Thermodynamics of Living Systems

Modern calorimetry of cell cultures and whole organisms (isothermal microcalorimetry, indirect calorimetry in clinical metabolic carts) directly measures the $\delta q/dt$ term of the steady-state balance, letting clinicians estimate basal metabolic rate and diagnose metabolic disorders (thyroid dysfunction, mitochondrial disease) from deviations in heat output relative to predicted values from body mass and composition. Systems biology and metabolic flux analysis use the open-system framework quantitatively: flux balance analysis (FBA) models a whole cell as a steady-state network in which every internal metabolite pool satisfies $d[X]/dt = 0$, the discrete analog of the continuous $dS_{sys}/dt=0$ condition used here, and computes which combination of reaction fluxes maximizes biomass production subject to that constraint — essentially an engineering application of the steady-state open-system idea to genome-scale metabolic reconstruction. In cancer metabolism and the Warburg effect, tumor cells shift toward aerobic glycolysis, producing lactate even in the presence of oxygen; understood thermodynamically, this reflects a different steady-state flux pattern with a different (higher) $\sigma$ per unit ATP produced, trading efficiency for speed of biosynthetic precursor supply — a live area connecting non-equilibrium thermodynamics directly to oncology. In mitochondrial bioenergetics, the proton-motive force across the inner membrane is a textbook example of $\Delta G$ maintained far from zero by continuous electron-transport-chain flux; uncoupling proteins (thermogenin in brown adipose tissue) deliberately increase $\sigma$ by short-circuiting the proton gradient, converting the free energy that would have driven ATP synthase directly into heat — a physiological, regulated dissipative structure used for non-shivering thermogenesis in hibernating and newborn mammals. Astrobiology and origin-of-life research use the same open-system entropy balance to argue that any candidate for "life" anywhere must be a persistent, entropy-producing, matter-and-energy-fluxing structure, giving a thermodynamic (rather than purely chemical) criterion for what to search for in planetary atmospheres and hydrothermal vent chemistry. Finally, synthetic biology and metabolic engineering explicitly compute cellular $\Delta G$ values (not just $\Delta G^{\circ\prime}$) when redesigning pathways, because a pathway step with a barely negative in-vivo $\Delta G$ is a flux bottleneck regardless of how favorable its standard free energy looks on paper — the practical, industrial payoff of taking the open-system, non-standard-state bookkeeping in this article seriously.

:::quiz[Cytosolic ATP hydrolysis has ΔG°′ = −30.5 kJ/mol. If cellular concentrations of ADP and Pi both happened to rise until [ADP][Pi]/[ATP] exactly equals exp(−ΔG°′/RT), what is the actual cellular ΔG for this reaction?]

• ΔG becomes more negative than −30.5 kJ/mol
• ΔG becomes exactly 0 kJ/mol, since the reaction has reached equilibrium ✓
• ΔG becomes positive, driving the reverse reaction spontaneously
• ΔG is unaffected by concentration and remains −30.5 kJ/mol

Ask the tutor: How does the steady-state open-system framework here connect quantitatively to the actual $\Delta G$ calculations used for individual metabolic reactions in Free Energy and Spontaneity?

Classical thermodynamics is usually taught for closed systems: a fixed amount of gas in a piston, a fixed number of moles of reactants in a sealed flask. The first law for such a system is $dU = \delta q - \delta w$, and the second law states that for any process, $dS \ge \delta q / T_{surr}$, with equality only for a reversible process. A living cell violates neither statement, but it is not a closed system in the relevant sense: nutrients, oxygen, ions, and water flow in; CO₂, urea, heat, and water flow out. To describe it we need the open-system (or "control volume") formulation, in which both energy and matter cross the boundary. Write the internal energy of the organism as $U_{sys}$. Over a small time interval $dt$, it changes for three reasons: heat exchanged with the surroundings $\delta q$, work done on/by the system $\delta w$, and energy carried in or out by matter flow, $\sum_k \mu_k \, dn_k$, where $\mu_k$ is the chemical potential of species $k$ and $dn_k$ is the net moles of $k$ crossing the boundary (positive in, negative out). The open-system first law is

