An international team of researchers led by Brookhaven National Laboratory (BNL) has elucidated and reported the mechanisms behind the giant isotropic negative thermal expansion (NTE) observed in yttrium-doped samarium sulfide (SmYS) below room temperature. The results, published in Physical Review Letters, suggest a path toward controlling the thermal expansion of SmYS and other materials with similar electronic structures.

One of the key challenges in materials design and engineering is precise control of thermal expansion across a wide range of temperatures. When materials that display typical positive thermal expansion are formed into composites with NTE materials, positive thermal expansion can be significantly reduced or eliminated. Although researchers have discovered several materials that exhibit NTE in the last few decades, a comprehensive understanding of its origins has remained elusive.

Giant NTE has been observed in SmS upon sufficient doping with yttrium or under hydrostatic pressure and at low temperatures. Under such conditions, the black-colored, semiconducting material collapses isotropically into a golden-colored metallic material of reduced volume.

While studying the electronic properties of SmYS using synchrotron-based x-ray spectroscopy, a research team led by BNL’s Daniel Mazzone and Ignace Jarrige saw an anomaly. “[W]e noticed that golden SmS showed an unusually large decrease of the Sm valence at low temperatures—more than anything we had seen before,” says Mazzone. SmS has a partially filled f-electron shell that can give rise to unusual materials properties. This prompted the team to probe a connection between the valence decrease and giant NTE.

The researchers synthesized single crystalline SmYS samples with Y substituting 0, 14, 23, and 33% of the Sm atoms. Samples with Y alloy concentrations of 0 and 14% relative to Sm were black, while samples with Y concentrations of 23% and 33% were golden-colored. This indicated that samples with 23% and 33% Y concentration were metallic in nature.

Utilizing synchrotron facilities in France, Japan, and the United States, the team characterized the samples and measured the unit cell volumes as a function of temperature from 10 K to 300 K. The black samples experienced a conventional positive thermal expansion of 0.6%. The golden samples experienced significant negative thermal expansion, with volume changes of 3.3% and 1.5% for the respective Y substitutions of 23% and 33%.

The team then examined how Sm valence varied with temperature and Y-substitution using a form of high-resolution x-ray absorption spectroscopy. The black samples had a stable valence across the whole temperature range. When the Y-substitution reached a threshold value between 14% and 23%, the valence increased sharply due to volume collapse, as expected, but then decreased with temperature. The decrease was gradual, but most pronounced in the golden sample with 23% Y substitution, the sample that also showed the largest NTE effect.

Because of its partially filled f-electron shell, SmYS is subject to the Kondo effect—wherein conduction electrons interact with magnetic impurities in the material and often lead to unusual material behaviors. Inspired by similarities to a Kondo effect-driven model that causes a key phase transition in metallic cerium, the team applied the same model to their SmYS data. The model was able to describe much of the observed behavior, with one significant exception. The model predicted a sharp transition point to NTE, but the experimental data showed continuous NTE over a 200 K range. However, when the researchers accounted for the fact that Y-doping increased local disorder, the transition broadened and the NTE was diluted. From this they concluded that the observed NTE was the result of both the Kondo effect and a strong dependence on alloy concentration.

Given these results, the team expects that NTE temperature and amplitude in SmYS can be tuned by adjusting the Y concentration. In addition, they suggest that NTE may be associated with the Kondo effect in other similar materials, such as thulium- and ytterbium-based compounds, and that it may be possible to tune the thermal expansion of these materials with alloy concentration as well.

“Through their careful measurements, Mazzone and co-workers have explicitly validated, perhaps more clearly than ever before, an important consequence of the coupling of electronic and lattice degrees of freedom in a material like SmS, namely that this coupling can produce a controllable negative thermal expansion,” according to Joe Thompson, a researcher at Los Alamos National Laboratory who specializes in the physics of strongly correlated electron materials.

“Perhaps more importantly, this paper raises the fundamental question of what crystal chemistry controls the extent to which these dominant degrees of freedom are coupled,” says Thompson. He notes that answering this question would offer a theoretically predictive framework that could pave the way for new and improved technologies.

Read the abstract in Physical Review Letters.