Site: École Polytechnique Fédérale de Lausanne (EPFL)
Institut de Physique Experimentale (IPE)
Département de Physique
CH-1015 Lausanne-Ecublens, Switzerland
Tel: (41) 21-693 3320; Fax: (41) 21-693 3604
Date Visited: 14 October 1997
WTEC: H. Goronkin (report author), M.C. Roco
The École Polytechnique Fédérale de Lausanne (EPFL) is one of three federally funded research institutions in Switzerland. The other two are the Paul Scherrer Institute and Eidgenössiche Technische Hochschule (ETH) in Zürich. EPFL was founded in 1853 as a technical institute of the University of Lausanne. The Laboratory of Experimental Physics was founded in 1947 and has traditionally explored the boundaries between pure and applied science. Starting with activities in piezoelectric properties of crystals, nuclear magnetic resonance, and surface and thin films, EPFL turned its attention to nanoclusters in the 1970s. Today, the main focus is on nanoscale physics of clusters, surfaces, and nanoscaled materials.
André Chatelain has been involved in cluster physics research for 30 years. He asks the questions, "How many atoms are needed for such bulk-like properties as melting, magnetism, conductivity?" "How many atoms are required for the Curie Temperature to be exhibited?" Clusters are size-selected from a molecular beam and characterized with a Stern-Gerlach magnet, after which the deflection is measured and the clusters are accelerated through a column in which time of flight measurements are made. Connections between cluster and bulk properties such as hysteresis and coercivity have not yet been made.
Chatelain's group has developed techniques for fabricating carbon nanotubes in higher concentrations than previous methods. By fixing a single nanotube to a scanning tunneling microscope (STM) tip, currents as high as 1 mA have been obtained from the tube tip. Typical currents are in the 10 pA to 1 nA range. Using a phosphor screen, fluctuations in the spot position have been related to fluctuations in electron density over the surface of the nanotube tip. It is not known whether this arises from thermally induced structural alteration of the tip or changes in electron density due to local charging.
Klaus Kern has a large group working on self-organized growth of nanostructure arrays. The novelty of his approach lies in his use of periodic dislocation arrays that serve to isolate nucleating adatoms. This has been demonstrated using a Pt(111) substrate precovered with 1.5 monolayer of silver that forms a pseudomorphic layer, and a second Ag layer that forms a trigonal dislocation network. Subsequent Ag adatoms are repelled by the dislocations and form into a network of regularly spaced individual islands. Deposition is performed below 110 K. Kern states that these experiments open a new method to create almost monodispersed, regularly spaced, superlattice nanostructures using the natural properties of crystals.
Jean-Philippe Ansermet uses polycarbonate membranes with 6 x 108/cm2 pores with gold sputtered on the back side as a template for magnetic nanowires. Magnetic material is plated into the pores, which are 20-200 nm in diameter and about 6 Ám high. Ni or Co is plated in order to study anisotropic magnetoresistance (AMR), and layered materials are used for giant magnetoresistance (GMR) structures. A structure consisting of 300 ten-nanometer layers of Cu and Co gave a 40% GMR ratio at room temperature. One of the difficulties of this approach is making contact to individual wires. Ansermet masks the top of the membrane and introduces gold into the plating solution. Plating is halted when contact is made to one wire, or perhaps, a few wires.
Ansermet is considering how AMR is related to GMR. His approach uses the curling spin wall to separate domains in the wire. He explains that the curl avoids surface charge along the wire. He claims that if the spin flip length (~50 nm in Co/Cu) is less than the length of a domain, the system is an appropriate analog-to-GMR structure. His prototype experiments show that the AMR ratio is enhanced by using the curl domain wall.