Coherent Raman detected electron paramagnetic resonance
The vast majority of optically detected magnetic
resonance experiments reported to date measure a change in some optical
property, such as circular dichroism or photoluminescence, when a magnetic
transition is power saturated by resonant microwave, or radio frequency,
radiation. Coherent Raman detected magnetic resonance experiments differ
fundamentally from this approach: instead of measuring changes in the
populations of the magnetic states, coherent Raman detected experiments measure
the microwave induced coherence between the magnetic states. Like a
conventional, microwave detected, electron paramagnetic resonance (EPR)
experiment, coherent Raman detected EPR measures the precession of the sample
magnetisation. These differences in the basic physics lead to a number of
important advantages in the application of the phenomenon as an analytical
spectroscopic technique as we discuss below.
There are, in principle, as many different types of
coherent Raman detected EPR as there are conventional incoherent experiments.
Each technique differs in the optical property that is measured and the
"coupling" mechanism between the magnetic and the optical transitions.
The experiments proposed here are the coherent analogue of the well established
method of magnetic circular dichroism (MCD) detected EPR. Indeed, provided that
the optical line width is larger than the microwave frequency, it is possible to
describe the experiments we propose as the microwave frequency modulation of the
circular dichroism by the precessing magnetisation. The theory of the experiment
is therefore a combination of the very well established theories of
magneto-optics and EPR.
The optical detection of electron paramagnetic resonance
EPR is a very well established and extensively used
technique for the investigation of condensed matter. It has the merit of being
non-destructive and provides sufficiently high spectral resolution for crystal
field effects and hyperfine interactions to be studied in detail. The
experiments thus provide information on the nature, location and charge states
of paramagnetic ions incorporated into organic and inorganic hosts, on defect
and dopant centres in crystalline materials and on the interactions between
magnetic centres. EPR is also one the most import probes of the electronic
structure of transition metal ion centres in proteins.
A major limitation of conventional EPR
spectroscopy, that is especially important in the study of modern semiconductor
devices, is the poor sensitivity when compared to many optical spectroscopies.
Commercial EPR spectrometers have sensitivities of order 1011 spins
at low temperatures for a g @
2 species with a linewidth of 10-4 Tesla (i.e. 1011 spins
per gauss). For epitaxial semiconductors, the active layers are typically of
thickness 10-6 m and of area 10-5 m2 or less.
Thus, magnetic centres in such specimens can only be observed if the
concentrations are of order of
1022 m-3, and then only if the linewidths are narrow.
Frequently, the concentrations of interest are less than this and the specimen
volumes smaller: as a result, conventional EPR has not been used extensively for
the study of semiconductor heterostructures. In contrast, MCD detected EPR has
been successfully used to study a wide range of paramagnetic species in
semiconductors.
Coherent Raman detected EPR as coherent spin-flip Raman spectroscopy
Spin-flip Raman spectroscopy is
described separately. Stokes and anti-Stokes radiation separated from the laser
excitation frequency by the energy of the magnetic splitting is generated in
this experiment. It is also possible to regard the coherent Raman detected EPR
experiment as an extension of spin-flip Raman in which the magnetic transition
is excited with microwave radiation to form a three wave mixing coherent Raman
experiment. Coherent Raman radiation is generated with frequencies equal to the
sum and difference of the laser and microwave frequencies. Such a modification
has three important effects. Firstly, the Raman radiation is often many orders
of magnitude more intense than the spontaneous experiment. Secondly, the
magnetic resolution of the experiment is limited by the microwave rather than
the laser line width. Thirdly, the Raman radiation is forward scattered; co-propagating
with the transmitted optical excitation. This allows the Raman radiation to be
measured with high efficiency. The coherent Stokes and anti-Stokes waves of the
Raman model are equivalent to modulation sidebands when the experiment is viewed
as a modulation of the light beam by the precessing magnetisation.
In the single previous report of coherent Raman
detected EPR in the literature, on donor spins in n-type CdS, the coherent
radiation was up to three orders of magnitude larger than the spontaneous
spin-flip radiation. Further, up to 6 % of the incident laser radiation was
converted to Raman light in a 1 mm thick crystal containing a donor
concentration of 2 x 1023 m-3. This could be detected with
a Fabry-Perot interferometer based instrument. The optical heterodyne instrument
can detect coherent Raman radiation @
1016 times weaker than a few mW of incident laser excitation. Noting
that the coherent Raman power is proportional to the square of the number of
chromophores, we find that the minimum detectable donor concentration in a @
1 mm
epitaxial layer of CdS would be @
1019 m-3. In comparison, conventional EPR would require a
concentration of @
4 x 1023 m-3 in a layer with an area of 10-5 m2.
Further, the optical experiment will have little difficulty in studying layers
with much smaller areas. Focusing to areas of @
10-11 m2 is possible with visible radiation. In principle
an absolute sensitivity of @
102 spins would be obtained, although in practice laser heating and
optical pumping effects may become limiting. Nevertheless, the combination of
extremely high absolute sensitivity and spatial resolution will be a very
powerful tool in studying the distribution of paramagnetic centres in
technological devices.
