State Functions and Path Functions

Climb a mountain by the steep north face or the gentle switchback trail. Either way, you end at the same summit, and your net change in altitude is identical. But your legs know the difference: the sweat and fatigue depend entirely on the path you took. This simple contrast — altitude versus exertion — is the heart of one of thermodynamics' most powerful ideas. Chemists need this distinction constantly. It tells us why internal energy and enthalpy can be tabulated in reference books, while heat and work cannot. It explains why chemists invent clever indirect routes, like Hess's law cycles, to calculate reaction enthalpies without ever running the reaction directly. This article defines state functions and path functions precisely. It introduces the mathematical tool that separates them: exact versus inexact differentials. You will see the physical picture, the core relations, and two fully worked examples. The ideas here set up everything that follows on work, heat, and internal energy.

Key takeaways

A state function depends only on the current state of a system, not on how that state was reached.
A path function depends on the specific route taken between two states.
Pressure, temperature, volume, internal energy, and enthalpy are state functions.
Heat and work are path functions; they are not properties of a state at all.
Exact differentials integrate to a state-function difference; inexact differentials do not.
Key idea: if a quantity's change is the same for every path between two fixed endpoints, it is a state function, and its infinitesimal change is an exact differential.

The core insight

A state function is a property whose value is fixed once the state of the system is fixed, so its change between two states is path-independent and given by simple subtraction of endpoint values, whereas quantities like heat and work depend on the process itself and must be calculated by integrating along the actual path taken.

The physical picture

Consider one mole of an ideal gas that starts at $p_1 = 1.00$ bar, $V_1 = 24.79$ L, $T_1 = 298$ K, and ends at $p_2 = 2.00$ bar, $V_2 = 12.40$ L, $T_2 = 298$ K (an isothermal compression). There are infinitely many ways to get from state 1 to state 2: a single-step compression against constant external pressure, a slow reversible compression, or a two-step process that first cools then reheats. The temperature, pressure, and volume at the end are identical in every case, so quantities built from them — internal energy $U$ , enthalpy $H$ , entropy $S$ — end up with the same total change no matter which route was used. Contrast this with the work done on the gas. Squeezing it in one abrupt step against a large constant external pressure does more work on the surroundings' account books than compressing it slowly and reversibly in tiny increments (this is developed quantitatively in the discussion of reversible versus irreversible work). The heat exchanged along the way differs correspondingly, because the First Law ties them together. Neither heat nor work has a "value" that belongs to state 1 or state 2 alone — they are properties of the journey, not the destination. A useful way to think about this: state functions are like altitude on a mountain — a single number attached to each location. Path functions are like the distance walked or the calories burned — they depend on the trail chosen, even between the same two endpoints. Systems, surroundings, and boundaries (covered earlier) define what we track; state versus path functions tell us how the properties of that system behave.

Figure

Two paths between the same two states, same ΔU, different q and w

The core relations

\Delta X = X_2 - X_1 \quad \text{(same for every path, if } X \text{ is a state function)}

where: $X$ stands for any state function such as $U$ , $H$ , $S$ , $p$ , $V$ , or $T$ ; the change depends only on endpoints 1 and 2.

\oint dX = 0

where: the cyclic integral of an exact differential vanishes because the system returns to its starting state.

q = \int_{\text{path}} \delta q, \qquad w = \int_{\text{path}} \delta w

where: $\delta q$ and $\delta w$ denote inexact differentials of heat and work; the notation $\delta$ (rather than $d$ ) flags that these are not exact differentials of any state function.

\Delta U = q + w

where: this is the First Law; $\Delta U$ is path-independent even though $q$ and $w$ individually are not.

dX = \left(\frac{\partial X}{\partial p}\right)_T dp + \left(\frac{\partial X}{\partial T}\right)_p dT \quad \text{is exact if} \quad \left(\frac{\partial}{\partial T}\left(\frac{\partial X}{\partial p}\right)_T\right)_p = \left(\frac{\partial}{\partial p}\left(\frac{\partial X}{\partial T}\right)_p\right)_T

where: this is the Euler reciprocity (mixed second-derivative) test for exactness, applicable to any state function $X(p,T)$ .

Derivation AWhy exactness guarantees path independence

Goal. Show that if $dX$ is an exact differential, the integral of $dX$ between two fixed states is identical for every path connecting them, while an inexact differential $\delta q$ generally is not.

Step 1. Suppose $X$ is a well-defined function of two independent variables, say $X(p, V)$ , describing the state of a closed system with fixed composition. Its total differential is by definition

dX = \left(\frac{\partial X}{\partial p}\right)_V dp + \left(\frac{\partial X}{\partial V}\right)_p dV.

Because $X$ is a genuine mathematical function of $p$ and $V$ , this differential automatically satisfies Euler's reciprocity relation: the mixed partial derivatives are equal regardless of order of differentiation. This is a property of any well-behaved function, not an extra assumption.

Step 2. Integrate $dX$ along an arbitrary path $\Gamma$ from state 1 $(p_1, V_1)$ to state 2 $(p_2, V_2)$ :

\int_\Gamma dX = X(p_2, V_2) - X(p_1, V_1).

