Physics of Semiconductor Devices
One goal of the Physics of Semiconductor Devices project is to measure captured carriers within the band-gap of a semiconductor. If a carrier (a hole or electron) is captured and then released in the band, the capture center is called a trap. These traps arise from defects or impurities in the material. One technique for measuring these traps is through Deep Level Transient Spectroscopy, or DLTS. This technique essentially measures the change in capacitance of the semiconductor.
We are using a high‐frequency capacitance transient thermal scanning method to observe traps in the depletion region of our diode. The capacitance transients can give us information about trap concentration, activation energy and electron‐ and hole‐capture cross sections for each trap. These properties are of interest to industry.
Our circuit is built on a copper plating encased in a steel container. This is to reduce stray capacitance and noise from the high frequency signal (our signal is 30 MHz.) The plating provides a large common ground that will reduce these effects. In the future copper tape soldered near the input and output ports will further reduce these unwanted effects. Our diode is a schottky diode, which is used for it's fast switching properties.
The premise behind DLTS is to use a short pulse to momentarily forward bias a diode that is in a steady state of reverse bias. This short pulse will inject carriers (in our case electrons) into the depletion zone of our semiconductor (putting the diode in forward bias will shrink the depletion zone and allow electrons to enter). This will change the capacitance of the diode. Semiconductors are characterized by their unique conductive behavior. A semiconductor has a fermi energy that lies inside a band gap in electron energy levels. This band gap is small in a semiconductor however, allowing electrons to cross the band gap at relatively low temperatures (as compared to an insulator.) The rate at witch electrons are emitted from a trap is what will characterize the capacitance transient. The emission rate is proportional to a boltzman factor and depends exponentially on the energy difference between the trap level and the conduction band. In essence we will inject the depletion zone with electrons, thus raising the capacitance of the diode. Some electrons will become trapped. Thermal energy will eventually kick these electrons out of the traps and back into the charged portion of the semiconductor. The capacitance will spike when the electrons are introduced and slowly return to normal as the electrons are kicked out the traps. If the traps were not present we would expect a linear return to steady state capacitance.
Our circuit will mix a known voltage signal with another voltage signal passing through the diode. We have created a circuit that will split the incoming signal. On one leg of our circuit the voltage signal will run through the diode. The other leg consists of a 180 degree phase shifter that will recombine with the signal coming out of the diode, thus canceling the signal and outputting a signal of 0 amplitude. In reality we cannot exactly cancel out the signal coming out of the diode, so after the phase shifter we have a variable capacitor. This capacitor allows us to very minutely shift the phase of the signal. This will allow us to fine tune the phase of the signal we are using to cancel the signal coming out of the diode. At the end of the circuit both legs are attached to a variable resistor. There are voltage drops across both legs of the circuit and the variable resistor allows us to change the amplitude of the signals to find the ideal for cancellation. By adjusting the variable resistor and capacitor an input signal measured at 248 mV resulted in a 1.6 mV output.
The original signal will input into a mixer with the signal coming out of the circuit. The signals are multiplied in the mixer and the resulting signal will be output on an oscilloscope. During steady state the output coming out of he circuit should be close to 0 so no signal will be measured. When we introduce the pulse the signal on the bottom leg of our circuit will no longer cancel the signal coming through the top wave and a signal of interest will be measured. This will result in a spike of voltage that will gradually return to zero as the electrons are kicked out of the traps in the depletion zone of the diode. We will observe a voltage transient and from that we will be able to calculate the capacitance transient we are looking for.
As stated earlier, thermal energy is what releases electrons from the traps. Electrons are released faster as the temperature increases, thus the capacitance transient is "thinner" with respect to time for higher temperatures (it takes less time for the capacitance to return to zero). To make measurements we will
consider emission rates which are what define the capacitance transient. We can define emission rate "windows", a certain rate in the capacitance transient, to measure only emission rates that we define. For each window we can take measurements at many different temperatures and we will see peaks at temperatures (because emission rate is temperature dependent) that match the window we defined. If we change the window and repeat the measurement we will see peaks at different temperatures. If we plot the width of the peaks against 1/Temperature we will observe a straight line, the length of which will be the rate window we defined. By measuring several different windows at varying temperatures we will observe a long straight line. The slope of that line is the activation energy for the trap. We will use a cryostat with liquid nitrogen to measure the diode at different temperatures.
