Beyond Born-Oppenheimer

A Visual Guide to Non-Adiabatic Molecular Dynamics (NAMD)

Why Do We Need NAMD?

In most of chemistry, we assume electrons and nuclei are decoupled (the Born-Oppenheimer Approximation). But in photochemistry, charge transfer, and at metal surfaces, this breaks down. NAMD simulates these complex processes where electronic states mix and molecules take unexpected paths.

⚡️

Photochemistry

Simulating how light energy is absorbed and dissipated, like in vision or DNA photostability.

↔️

Charge Transfer

Modeling electron movement in solar cells, batteries, and organic electronics.

🔬

Material Science

Understanding reactions at metal surfaces, catalysis, and energy conversion.

The Landscape of Chemical Change

Potential Energy Surfaces

The Born-Oppenheimer breakdown is driven by the topography of Potential Energy Surfaces (PESs). Where these surfaces for different electronic states get close or intersect, non-adiabatic transitions happen.

An illustration of an avoided crossing and a conical intersection (CI), the primary funnels for photochemical reactions.

Adiabatic vs. Diabatic

Simulations require choosing a representation. The "adiabatic" basis is natural but has problematic singularities at intersections. The "diabatic" basis is smooth and numerically stable, making it ideal for dynamics.

Feature Adiabatic Diabatic
PES Shape Cusps / Avoided Crossings Smooth / Crossing
Coupling Derivative (Singular) Potential (Smooth)
Use Case Static Calculations Wavepacket Dynamics

The NAMD Simulation Toolkit

No single method is perfect. The choice depends on a trade-off between computational cost and the accurate treatment of quantum effects like branching and coherence.

Hierarchy of NAMD Methods

Methods range from efficient but approximate independent-trajectory approaches (like FSSH) to highly accurate but expensive wavepacket-based methods (like AIMS). The "best" method balances these factors for the specific problem.

The Machine Learning Revolution

The biggest bottleneck is calculating electronic properties. ML models can learn these properties, offering a speedup of over...

1,000,000x

... enabling simulations of larger systems for longer times than ever before. This is the future of NAMD.

How a Simulation Works: Trajectory Surface Hopping (FSSH)

Fewest Switches Surface Hopping (FSSH) is the most popular NAMD method. It balances efficiency and accuracy by evolving classical trajectories on single potential energy surfaces, allowing them to stochastically "hop" between surfaces in regions of strong coupling.

1. Propagate

Evolve nucleus on active PES. Evolve electron wavefunction.

2. Calculate Probabilities

Determine probability of hopping to other surfaces.

3. Stochastic Decision

Use a random number. Attempt a hop?

4. Hop!

If yes & allowed, adjust velocity, switch active surface.

NAMD in Action: From Sunlight to Life

Case Study: Organic Solar Cells

NAMD resolves the ultrafast processes in organic solar cells. Simulations reveal the competition between productive Singlet Fission (one photon → two excitons) and undesirable charge transfer, guiding the design of more efficient materials.

Case Study: DNA Synthesis

The enzyme RNR is essential for life, using a chain of proton-coupled electron transfer (PCET) steps. NAMD simulations classify each step, revealing a mix of non-adiabatic, adiabatic, and intermediate regimes, which is crucial for understanding the enzyme's kinetics.