This follows directly from the fundamental theorem of calculus for line integrals: because $dX$ is the differential of an actual scalar function $X$ , the integral telescopes to the difference of the function's values at the endpoints. The specific route taken by $\Gamma$ never enters the right-hand side.

Step 3. Therefore, for two different paths $\Gamma_1$ and $\Gamma_2$ joining the same states 1 and 2,

\int_{\Gamma_1} dX = \int_{\Gamma_2} dX = X_2 - X_1,

so the change in $X$ is identical for both paths. Equivalently, going around any closed loop (state 1 back to state 1) gives $\oint dX = 0$ .

Step 4. Now consider heat, $\delta q$ . There is no function $q(p, V)$ whose differential reproduces $\delta q$ in general — heat is defined operationally as energy transferred by virtue of a temperature difference, and its magnitude depends on how the process is carried out (reversibly or irreversibly, at constant $p$ or constant $V$ , and so on). No reciprocity relation of the form in Step 1 holds for $\delta q$ written as a combination of $dp$ and $dV$ ; the "mixed partial derivative" test fails. Consequently

\int_{\Gamma_1} \delta q \neq \int_{\Gamma_2} \delta q \quad \text{in general},

even though both integrals start and end at the same states 1 and 2. The same argument applies to $\delta w$ .

Step 5. Combine with the First Law, $dU = \delta q + \delta w$ . Experiment (Joule's and later, more precise, calorimetric work) shows that $dU$ is exact — the sum $\delta q + \delta w$ always integrates to the same $\Delta U$ for a given pair of states, even though the two pieces individually vary path by path. This is the deep empirical content of the First Law: heat and work are individually path functions, but a specific linear combination of them is not.

\boxed{\Delta U = \int_{\Gamma} dU = U_2 - U_1 \ \text{(any path)}, \qquad q_\Gamma \text{ and } w_\Gamma \ \text{individually depend on } \Gamma}

Worked examples

Example 1 — Confirming path independence of $\Delta U$ for an ideal gas, two different routes. Take 1.00 mol of an ideal monatomic gas ( $C_{V,m} = \tfrac{3}{2}R$ ) at $T_1 = 300$ K, $V_1 = 10.0$ L. It is taken to $T_2 = 300$ K, $V_2 = 20.0$ L (isothermal expansion, so $\Delta T = 0$ ) by two different routes. Route A: reversible isothermal expansion. For an ideal gas, $U$ depends only on $T$ , so

\Delta U_A = n C_{V,m}\,\Delta T = (1.00\ \text{mol})\left(\tfrac{3}{2}\right)(8.314\ \text{J K}^{-1}\text{mol}^{-1})(0\ \text{K}) = 0\ \text{J}.

The reversible work is

w_A = -nRT\ln\frac{V_2}{V_1} = -(1.00)(8.314)(300)\ln\left(\frac{20.0}{10.0}\right) = -(2494.2\ \text{J})(0.6931) = -1729\ \text{J}.

From the First Law, $q_A = \Delta U_A - w_A = 0 - (-1729\ \text{J}) = +1729$ J. Route B: free expansion (against zero external pressure) then isothermal. Suppose instead the gas first expands freely (irreversibly) from 10.0 L to 20.0 L doing no work ( $w = 0$ , and for an ideal gas $q=0$ too, since $T$ is unchanged and there's no restoring pressure), landing at the same final state.

\Delta U_B = n C_{V,m}\,\Delta T = 0\ \text{J} \quad (\text{same final } T),

w_B = 0\ \text{J}, \qquad q_B = \Delta U_B - w_B = 0\ \text{J}.

Check: $\Delta U_A = \Delta U_B = 0$ J — identical, confirming $U$ is a state function. But $w_A = -1729$ J while $w_B = 0$ J, and $q_A = +1729$ J while $q_B = 0$ J — completely different. This is exactly the behavior expected: the state function's change is route-independent; the path functions are not.

Example 2 — Using path independence to find an unmeasurable ΔH via a Hess's-law cycle. Suppose we want $\Delta H$ for the direct reaction

\text{C(graphite)} + \tfrac{1}{2}\text{O}_2(g) \longrightarrow \text{CO}(g)

which cannot be measured cleanly by direct combustion (some CO always further oxidizes to CO₂ in a bomb calorimeter). Because $H$ is a state function, we may instead combine two measurable reactions that connect the same overall reactants and products, regardless of the mechanistic path. Given (standard enthalpies of combustion/formation, 298 K):

\text{(i)}\quad \text{C(graphite)} + \text{O}_2(g) \rightarrow \text{CO}_2(g), \qquad \Delta H_i^\circ = -393.5\ \text{kJ mol}^{-1}

\text{(ii)}\quad \text{CO}(g) + \tfrac{1}{2}\text{O}_2(g) \rightarrow \text{CO}_2(g), \qquad \Delta H_{ii}^\circ = -283.0\ \text{kJ mol}^{-1}

Reverse reaction (ii) to put CO₂ on the left and CO on the right (reversing flips the sign):

\text{CO}_2(g) \rightarrow \text{CO}(g) + \tfrac{1}{2}\text{O}_2(g), \qquad \Delta H = +283.0\ \text{kJ mol}^{-1}.

