"Correlated Electron" Technology
|
How can a correlated material undergo a metal-to-insulator transition?
Suppose we have a crystal with a simple band structure consisting
of a single half-filled valence band, and we start to increase
the lattice spacing between the atoms. Band theory then predicts
metallic conduction, no matter how narrow the valence band becomes
(as measured by the band width W).
|
Top: Schematic representation of the lattice-constant
dependence of a simple band structure. Band theory predicts metallic
conduction, no matter how narrow the band becomes. Bottom: At a critical
lattice constant, itinerant electrons become mostly localized at individual
sites, but can still hop from site to site with a nonnegligible probability. |
What happens instead, at some
critical band width, is that the itinerant (delocalized) valence
electrons become mostly localized at individual atomic sites, such
that they only hop from site to site with a small, but nonnegligible
probability. This breakdown of metallic conduction is reflected
in the formation of the so-called Hubbard subbands, which are separated
by U, the energy cost an electron has to pay when it hops on an
already occupied atom. Quite generally, these Hubbard subbands
form gradually as the ratio U/W is increased.
Left:
Hopping of a spin-up electron and a spin-down hole on the background
of spin-down electrons gives rise to the so-called Hubbard subbands,
separated by the Hubbard interaction term U. Right: The Hubbard subbands
form gradually as the ratio U/W is increased.
To reveal such drastic changes in the electronic structure of
a material, ARPES is the ideal tool, since it directly probes the
valence-electron distribution in momentum space. The Kiel–Berkeley
collaboration performed their ARPES measurements at the Electronic
Structure Factory on ALS Beamline 7.0.1. Clean surfaces of layered
TaS2 were prepared by cleavage in ultrahigh vacuum, and an alkali,
rubidium in this case, was evaporated live during the photoemission
measurements.
A "photoemission movie" recorded during the deposition
process shows the intriguing evolution of the valence spectra as
a function of deposition time. At the beginning of the movie, the
high spectral intensity near the highest occupied electron energy,
the Fermi energy EF, identifies TaS2 as a metal. Then rubidium
deposition is started and spectral weight is continuously removed
away from EF, gradually transforming the material into an insulator.
The similarity of these results with the simple sketch of Hubbard
subbands strongly supports the interpretation as a correlation-driven
metal-to-insulator transition (note that ARPES can only probe the
lower Hubbard subband). But what is the reason for this transition,
as alkali adsorption onto metallic systems usually causes an increased
band filling and thus more metallic behavior?
Left: ARPES movie recorded during the deposition of rubidium
atoms on the surface of the layered transition-metal compound TaS2.
Right: An artist's impression of the Rb intercalation process in
TaS2. The possible solution lies in the characteristic layer structure
of TaS2. Remarkably, alkali deposition on such a material can result
in the intercalation of some alkali atoms in the gaps between the
layers, which increases the layer separation and thereby reduces
the electronic band width perpendicular to the layers. Thus, the
ratio U/W can be driven through the critical value for the metal-to-insulator
transition.
Besides electronic correlations, there are more facets of this
particular metal-to-insulator transition, namely, disorder introduced
by the rubidium atoms, charge transfer from rubidium to TaS2, and
electron–lattice coupling, as TaS2 also undergoes a charge-density-wave
reconfiguration. But electron–electron interaction appears
as the dominant cause in this case, making the rubidium/TaS2 system
an ideal test object to study in detail the continuous evolution
of the valence electron structure of a correlated material across
the passage from the metallic to the insulating regime.
Research conducted by K. Rossnagel and L. Kipp (University of
Kiel, Germany) and H. Koh, E. Rotenberg, and N. Smith (ALS).
Research funding: Deutsche Forschungsgemeinschaft, Alexander von
Humboldt Foundation, and U.S. Department of Energy, Office of Basic
Energy Sciences (BES). Operation of the ALS is supported by BES.
Publication about this research: K. Rossnagel, H. Koh, E. Rotenberg,
N.V. Smith, and L. Kipp, "Continuous tuning of electronic
correlations by alkali adsorption on layered 1T-TaS2," Phys.
Rev. Lett. 95, 126403 (2005).
ALSNews Vol. 263, March 29, 2006 |