1988 - the AMTMP (at that
time I did not call it AMTMP!) was developed by S.Grabtchak and
G.Novikov in the Institute of Chemical Physics of Russian Academy
of Science, Chernogolovka, Moscow Region, Russia. Worked well
for the first realization. (Is still there, I guess...) It took
several years to built it...
1993 - the second realization
of AMTMP. The method and the algorithm were improved by S.Grabtchak
and M.Cocivera in the University of Guelph, Canada. This setup
was designed in tree months.
1994 - a lot of work on
automation and S/N improvement (same authors, same place).
??? - where and when will be the next setup? (It's up to you!)
A semiconductor sample inserted into a microwave cavity can be
characterized by two parameters of a resonance curve namely a
resonance frequency and a cavity quality factor which is related
to a bandwidth at a half maximum of amplitude. This curve is shown
in black and corresponds to time t0 before the illumination. The
short illumination pulse creates extra carriers in the sample
and changes the real and an imaginary parts of dielectric constant.
This causes a perturbation of the resonance curve and manifests
itself in a shift of the resonance frequency and in the broadening
of the resonance curve. The resonance curve immediately after
the illumination is shown in magenta and denoted as time t1. During
the decay of excess carriers due to recombination processes the
resonance curve gradually returns to its initial position and
parameters. Two intermediate curves corresponding to instants
t2 and t3 after the excitation are shown in green and in red correspondingly.
One can see that there are two contributions to the signal observed
at the initial resonance frequency and this single kinetics is
what is measured in the traditional method of microwave photoconductivity.
The first due to the change in the bandwidth is known as a photoconductivity
and usually is related to a change in the imaginary part of dielectric
constant. The second due to the shift of the resonance frequency
is known as a photodielectric effect and usually is related to
a change in the real part of dielectric constant. If the precautions
against the shift of the resonance frequency are not made the
interpretation of the observed kinetics will be ambiguous. Even
if this is done and the observed signal is proportional only to
the change of the cavity quality factor it is not possible to
distinguish the conductivity due to free electrons before the
first trapping occurs from the conductivity of electrons after
reaching the thermal equilibrium with traps. This distinction
is of great importance because in the first case the kinetics
gives the true lifetime but in the second the observed time constant
is related to the rate of release from traps present.
The Advanced Method of Transient Microwave Photoconductivity (AMTMP)
permits registration of both the photoconductivity and the photodielectric
effect. Photoresponses are measured at a number of frequencies
around the initial resonance frequency. The typical behavior of
kinetics is shown at some frequencies. Next, the values corresponding
to the same time are chosen from all kinetics and plotted versus
frequency. As an example three such curves for time t1, t2, t3
after the excitation are shown. The last step is to realize that
each curve is a difference signal between the original resonance
curve with known parameters and the current resonance curve with
unknown parameters. Fitting the difference signal to the difference
between two Lorentzian curves gives values of resonance frequency
and cavity quality factor for each instant. Therefore, two kinetics
one of the shift of the resonance frequency and another the change
of the cavity quality factor are the final results of the AMTMP.
In fact the central part
of this method can be thought of as a kind of time-resolved spectrometer
with an excitation at one frequency (UV or visible range) and
responses measured at microwave frequencies (but within very narrow
range compare with typical spectrometers). The signal is produced
by the signal generator (illumination of sample in the microwave
cavity). It causes the absorption of microwave power generated
by a microwave source. The changes in a microwave power in time
are registered by a detector diode in an appropriate detector
mount (microwave power-to-voltage transducer). The voltage produced
by a diode is registered by the oscilloscope. The kinetics corresponding
to various microwave frequencies of interest must be measured
sequentially by the scope and stored for the subsequent analysis.
The schematic diagram of
the microwave system is shown in the next picture. The microwave
source was a 20 mW Wiltron 68137B synthesized sweep generator
(2.0-20.0 GHz) connected to a PC 486DX4 by an IEEE-488 interface.
A three port circulator WFX-C (Microwave Resources Inc.) was used
which simplified the tuning procedure of getting the Lorentz form
of a dark resonance profile. All microwave signals were measured
using a DT8012 (Herotek Inc.) detector diode with a time constant
corresponding to ~ 1.3 nsec. The signal from the detector diode
was recorded by a 100 MHz TDS 320 (Tektronix) oscilloscope connected
to the computer by an IEEE-488 interface. The time resolution
in the experiment was limited by a time constant of the cavity
used (about 60 nsec). Since it is possible to determine a time
constant for a physical process that is at least 2.2 times larger
than the maximum instrumental time constant, the minimum decay
time constant that could be determined was 130 ns. The light source
used in these experiments was an LN 1000 nitrogen (337 nm) pulsed
laser (Photochemical Research Assoc. Inc.) with a pulse width
of 600 ps. The maximum incident laser power was 6.5E15 photons/pulse
cm2 . The illuminated area of the sample had a rectangular rather
then a circular shape with dimensions 1.7x5.5 mm. It was chosen
to satisfy more the requirement to the specific sample geometry
(long narrow strip) after the illumination. Laser power was measured
using an ED-100A joulemeter (Opticon Corp.).