Ernst Ising
Solved a model in one dimension, concluded it had failed, and learned decades later that it had made him famous.
Biography
Ernst Ising was born in Cologne in 1900 and studied physics at Göttingen and Hamburg, where Wilhelm Lenz gave him a thesis problem: determine whether a chain of two-state elements favouring agreement with their neighbours could produce ferromagnetism. Ising solved the one-dimensional case exactly and published the result in 1925. The answer was no. The chain has no spontaneous magnetisation at any temperature above absolute zero — a domain wall costs a fixed energy but can sit anywhere, so entropy always wins and order never survives.
His result was correct, and his extrapolation was not. Ising reasoned that higher dimensions would behave similarly, concluded the model failed to describe ferromagnetism, and left the subject. He was right about the chain: the argument turns on a domain wall in one dimension costing a fixed energy, whereas in two dimensions the wall is a closed loop whose cost grows with the domain it encloses, which changes the verdict entirely. Onsager showed this exactly in 1944.
Ising taught in Germany until the Nazi racial laws cost him his position in 1933; he then taught at a Jewish school in Caputh until Kristallnacht, when it was destroyed. He and his wife Johanna fled to Luxembourg in 1939, where he survived the occupation working as a forced labourer on the railways. They emigrated to the United States in 1947.
He became a professor of physics at Bradley University in Peoria, Illinois, in 1948, where he taught for two decades and was by all accounts a devoted teacher. Only after arriving did he discover that in the years he had spent surviving, his name had become attached to what was by then among the most-studied models in physics; he had published nothing further in the field and never did. He died in 1998 at the age of ninety-eight.
Contributions
- 01Exact solution of the one-dimensional Ising model (1924 thesis, published 1925), proving no phase transition occurs at any T > 0
- 02The free-energy argument for why one-dimensional systems cannot sustain long-range order