The corresponding open-system second law tracks entropy the same way. Entropy of the system changes due to heat exchange (entropy carried by heat crossing the boundary, $\delta q/T$), matter flow (each mole of species $k$ carries its own partial molar entropy $\bar{s}_k$ across the boundary), and — critically — entropy produced internally by irreversible processes (chemical reactions running away from equilibrium, diffusion down gradients, heat conduction across finite temperature differences). Denote that internal production rate $\sigma \ge 0$; by the second law it can never be negative. Then

This is the master equation of nonequilibrium biological thermodynamics. It says the system's own entropy can go up, down, or stay flat — there is no restriction on the sign of $dS_{sys}$ — because entropy can be exported through the boundary via heat and matter flux. What the second law actually forbids is $\sigma < 0$: the internal production term must be non-negative, always. A cell that builds a highly ordered protein (locally lowering its configurational entropy) is completely consistent with the second law provided the associated $\sigma\,dt$ from ATP hydrolysis, coupled to a larger $\bar{s}_k\,dn_k$ export of heat and small molecules, keeps the total entropy of system-plus-surroundings increasing. A useful special case is the steady state, which is the operating regime of a healthy adult organism, a growing bacterial culture in balanced growth, or any cell maintaining constant composition. In steady state, all state variables of the system — temperature, volume, and the concentration of every metabolite — are constant in time:

Rearranging the second law at steady state gives the balance that every textbook chapter on bioenergetics is secretly built on:

That is: at steady state, the organism must export entropy to its surroundings at exactly the rate that internal irreversible processes generate it. This export is overwhelmingly dominated by (i) heat released to the environment (basal metabolic heat) and (ii) the entropy carried away by small, high-entropy waste molecules (CO₂, H₂O, urea) relative to the low-entropy, highly ordered nutrient molecules (glucose, amino acids, fatty acids) taken in. Erwin Schrödinger's famous phrase that organisms "feed on negative entropy" is a loose but essentially correct way of saying: nutrients arrive with low entropy per unit free energy, and waste leaves with high entropy per unit free energy, and this gradient is what is consumed.

The open-system first law, restated for constant-pressure biology where $\delta w = P\,dV$ plus any non-expansion (e.g., electrical, osmotic) work $\delta w'$:

where:

• $U_{sys}$ — internal energy of the organism/cell (J)
• $\delta q$ — heat exchanged with surroundings (J), positive into the system
• $P\,dV$ — pressure–volume work (negligible for most soft, incompressible biological work)
• $\delta w'$ — non-expansion work (e.g., ion pumping against a gradient, mechanical work of muscle)
• $\mu_k, dn_k$ — chemical potential and net molar flux of species $k$ crossing the boundary

At constant T and P, enthalpy $H = U + PV$ is the natural energy bookkeeping variable, and Gibbs free energy $G = H - TS$ the natural spontaneity criterion, exactly as for closed systems — but now applied along a continuous flux rather than to a single closed reaction vessel:

At constant $T$ this reduces to $dG_{sys} = dH_{sys} - T\,dS_{sys}$, and using the open-system entropy balance from above:

The entropy production rate $\sigma$ is the single quantity that separates a living, dissipative steady state from an equilibrium one:

where:

• $J_j$ — the flux of process $j$ (a reaction rate, a diffusive flux, a heat flux)
• $X_j$ — the thermodynamic force driving that flux (a chemical affinity $-\Delta G_j/T$, a concentration gradient, a $1/T$ gradient)
• the sum is over every irreversible process occurring inside the organism simultaneously (thousands of coupled reactions in a real cell)

Each term $J_j X_j$ is individually non-negative for an uncoupled process relaxing toward equilibrium, but in a living cell many processes are coupled: ATP hydrolysis (a hugely negative-affinity, fast process) is mechanically linked via an enzyme to protein folding or ion pumping (a positive-affinity, "uphill" process). The coupling permits an individual sub-process to run against its own local gradient, exactly as in The Hydrophobic Effect where burial of a nonpolar surface is entropically unfavorable for solvent-ordering alone but is more than compensated once counted with the rest of the system. What is protected is only $\sigma_{total} \ge 0$, the sum over all coupled processes — never any single term. Finally, the basal steady-state condition already stated is worth writing as an explicit rate equation, because it is what converts "the second law" into a testable number — a calorimetric heat output:

where:

