Thermodynamics & Heat Transfer — Carnot, Blackbody Radiation and Phase Diagrams

Why Thermodynamics Is Unavoidable

Every process in the universe — chemical, biological, mechanical, electromagnetic — is ultimately governed by thermodynamics. Thermodynamics cannot be derived from Newton's laws; it stands on its own logical foundation, built from observations about heat, work, and the arrow of time. Its laws are simple but their consequences are profound: machines cannot be perfectly efficient, disorder increases, and absolute zero is forever out of reach.

What makes thermodynamics uniquely hard to teach is that its key concepts — entropy, temperature, heat capacity — are statistical averages over enormous numbers of particles. No single particle has a temperature. An interactive simulation that lets you watch Maxwell-Boltzmann distributions reshape as you drag a temperature slider, or observe convection rolls form spontaneously in a heated fluid, makes these averages visceral in a way that equations alone cannot.

This spotlight tours the six thermodynamics simulations on the platform, showing you what each teaches and how they connect to the underlying physics.

Layer 1: Newton's Cooling Law

Coffee Cooling Simulation

The simplest thermodynamics experiment is also one of the most illuminating: a hot object placed in a cooler environment. Newton's cooling law states that the rate of heat loss is proportional to the temperature difference between the object and its surroundings. The result is exponential decay — the same mathematical form that governs radioactive decay, RC charging, and population growth toward a carrying capacity.

Newton's cooling:
  dT/dt = -k(T - T_ambient)
  Solution: T(t) = T_ambient + (T₀ - T_ambient)·e^{-kt}
  k = h·A / (m·c_p)       [cooling constant, s⁻¹]

Fourier's heat conduction (1D):
  q = -λ · dT/dx           [W/m², heat flux]
  ∂T/∂t = α · ∂²T/∂x²     [heat equation, α = λ/(ρ·cₚ)]

Thermal diffusivity:
  α = λ / (ρ·cₚ)           [m²/s]
  Copper: α ≈ 1.17×10⁻⁴ m²/s
  Water:  α ≈ 1.43×10⁻⁷ m²/s (820× slower)

Newton's law of radiation (Stefan-Boltzmann):
  P = ε·σ·A·(T⁴ - T_amb⁴)   [dominant at high T]
  σ = 5.67×10⁻⁸ W/(m²·K⁴)

The Coffee Cooling simulation lets you adjust initial temperature, ambient temperature, cup material (conductivity), and surface area. Watch the cooling curve shift from near-linear at small ΔT to a steep exponential at large ΔT. Toggle between pure conduction and radiation-dominated cooling to see why the Stefan-Boltzmann T⁴ term only matters at very high temperatures.

Layer 2: The Heat Engine Limit

Carnot Cycle Simulation

In 1824, Sadi Carnot proved something remarkable without any knowledge of atoms: no heat engine operating between two reservoirs at temperatures T_H and T_C can ever exceed the Carnot efficiency η = 1 − T_C/T_H. This bound is absolute. It does not depend on the working fluid, the engine design, or the ingenuity of the engineer — it follows from the second law alone.

Process        Type            Heat       Work
─────────────────────────────────────────────────────
1→2  Isothermal expansion    Q_H > 0    W₁₂ = Q_H
2→3  Adiabatic expansion     Q = 0      W₂₃ = -ΔU
3→4  Isothermal compression  Q_C < 0    W₃₄ = Q_C
4→1  Adiabatic compression   Q = 0      W₄₁ = -ΔU

Net work output:
  W_net = Q_H - |Q_C|

Carnot efficiency:
  η_C = W_net/Q_H = 1 - T_C/T_H   (temperatures in Kelvin)

Second law constraint:
  ΔS_universe ≥ 0
  |Q_C|/T_C ≥ Q_H/T_H  (Clausius inequality)

Real engine examples (η_C vs actual):
  Steam turbine:  T_H=600K, T_C=300K → η_C=50%, actual ~42%
  Car engine:     T_H=900K, T_C=300K → η_C=67%, actual ~25-35%

The Carnot Cycle simulation renders the full P-V diagram and T-S diagram simultaneously. Drag T_H and T_C sliders and watch the enclosed area — the net work — expand or shrink. Compare against an Otto cycle (the idealized gasoline engine) to see why diesel engines, with their higher compression ratios, achieve better real-world efficiency.

