The Problem with Infrared Crystallography

Calculated Raman spectra comparing H and HD in the same  crystal phase.  The three groups of peaks in HD are expected from the different molecules H2, D2  and HD, but previous expectation was that each group should split like the pure hydrogen.
Calculated Raman spectra comparing H and HD in the same crystal phase. The three groups of peaks in HD are expected from the different molecules H2, D2 and HD, but previous expectation was that each group should split like the pure hydrogen.

Spectroscopy is often used to detect phase transitions in solids. The appearance or disappearance of a peak is regarded as evidence for a change in symmetry - a different crystal phase. This general assumption has been disproven in a recent paper by PhD student Ioan Magdau and supervisor Prof. Graeme Ackland.

Solids are held together by their electrons, so it is not expected that different isotope would have different crystal structures. For example, hydrogen and deuterium have very similar phase diagrams. However, in a recent paper it was shown that an isotopic mixture of hydrogen and deuterium had completely different Raman spectra from the pure elements.  This result is so outlandish that some people assumed the experiment was flawed: different numbers of vibrational modes are a classic signature of different crystal symmetry.  However, calculations in CSEC showed a different story: the disordered masses of the nuclei meant that one of the most fundamental assumptions of solid state physics, the so-called Bloch theorem - broke down.  In simple terms, instead of vibrations spreading throughout the crystal as waves, they become localised on a few molecules - and the number of modes seen depends on the ways that H and D atoms can be arranged nearby, rather than the crystal symmetry.

The paper is the first all-Edinburgh study to appear in the top physics journal Physical Review Letters this year.