laws had already been named, and it logically precedes all of them. It establishes that temperature is a consistent property, which every later law relies on.