This is a short article which was published in Oxford’s student popular science magazine, Bang!. See it in its illustrated glory here (page 20).

Things expand when heated and contract when compressed. Obvious – but not always true. In the past fifteen years researchers have produced some striking examples of counterintuitive materials that contract on heating (negative thermal expansion, NTE) or expand in one direction when uniformly compressed (negative linear compressibility NLC).

Bonds between atoms are like springs, so when given more energy by heating they stretch and a material expands giving ordinary, positive, thermal expansion. But this is only a first approximation, as the structure of a material is extremely important to how it behaves when heated or compressed. Thus, by tuning the structure, unusual properties including NTE or NLC can be produced.

Zirconium tungstate, Zr(WO4)2, was the material that reignited interest in NTE, as it displays it for over 1000K (a far wider range than previous materials). NTE in zirconium tungstate is explained by a ‘skipping rope’ mechanism. The individual bonds (the rope) still lengthen when heated, but an additional vibration twirls the bonds around, bringing the atoms at the end of the rope closer together producing an overall contraction.

The investigation of silver cobalt cyanide (Ag3CoCN6) revealed a material that not only NTE but also NLC. The ‘winerack’ structure of Ag3CoCN6 is key: the silver is bonded to two cyanides to form the ‘struts’ and the cyanides then attach to the cobalt ‘joints’. The rigid struts means that a contraction in one direction, whether due to pressure or contraction of weak silver-silver bonds on cooling, is accompanied by an equivalent expansion in the other direction. As the framework of the structure is extremely flexible, both expansion and contraction are colossal!

Materials that shrink when others expand allow fabrication of composites that do not change shape with temperature or pressure, useful for high precision in extreme conditions, e.g. satellite parts. More prosaically, zero thermal expansion materials would make for more robust dental fillings, though metal cyanides have an obvious drawback. NTE and NLC materials demonstrate how much novelty and strangeness can be found in the structures of solids and how the structures can be linked to surprising properties.