density-functional theory
/ DFT /
Why is diamond hard and graphite soft, when both are pure carbon? Why does one candidate molecule make a good battery electrode and another a poor one? The answers live with the electrons — how they arrange themselves around atomic nuclei sets nearly every chemical and material property. Density-functional theory, DFT, is the workhorse method for computing that electronic arrangement on a computer, and it is one of the most-used calculations in all of science.
The exact quantum description of N electrons is a wavefunction in 3N dimensions — utterly intractable beyond a handful of electrons because of the curse of dimensionality. DFT's brilliant pivot, due to Hohenberg, Kohn, and Sham, is to prove that you do not need the full wavefunction: every ground-state property is fixed by the electron DENSITY, a single function of just three spatial dimensions. The Kohn-Sham scheme then replaces the interacting electrons with a set of fictitious non-interacting ones that share the same density, and solves their much simpler equations self-consistently: guess a density, build the effective potential it creates, solve the one-electron equations to get a new density, and repeat until input and output densities agree. Numerically this is a nonlinear eigenvalue problem solved by iteration, leaning on every tool in this subject — large eigensolvers, FFTs for plane-wave methods, and careful linear algebra.
DFT predicts molecular geometries, reaction energies, crystal structures, and electronic properties across chemistry, materials science, and physics, and earned Walter Kohn a share of the 1998 Nobel Prize in Chemistry. The crucial honest caveat is that DFT is exact in principle but approximate in practice: one ingredient, the exchange-correlation functional, is unknown and must be approximated. Standard approximations are remarkably good for many systems but can fail systematically — for band gaps, weak van der Waals forces, and strongly correlated materials — so a DFT number is a high-quality estimate, not a guaranteed truth.
Searching for a better battery, materials scientists screen thousands of candidate lithium compounds by DFT: each calculation predicts the voltage and stability of a crystal without ever synthesizing it. The few that look promising on the computer are then made and measured in the lab — turning a years-long trial-and-error hunt into a guided search.
DFT screens battery materials in silico — voltage and stability predicted before anything is synthesized.
DFT is exact in principle but its exchange-correlation functional is unknown and approximated, so it can fail systematically (band gaps, dispersion forces, strongly correlated systems). Treat a DFT result as a strong estimate to be checked, not a settled fact.