Layer 3: The Speed of Molecules

Maxwell-Boltzmann Distribution

Temperature is the average kinetic energy of molecules — but not all molecules move at the same speed. James Clerk Maxwell and Ludwig Boltzmann derived the probability distribution of molecular speeds in an ideal gas. This distribution is not Gaussian: it has a long high-speed tail, which explains why chemical reactions are possible even when the average molecule lacks sufficient energy.

Probability density:
  f(v) = 4π·n·(m/2πk_BT)^{3/2}·v²·exp(-mv²/2k_BT)

Characteristic speeds:
  v_p  = √(2k_BT/m)          most probable
  v̄   = √(8k_BT/πm)         mean
  v_rms = √(3k_BT/m)         root-mean-square

Relationship: v_p < v̄ < v_rms  (ratio ≈ 1 : 1.128 : 1.225)

Equipartition theorem:
  ½m⟨v_x²⟩ = ½m⟨v_y²⟩ = ½m⟨v_z²⟩ = ½k_BT
  Monatomic ideal gas: U = (3/2)Nk_BT
  Diatomic ideal gas:  U = (5/2)Nk_BT  (+ 2 rotational DOF)

Boltzmann factor:
  P(E) ∝ exp(-E/k_BT)         Arrhenius: k = A·exp(-E_a/k_BT)

The Maxwell-Boltzmann simulation lets you set gas temperature and molecular mass, watching the distribution curve shift and the live histogram of simulated molecular speeds align with the theoretical curve. The high-speed tail becomes visceral: even at modest temperatures, a small fraction of molecules carry enough energy to escape the liquid phase (evaporation) or overcome reaction barriers (chemistry).

Layer 4: Light from Hot Objects

Blackbody Radiation Simulation

Every object above absolute zero emits electromagnetic radiation. The spectrum of that radiation depends only on the object's temperature — a fact that classical physics could not explain, producing the "ultraviolet catastrophe." Max Planck resolved the crisis in 1900 by assuming energy is quantised in discrete packets — a revolutionary hypothesis that launched quantum mechanics.

Planck spectral radiance:
  B(λ,T) = (2hc²/λ⁵) · 1/(exp(hc/λk_BT) - 1)

Wien's displacement law:
  λ_max · T = 2.898×10⁻³ m·K
  Sun (T≈5778K):  λ_max ≈ 502 nm (green)
  Human body (310K): λ_max ≈ 9.35 μm (infrared)

Stefan-Boltzmann law:
  P = ε·σ·A·T⁴
  σ = 5.670×10⁻⁸ W/(m²·K⁴)
  Doubling T → 16× more power

Classical Rayleigh-Jeans (fails at short λ):
  B_RJ(λ,T) = 2ck_BT/λ⁴    → ∞ as λ → 0 (UV catastrophe)

Planck's fix: quantised oscillators
  ⟨E⟩ = hν/(exp(hν/k_BT) - 1)  instead of k_BT

The Blackbody Radiation simulation renders the Planck spectrum alongside the Rayleigh-Jeans classical prediction. Slide the temperature from 300 K to 30,000 K and watch the peak wavelength sweep from infrared through visible red, yellow, and white — exactly the colours of heating metal. The classical curve diverges catastrophically at short wavelengths while Planck's formula stays physical, making the quantum revolution feel inevitable.