Add this to reaction (i). The CO₂ produced in (i) is exactly consumed here, and the $\text{O}_2$ partially cancels ( $1 - \tfrac12 = \tfrac12$ mol O₂ net), leaving

\text{C(graphite)} + \tfrac{1}{2}\text{O}_2(g) \rightarrow \text{CO}(g),

\Delta H^\circ = \Delta H_i^\circ + (-\Delta H_{ii}^\circ) = (-393.5\ \text{kJ mol}^{-1}) + (+283.0\ \text{kJ mol}^{-1}) = -110.5\ \text{kJ mol}^{-1}.

Sanity check: the accepted literature value for $\Delta_f H^\circ(\text{CO}, g)$ is $-110.5$ kJ mol⁻¹, matching perfectly. This works only because $H$ is a state function: the enthalpy change for going from graphite plus oxygen to CO is the same whether the reaction proceeds directly (impossible to measure) or via the two-step cycle through CO₂ (both steps independently measurable). If $H$ were a path function like $q$ alone, this substitution trick would be mathematically illegitimate.

Limiting cases & common traps

Situation	Correct view
Ideal gas, isothermal process	$\Delta U = 0$ always, but $q$ and $w$ are generally nonzero and route-dependent
Cyclic process (system returns to start)	$\Delta U = \Delta H = \Delta S = 0$ , but net $q$ and net $w$ need not be zero
Constant-pressure process	$q_p = \Delta H$ only under these specific conditions — this makes $q$ look state-function-like, but it is a special-case equality, not a general one
Constant-volume process	$q_V = \Delta U$ , again only under these specific conditions, with no non-expansion work
Reversible vs irreversible between same states	$\Delta U$ , $\Delta H$ , $\Delta S$ identical; $q_{rev} \neq q_{irrev}$ and $w_{rev} \neq w_{irrev}$

The chemist behind it

The formal separation of state functions from path functions grew out of nineteenth-century arguments about heat engines and the nature of heat itself, but it was Rudolf Clausius who gave the concept its sharpest mathematical form. Building on Sadi Carnot's engine analysis and James Joule's mechanical-equivalent-of-heat experiments, Clausius recognized that while heat and work individually depend on process details, their sum — internal energy — behaves as a true function of state. He pushed the idea further by identifying an inexact differential, $\delta q_{rev}/T$ , that becomes exact when divided by temperature along a reversible path, defining what he named entropy. That later development builds directly on the exact/inexact distinction introduced here.

Origins · 1 minRudolf Clausius and how this was discovered

Rudolf Clausius

1822–1888 · German physicist who formalized the First and Second Laws and coined the term "entropy"

Ask the tutor: Can you show me how reversible versus irreversible compression of the same gas gives different work values, even though ΔU is identical?

What exactly makes a differential "exact"?

A differential $dX$ is exact if it is the total differential of some genuine function $X$ of the state variables — meaning its line integral between two points is path-independent and depends only on the endpoints. Mathematically, this is tested via Euler reciprocity: mixed second partial derivatives of $X$ must be equal regardless of the order of differentiation. If no such underlying function exists, the differential is "inexact," written with $\delta$ instead of $d$ , and its integral genuinely depends on the path chosen, as is the case for heat and work.

Is enthalpy always a state function, even for irreversible processes?

Yes. Enthalpy is defined as $H = U + pV$ , and since $U$ , $p$ , and $V$ are all state functions, any combination of them is also a state function, regardless of whether the process connecting two states is reversible or irreversible. The change $\Delta H$ between two equilibrium states is fixed by those states alone. What changes with reversibility is whether $q_p$ (heat at constant pressure) numerically equals $\Delta H$ , which requires no non-expansion work being done.

Why can't we just define a "state function" version of heat?

Heat is fundamentally a mode of energy transfer across a boundary, not a property possessed by a system in a given state. A system does not "contain" heat; it contains internal energy. Asking how much heat a system "has" is like asking how much distance a mountain "has" rather than how much distance a hiker walked to reach its summit. Only under the special lens of reversible transfer divided by temperature does a path-dependent quantity like $\delta q_{rev}$ become the exact differential of a new state function, entropy.

How does this connect to Hess's law?

Hess's law is simply the statement that enthalpy (and internal energy) are state functions, applied to chemical reactions. Because $\Delta H$ for going from reactants to products depends only on the identities and states of reactants and products, not on the reaction mechanism, chemists can add up enthalpies of intermediate steps — even hypothetical ones — to get the enthalpy of an overall reaction. This is exactly the calculation used in Example 2 to find $\Delta_f H^\circ$ of carbon monoxide.

Does path independence mean the process itself doesn't matter physically?

Not at all. The process matters enormously for heat, work, entropy production, and efficiency — reversible and irreversible paths between the same two states can have wildly different $q$ and $w$ , and only reversible paths achieve the maximum possible work extraction. Path independence applies strictly to state functions like $U$ , $H$ , and $S$ . The physical journey is still where all the interesting engineering and chemistry happens; it just doesn't affect the endpoint values of state properties.