Simulating Phase Change | Guest video by Vilas Winstein

Key Concepts:

• Phase Transitions: Changes in the state of matter (e.g., solid, liquid, gas) due to changes in parameters like temperature and pressure.
• Liquid Vapor Model: A discretized fluid model used to simulate phase transitions.
• Boltzmann Distribution: A probability distribution that describes the probability of a system being in a particular state as a function of the state's energy and the system's temperature.
• Microstate: A specific configuration of all the particles in a system, including their positions and velocities.
• Macrostate: The large-scale, observable properties of a system, such as temperature, pressure, and density.
• Entropy: A measure of the number of possible microstates corresponding to a given macrostate.
• Free Energy: A thermodynamic potential that combines energy and entropy, minimized at equilibrium.
• Chemical Potential: A measure of the change in free energy when particles are added to a system.
• Kawasaki Dynamics: A Monte Carlo algorithm used to simulate the liquid-vapor model by swapping particles between neighboring sites.
• Metastability: A state where a system remains in a non-equilibrium state for an extended period due to a lack of sufficient perturbation.
• Critical Point: The point on a phase diagram where the phase transition line ends, and the properties of the two phases become indistinguishable.
• Ising Model: A mathematical model of ferromagnetism, which is mathematically equivalent to the liquid-vapor model.
• XY Model: A model similar to the Ising model, but where the magnetic elements can point in any direction in a two-dimensional plane.
• Universality: The principle that the macroscopic behavior of a system is often independent of the microscopic details.

1. Introduction to Phase Transitions and the Liquid Vapor Model

• The video focuses on phase transitions, specifically using a discretized fluid model called the liquid vapor model.
• Phase transitions are not chemical reactions; they involve changes in how molecules interact (e.g., ice, water, steam are all H2O).
• Ice has long-range interactions due to its rigid structure, water has some (ripples), and steam has almost none.
• Phase diagrams show the phases of matter at different temperatures and pressures; crossing lines indicates a phase transition.
• Supercritical fluids exist at high temperatures and pressures, allowing a continuous transition between liquid and gas.
• The liquid vapor model simulates a fluid with blue pixels representing molecules and white pixels representing empty space.
• The simulation is controlled by two parameters: temperature (T) and chemical potential (μ).
• Chemical potential is used as a stand-in for pressure because pressure is difficult to implement directly in the simulation.
• Temperature controls the importance of energy, which in this simulation corresponds to how clumped up the molecules are.
• At high temperatures, the density of molecules varies smoothly with chemical potential, resembling a supercritical fluid.
• At low temperatures, a phase transition occurs, with the density abruptly changing from gas-like to liquid-like.

2. The Boltzmann Distribution and Its Implications

• The Boltzmann distribution is a key formula in statistical physics that gives the probability of a microstate (x) as proportional to exp(-E(x)/T), where E(x) is the energy of the microstate and T is the temperature.
• The video addresses why randomness is used, clarifying it's not quantum randomness but a proxy for uncertainty about the true microstate.
• It's impossible to track every particle in a system (e.g., 10^23 molecules in a teaspoon of water), so statistical approximations are necessary.
• The Boltzmann distribution explains how phase transitions arise by balancing energy minimization and entropy maximization.
• States with higher energy have lower probability, but the number of microstates at a given energy level (Ω(E)) also matters.
• The probability of observing a microstate with energy E is proportional to exp(-E/T) * |Ω(E)|.
• Entropy (S) is defined as the logarithm of the number of microstates at a given energy level: S = ln(|Ω(E)|).
• The most likely energy level is the one that minimizes the free energy (F = E - TS).
• At low temperatures, minimizing energy is dominant, leading to ordered phases. At high temperatures, maximizing entropy is dominant, leading to disordered phases.
• The simulation models intermolecular forces with a "sweet spot": molecules prefer to be close but not too close.
• Each pair of adjacent molecules has an energy of -1 (represented by a green diamond).
• At high temperatures, molecules behave like a gas (high energy, high entropy). At low temperatures, they form a droplet (low energy, low entropy).
• Energy and entropy are in competition, and temperature mediates this competition.