Layer 5: Spontaneous Pattern Formation

Bénard Convection Simulation

Heat a flat layer of fluid from below and it remains conductive — heat diffuses upward without fluid motion — until a critical threshold, the Rayleigh number, is exceeded. Above that threshold, the stationary state becomes unstable and spontaneous convection rolls emerge, forming the beautiful hexagonal Bénard cells seen in heated oil or the solar surface granules.

Rayleigh number:
  Ra = g·β·ΔT·d³ / (ν·α)
  g  = gravity [m/s²]
  β  = thermal expansion coefficient [1/K]
  ΔT = temperature difference [K]
  d  = layer thickness [m]
  ν  = kinematic viscosity [m²/s]
  α  = thermal diffusivity [m²/s]

Prandtl number:
  Pr = ν/α   (momentum/thermal diffusivity ratio)
  Water: Pr ≈ 7    Air: Pr ≈ 0.71    Silicone oil: Pr ≈ 1000

Onset of convection:
  Ra_c ≈ 1708  (for rigid-rigid boundary conditions)

Roll wavelength at onset:
  λ ≈ 2d  (horizontal cells ~same width as layer depth)

Nusselt number (convective enhancement):
  Nu = Q_conv/Q_cond = 1 + C·(Ra/Ra_c - 1)^n   for Ra > Ra_c

The Bénard Convection simulation uses a finite-difference Navier-Stokes solver with Boussinesq approximation. Turn up the temperature gradient past the critical Rayleigh number and watch disorder suddenly crystallise into rolling structures. Increase it further and the rolls become turbulent — a vivid demonstration of how order emerges from boundary conditions, not design.

Layer 6: Materials and Phase Transitions

Alloy Phase Diagram Simulation

When two metals are mixed, their thermodynamic interactions create phase diagrams that govern which crystal structures are stable at each composition and temperature. Phase diagrams are engineering blueprints: steelmakers, aerospace alloy designers and microelectronics manufacturers all depend on them to choose processing temperatures and predict microstructure.

Phase diagram regions:
  Liquid        — homogeneous melt
  Solid L (α)   — A-rich solid solution
  Solid R (β)   — B-rich solid solution
  Liquid + α    — two-phase coexistence
  Liquid + β    — two-phase coexistence
  α + β         — sub-solidus two-phase region

Lever rule (fraction of phases in two-phase region):
  f_α  = (C_β - C₀) / (C_β - C_α)
  f_β  = (C₀ - C_α) / (C_β - C_α)
  C₀ = overall composition, C_α, C_β = phase boundaries

Eutectic point:
  Minimum melting temperature in the phase diagram
  At eutectic: liquid → α + β simultaneously
  Sn-Pb solder eutectic: 63% Sn, 37% Pb, T_e = 183°C

Gibbs phase rule:
  F = C - P + 2   (C components, P phases, F degrees of freedom)
  At eutectic:    F = 2 - 3 + 2 = 1  (fix P, T is determined)

The Alloy Phase Diagram simulation renders a full binary eutectic diagram and traces your cursor through phase regions with a real-time Lever-rule readout. Set initial liquid composition, set cooling rate, and watch the microstructure form: primary α dendrites crystallise first, enriching the remaining liquid in component B until it hits the eutectic and freezes simultaneously.

Complete Thermodynamics Collection

Six simulations cover the major branches of classical thermodynamics:

Cross-Collection Connections

Thermodynamics reaches into nearly every other category on the platform. The same Maxwell-Boltzmann distribution that governs gas kinetics also determines the rate constants in the Enzyme Kinetics simulation via the Arrhenius equation. The Fourier heat equation is mathematically identical to the diffusion equation driving the Crystal Diffusion simulation — replace temperature with concentration and the physics is the same. Bénard convection is a close cousin of the Atmospheric Convection simulation, where the Coriolis effect breaks the symmetry of convection cells into cyclones. Even the Ising Model is a thermodynamic system — its phase transition at the Curie temperature is a textbook second-order transition governed by the Boltzmann factor.

Algorithms & Methods in This Collection