Measurable macroscopic magnetic properties are controlled by microscopic
properties for which there are insufficient direct measurement techniques.
Technological advances leading to higher density magnetic information storage
and smaller magnetic devices are driving magnetic measurement requirements
further towards nanometer resolution. The development of a general method of
adding magnetic contrast to scanning tunneling microscopy is a widely
recognized challenge that would not only allow high spatial resolution magnetic
imaging, but would also allow correlation of magnetic microstructure with
topographic and spectroscopic properties measured by the STM.
In light emission from a metal-metal STM junction, electromagnetic modes of the
coupled tip-sample system are excited by inelastic electron tunneling as shown
in Fig. 1. The resulting dipolar radiation is expected to be linearly
polarized in a plane defined by the direction of the tunneling current and the
emitted light. Alvarado and coworkers reported that STM-stimulated luminescence
emitted in tunneling between a W tip and a Co ferromagnetic film had an
unexpected circularly polarized component, the sign of which was related to the
direction of magnetization of the Co. The promise of this intriguing result for
magnetic imaging on the nanometer scale led us to make further measurements of
this type. Our goals were 1) to test the generality of this effect by
measuring a different ferromagnetic material, 2) to eliminate surface
roughness as a possible source of change in circular polarization, and
3) to fully characterize the polarization of the tunneling-induced
luminescence to try to understand the underlying mechanism. To accomplish the
first two of these goals we use an Fe(001) whisker which we have shown with
RHEED and STM to be a high quality single crystal that has a very flat surface
with terrace widths of approximately 1 micrometer.
The polarization of the STM-induced luminescence for W(111) tips and samples of
Au(111) and Fe(001) was fully characterized by measuring the Stokes parameters
of the radiation using the apparatus shown in Fig. 2. For both Au(111) and
Fe(001), the luminescence is fully polarized within experimental uncertainty
and predominantly linearly polarized as expected for dipole radiation. There is
a tip-dependent circular polarization for both Au and Fe, which we attribute to
tip asymmetries. The small circular polarization component in the Fe data is
shown in Fig. 3, and is seen to be independent of sample bias. Most
importantly, there is no change in the circular polarization that can be
associated with a change in magnetization of the Fe(001) within an experimental
uncertainty of approximately ± 2%. A number of factors make the
measurement of the circular polarization of the luminescence vs sample
magnetization difficult: 1) low count rates, 2) circular polarization
from tip dependent asymmetries, and 3) necessity of protecting the tip
during sample magnetization reversal. Because of these factors, we have not
investigated other combinations of tip materials and ferromagnetic samples. We
conclude that measuring the circular polarization of STM-stimulated photon
emission from a metal-metal junction does not provide a general means to
achieve magnetic contrast in STM.

Figure 1: Schematic of the inelastic tunneling process leading to photon
emission in a metal-metal STM junction.

Figure 2: (a) rotaing quarter-wave retarder/fixed linear analyzer
techniqye for determining the Stokes parameters. (b) Schematic of the
experimental apparatus for polarization measurements of the STM-induced
luminescene.

Figure 3: The circular polaization measured for tunneling between W(111)
and Fe(001) as a function of sample voltage and sample magnetization is shown
by the filled squares and circles. The difference in the circular polarization
on reversing the sample magnetization is plotted against sample voltage at the
bottom. There is no magnetization-dependant effect within experimental
uncertainty.
![]()
Polarized light emission from the metal-metal STM junction
![]()
Daniel T. Pierce
Joseph A. Stroscio
Robert J. Celotta
![]()
Angela Davies - University of North Carolina (Charlotte)
Online: July 1999
Last Updated: February 2008
egpwebmaster@nist.gov