3. Derivation of the Boltzmann Distribution

• The derivation starts with an isolated system where all microstates have the same energy and are equally likely.
• When two isolated systems are brought into contact, their temperatures equalize.
• Temperature is defined as the quantity that equalizes when systems exchange energy.
• The number of combined microstates with energies E1 and E2 is |Ω(E1, E2)| = |Ω(E1)| * |Ω(E2)|.
• The combined entropy is S12 = S1 + S2.
• Energy flows from system 2 to system 1 if dS12 > 0, which leads to the condition dS1/dE1 > dS2/dE2.
• Equilibrium is reached when dS1/dE1 = dS2/dE2.
• Therefore, 1/T = dS/dE, meaning temperature is the inverse of the rate of change of entropy with respect to energy.
• To derive the Boltzmann distribution for a system at fixed temperature, it's placed in contact with a heat bath.
• The probability of a microstate X in the small system is proportional to the number of microstates in the heat bath compatible with X.
• This leads to P(X) ∝ exp(-E(X)/T), which is the Boltzmann distribution.

4. Simulation Implementation and Kawasaki Dynamics

• Sampling directly from the Boltzmann distribution is computationally expensive due to the exponential number of microstates.
• The simulation uses Kawasaki dynamics, a Markov chain Monte Carlo algorithm, to generate samples.
• At each time step, two pixels are chosen, and if one has a molecule and the other is empty, they are potentially swapped.
• The probability of swapping is determined by the ratio of the Boltzmann probabilities of the two microstates: P(X')/P(X) = exp(-(E(X') - E(X))/T).
• The probability of swapping is then P(swap) = exp(-ΔE/T) / (1 + exp(-ΔE/T)), where ΔE is the energy difference.
• The energy difference only depends on the neighboring pixels, making the calculation efficient.
• The algorithm can be thought of as a random walk through the space of possible microstates.

5. Introducing Chemical Potential and Exploring the Phase Diagram

• To create a two-dimensional phase diagram, chemical potential (μ) is introduced as a second parameter.
• Chemical potential is defined as the quantity that equalizes when systems exchange particles.
• μ = -T * (dS/dN), where N is the number of molecules.
• The probability of a microstate is now proportional to exp(-(E(X) - μN(X))/T), where N(X) is the number of molecules in microstate X.
• A positive chemical potential favors microstates with more molecules.
• The simulation is modified to allow adding or removing molecules at each pixel.
• The phase diagram is explored as a function of T and μ, showing gas, liquid, and supercritical fluid phases.
• At high temperatures, the density changes smoothly with chemical potential (supercritical fluid).
• At low temperatures, a phase transition line separates the gas and liquid phases.

6. Advanced Phenomena and Connections to Other Models

• Droplets or bubbles form during the phase transition because the density cannot change smoothly.
• The shape of the droplets depends on the temperature and the lattice structure.
• Metastability occurs when the system remains in the "wrong" phase due to a lack of sufficient perturbation.
• At the critical point, the system exhibits fractal-like structures and self-similarity.
• The Ising model is mathematically equivalent to the liquid-vapor model, with "up" spins representing molecules and "down" spins representing empty spaces.
• The XY model is similar to the Ising model but allows the magnetic elements to point in any direction, leading to the formation of vortices.
• The video concludes by noting that while simplified models can capture essential features of phase transitions, more complex models are needed for quantitative accuracy.

7. Conclusion

• The video provides a comprehensive overview of phase transitions, the liquid vapor model, and the Boltzmann distribution.
• It explains how phase transitions arise from the competition between energy minimization and entropy maximization, mediated by temperature.
• The video also demonstrates how to implement a simulation of the liquid vapor model using Kawasaki dynamics.
• Finally, it connects the liquid vapor model to other important models in statistical physics, such as the Ising model and the XY model.
• The principle of universality suggests that the macroscopic behavior of a system is often independent of the microscopic details.
• The second video will further simplify the liquid vapor model to the point where it can be completely understood using basic mathematics.