University Oral Examination: Scott Andrews (Tuesday, December 7, 10:00AM)
Scott D. Andrews
sandrew at stanford.edu
Fri Nov 26 16:04:12 PST 2004
University Ph.D. Oral Examination
Department of Materials Science and Engineering
Fabrication of Magnetic Nanopillars and X-ray Imaging of Spin-Transfer
Tuesday, December 7, 2004, 10:00 AM (Refreshments will be served at 9:45
AM) Center for Integrated Systems Annex (CISX), room 101
ABSTRACT Spintronics has generated much interest and research in recent
years. Conventional technology uses the electron charge to store and
transmit information. Novel devices with additional functionalities can
be made using the quantum mechanical spin in addition to the electron
charge. Potential spintronic applications include spin transistors,
quantum computers, and magnetic random access memory (MRAM). Currently,
giant magnetoresistance read heads, which transmit information using spin
polarized currents, are used extensively in the magnetic storage industry.
MRAM has started to impact the storage industry with its recent
introduction to the consumer market. It is expected to compete with flash
memory, and, to a lessor extent, dynamic random access memory (DRAM). MRAM
does not deteriorate with use like flash but both are nonvolatile. MRAM's
relatively fast speed will allow it to compete with DRAM in applications
where information needs to be stored without a continuous power supply.
Since MRAM and other spintronic applications are dependent on the
interaction of spin polarized currents with magnetic materials, this
phenomenon deserves further investigation in order to gain a more complete
The main focus of this work is to investigate spin-transfer torque, a
magnetic phenomenon in which spin polarized currents can be used to alter
the magnetic state of a ferromagnet. This torque can be used as the
dominant switching mechanism in systems where the current density is high.
However, at low current densities, this effect is overpowered by the
Oersted field, the classical magnetic field created by an electric
current. Unlike the spin-transfer torque which is proportional to the
current density, the Oersted field is proportional to the total current.
Thus, for the spin-transfer torque to be dominant, small structures are
necessary. In an attempt to meet the design criteria of samples that would
carry high current density but low total current, holes were drilled with
focused ion beam (FIB) into silicon nitride films and filled with a stack
of two ferromagnets separated by a nonmagnetic spacer. These samples were
analyzed using x-ray photoemission electron microspectroscopy (X-PEEM).
Due to magnetic uniformity issues in these samples, two other structures
were subsequently investigated. One uses a stencil method that allows easy
variations of the exact materials chosen. However, this sample design has
the disadvantage that the magnetic material that should be switched using
the spin polarized currents is not localized to the pillar structure being
tested. This creates extraneous GMR signals during electrical
measurements. Another sample, which was fabricated with the help of
Hitachi, was designed to isolate the magnetic materials to a small region
and does not show the convolved signal during electrical measurements.
Both of these samples were analyzed with a scanning transmission x-ray
microscope (STXM). Using a pump-probe configuration, the time dependent
effects of the spin-torque transfer were examined. In the last sample,
spin-transfer torque, combined with conventional Oersted switching, was
observed and analyzed. Such a direct observation of spin injection and its
time characteristics has never been achieved previously.
More information about the labmembers