• $\delta q/dt$ — rate of heat exchange with environment (negative for a warm organism in a cooler environment: heat flows out)
• $dn_k/dt$ — molar flux rate of species $k$ (nutrients in, negative; wastes out, positive by convention if defined as net outflow)
• $\sigma$ — entropy production rate (W/K), strictly positive for any living, metabolizing cell

This equation is the thermodynamic definition of basal metabolic rate: it is the continuous heat-and-matter export required to keep $dS_{sys}/dt = 0$ while $\sigma > 0$ inside.

Example 1 (foundational) — Steady-state entropy balance for a resting human. A resting adult has a basal metabolic rate of about 90 W (roughly 1900 kcal/day; this is a mid-range adult value — the classic 65 kg reference man is often quoted near 80–100 W depending on activity assumptions). At body temperature $T = 310$ K, essentially treat almost all of this power as heat released to a $T_{surr}=293$ K (20 °C) room. What is the human's rate of entropy export, and what does this say about $\sigma$? Set up the steady-state balance from Step 6 of the first derivation. The dominant term on the left is the heat-exchange term, $\dot q/T$, evaluated using the temperature at which heat actually leaves the body surface — take the skin/room boundary at an intermediate value, but for an order-of-magnitude estimate use core temperature $T=310$ K directly:

This is not $\sigma$ itself (the matter-flux terms, food in and waste out, also contribute, and doing the heat-transfer bookkeeping fully at the true skin–air temperature difference changes the exact number), but it is the dominant contribution and gives the right order of magnitude directly from the definition. Since the human is in steady state ($dS_{sys}/dt \approx 0$ over a day, ignoring growth or fat storage), the balance derived above requires

So a resting human generates roughly $0.3$ W/K of entropy internally, every second, for as long as they live, and exports it as body heat plus the entropy difference between food and waste. Limiting check: if $\dot q \to 0$ (say, in the limit of zero metabolic rate), $\sigma \to 0$ as well — consistent, since a system generating no heat and doing no chemistry is not producing entropy; this is the (unattainable, for a living organism) equilibrium limit.

Example 2 (applied) — Free energy of ATP synthesis against a physiological phosphorylation potential in exercising muscle. During intense exercise, muscle cytosolic concentrations shift: $[\text{ATP}] \approx 4$ mM, $[\text{ADP}]$ rises to about $\approx 0.02$ mM (still much less than ATP but higher than resting), and $[\text{P}_i]$ rises sharply to $\approx 8$ mM as ATP is consumed faster than oxidative phosphorylation can replace it. Compute the actual $\Delta G$ of ATP hydrolysis under these conditions at $T = 310$ K, using $\Delta G^{\circ\prime} = -30.5$ kJ/mol.

So exercising muscle, despite depleting ATP and accumulating Pi, actually holds ATP hydrolysis at a more negative free energy than resting cytosol (−56.6 vs. the resting estimate of about −49 kJ/mol computed earlier), because the drop in ADP dominates the log term less than the fact that Pi rose while ADP stayed comparatively low relative to ATP. Physiologically this matters: the immediate free-energy "punch" available per ATP hydrolyzed to drive the myosin power stroke is preserved, even as the absolute ATP pool shrinks — the cell trades pool size for a maintained, steeply negative $\Delta G$, consistent with $\sigma>0$ being sustained by continued (if lagging) oxidative and glycolytic flux. Limiting check: if ADP and Pi both rose to make the ratio equal the equilibrium constant $K_{eq} = \exp(-\Delta G^{\circ\prime}/RT)$, $\Delta G \to 0$ — the cell would be at true equilibrium for this reaction, unable to do further work; this is exactly the pathological end-state of severe ischemia, when ATP is nearly exhausted.

Example 3 (exam / competition level) — Minimum entropy export rate for a bacterium doubling every 20 minutes, and consistency with observed heat output. A bacterial cell ($E.\ coli$) has a doubling time of $\tau = 20$ min under optimal conditions, dry mass $\approx 3\times10^{-13}$ g per cell (about 0.3 pg), and roughly 60% protein by dry mass with average protein molecular weight 30 kDa. To make one new cell, the bacterium must synthesize on the order of $N_{aa} \approx 3\times10^{6}$ amino-acid peptide bonds (a standard estimate for the total protein content of E. coli, distributed among roughly 2.4 million protein molecules of ~250 residues average — consistent with total protein mass of about 1.8×10⁻¹³ g at 110 Da/residue). Estimate the minimum rate of entropy production associated purely with the informational/configurational ordering of translation, and compare it with measured bacterial heat output (~ tens of pW per cell during rapid growth), to see whether the biosynthetic ordering term is a small or large fraction of the total dissipation. Each peptide-bond formation during translation is coupled to hydrolysis of (at minimum) 4 high-energy phosphate bonds — 2 ATP-equivalents for amino acid activation (aminoacyl-tRNA synthetase, using ATP→AMP + PPi, formally 2 phosphoanhydride bonds) plus 2 GTP hydrolyzed on the ribosome (one at EF-Tu-mediated delivery, one at EF-G-mediated translocation). Using $\Delta G^{\circ\prime}_{ATP} \approx -30.5$ kJ/mol as a representative phosphoanhydride bond energy for all four:

Over one generation, total chemical free energy spent on translation alone:

If essentially all of this appears as dissipated heat (a defensible upper bound, since a large fraction of phosphoanhydride bond energy in translation is indeed dissipated rather than stored as configurational free energy of the folded protein, which is only ~kJ/mol per protein — a tiny correction), the associated entropy production per generation at $T=310$ K is

Convert to a rate using $\tau = 20$ min $= 1200$ s:

Compare with measured total heat output of a rapidly growing E. coli cell, of order 10–30 pW per cell (calorimetric estimates in the literature); at $T=310$ K this corresponds to a total entropy production rate of

The translation-associated term computed above ($1.6\times10^{-18}$ W/K) is roughly four orders of magnitude smaller than the total measured entropy production rate ($6.5\times10^{-14}$ W/K). Interpretation: even though protein synthesis is the biosynthetic process most often cited as "building order," the free energy spent on it is a small fraction of the cell's total dissipative budget; the overwhelming majority of $\sigma$ is generated by "housekeeping" dissipation — maintaining ion gradients, futile cycles, proofreading/error-correction steps (themselves entropy-producing by design, since kinetic proofreading trades extra ATP hydrolysis for lower error rate), and simple heat loss from thousands of independently near-irreversible enzymatic steps running continuously. Limiting check: if the cell stopped growing entirely ($\tau\to\infty$), $\sigma_{translation}\to 0$ while $\sigma_{total}$ would drop but not vanish (basal maintenance metabolism continues even in stationary phase), correctly separating growth-associated from maintenance-associated entropy production.

Ilya Prigogine (1917–2003) built the mathematical machinery that makes rigorous sense of "order maintained by dissipation." Trained in Brussels under Théophile de Donder, who had introduced the concept of chemical affinity as the thermodynamic driving force for reaction rate, Prigogine extended classical equilibrium thermodynamics into the regime of systems held permanently away from equilibrium by continuous flux — exactly the regime a living cell occupies. His central contribution, later recognized with the 1977 Nobel Prize in Chemistry, was the theory of dissipative structures: stable, ordered patterns (chemical oscillations, convection cells, and by extension living organisms) that exist only because they continuously dissipate free energy and export entropy, vanishing the instant the flux driving them is switched off. Prigogine's minimum entropy production theorem, valid near equilibrium in the linear regime, showed that steady states of open systems tend to settle into states that minimize $\sigma$ subject to the constraints imposed on them — a result that, while not universally applicable far from equilibrium (where living systems mostly operate), gave the first quantitative bridge between the abstract Clausius inequality and concrete, structured, flowing systems like a candle flame, a Bénard convection cell, or a bacterium. His 1955 monograph Thermodynamics of Irreversible Processes and later popular writing with Isabelle Stengers (Order out of Chaos, 1984) directly shaped how biochemists now talk about metabolism as a steady-state flux problem rather than an equilibrium one, replacing loose talk of organisms "fighting entropy" with the precise, testable open-system balance derived above.