Schottky Diode Lab Report

INTRODUCTION

A Schottky diode is the common name for a metal-semiconductor junction, it is the work function between the semiconductor and metal that determines whether the junction is ohmic or rectifying [1]. This lab will focus on the fabrication of a Schottky diode and the characteristics they possess. In principle, Schottky diodes are rectifying as the current can only flow one way. They can also handle high frequencies and have lower power loss which makes them ideal for many applications [2]. These include being extensively used in power electronics, general purpose rectifiers and due to its non-linear I-V curves, it is able to be used as a varistor which in turn can be used for voltage suppression [2]. Also there is a low capacitance of the device which makes it more ideal than alternative diodes.

OVERVIEW OF THEORY

An Ohmic contact is needed on the bottom part of the sample as without this, any measurements taken would create a very high resistance as the semiconductor and the metal probe used to measure would create two diodes facing each other, therefore no current could flow. This is overcome by diffusing indium and germanium onto the semiconductor wafer. Gold is also used to prevent any oxidisation on the sample. The wafer is firstly placed into the tube furnace with the gold, indium and germanium on it then heated at 420°C for 90 seconds. This is to enable quantum tunnelling so there is little resistance and only the semiconductor and top metal are measured.

Figure 1 shows the energy band diagram of a metal-semiconductor junction, which is rectifying as the work function of semiconductor is higher than that of the metal.

Figure two is the Schottky diode characteristics and when compared to figure 3, the PN diode characteristics, it shows that the turn on point is lower in the Schottky diode, which means less power is needed to operate, however there is a higher leakage current in reverse bias.

DEVICE FABRICATION

When working with such easily influenced devices, safety must be upheld at every milestone. This is first enforced by the dress code that is required in the laboratory. Each person must wear an overcoat, safety glasses, gloves, overshoes and a hair net. Due to e

ven the smallest particles being able to affect each sample, the dress code is necessary to ensure that very little dust gets into the clean room. Also, there are many different chemicals in the room therefore more precautions are needed when working with them. All guidelines set out by the Control of Substances Hazardous to Health (COSHH) regulations must be vigorously followed. Using these guidelines, all chemicals must be used inside a fume cupboard as any spillages will be contained and the fumes can be collected. There are also hot plates and other specialised equipment that should be handled with care to ensure no damage comes to either the user or the machine itself.

The cleanroom is comprised of 3 different areas. The first is where the entrants of the cleanroom put the safety equipment on, the second is the main laboratory area and the final is the yellow room. Each room is connected to the next with interlocking doors that only open when all others are closed.  This is because each has a different pressure level due to the need of having to keep as little dust in the main room to stop the contamination of samples. The high pressure level in the main room help to filter out any unwanted particles in the room. The air flow is also monitored to ensure that if there is a drop in pressure, the right actions can be taken to rectify any issues. The yellow room has special lighting due to the sensitivity of the devices that will be later described, however the main difference is that both red and blue colours have been taken out as the devices are easily affected by UV lights. The clean room being used for this experiment is classified as Class 6, this refers to the particle size compared to the maximum particle density.

Once all safety aspects have been covered, the process to create the device can begin. The first step is cleaving. This is where the wafer is cut into small sample sizes, usually 3x3mm, as the wafer is very expensive and only a small amounted is needed to complete the experiment. A special machine can be used to perform this however it is also achievable by hand.

When the sample has been cut to the necessary size, it is essential that a three stage solvent cleaning method is used. This is to ensure that there is little to no dust on the sample. To complete the three stages, the sample must first be placed into a beaker of N-Butyl Acetate (NBA) that has been heated on a hotplate. Once it has been in the beaker for a short amount of time it must be removed and then cleaned with a cotton swab that has also been dipped in the NBA. This is done by rolling the swab forward whilst pulling back so that as many particles as possible are collected. The sample is the turned 90° and swabbed again to collect any particles that may have been missed. This process is again completed but with the sample being dipped in Acetone and then Isopropyl Alcohol to complete the three stage cleaning. After this has happened, the sample is then dried with Nitrogen gas.

Due to the small size of the sample, it is much easier to handle on a larger material, such as a glass slide. This is done by heating the slide on a hot plate then melting wax onto it then placing the sample on that. This is then left to cool and will now be ready for the next stage.

The next stage is photolithography. This is required to put a pattern onto the sample so that testing of the device can happen and is exclusively completed in the yellow room. As mentioned earlier, the stages throughout this part will mean that the sample will be affected by UV lights so it is necessary to complete this in the designated room. Firstly, the sample is placed on a hot plate for one minute. Then it is placed onto the spinner and spun at 4000rpm for 30 seconds and dried with nitrogen gas. Now a few drops of photoresist are added, this is the solution our sample will be printed with. Again the sample is placed in the spinner 30 seconds and then soft baked on a hot plate for exactly 1 minute. If the sample is baked for too short, the sample will stick to the machine used in the next part and if baked for too long the sample may become damaged. Once the sample is completely dry, it is placed into the mask aligner. Each corner is lined up using the microscope and the camera so that a complete pattern is across the sample. Once completed, the machine is set for 6 seconds and exposes the sample to UV light, the time needed varies with different materials. The sample is then washed in a beaker of developer solution for exactly one minute as again if the time is not strictly adhered to, it can have adverse effects on the sample. Finally it is washed in deionised water and dried with nitrogen gas. Now the sample has the same pattern which was imprinted from the mask aligner as the photoresist has been taken away from the parts that are needed.

Now metallisation must happen as the sample needs a metal layer placed onto it. This is completed by placing the sample in a vacuum chamber and placing a small amount of aluminium inside a tungsten coil. Tungsten is used as the process involves heating the chamber to a point where the aluminium will evaporate but the other materials will not. As the chamber heats, the aluminium creates a thin layer across all of the inside of the chamber, this layer is approximately 0.2μm which can be calculated by knowing the amount of aluminium used. As the chamber is a vacuum, there will be no chance of an oxide layer being created which would ruin the sample. Using the vacuum also means that less pressure is needed and that a lower temperature can be used.

Due to the evaporation covering the entire surface of the sample, the Lift-off process is needed to remove any parts of the metal that is not needed. The photoresist has protected some parts of the semiconductor from the metal and this needs to be removed to leave the parts where there is a direct contact between the metal and semiconductor. This is done by placing the sample in a beaker of acetone and syringing the sample until the photoresist, and the metal on it, is removed. This process is usually quite quick but can take anywhere up to 20 minutes.

The final stage before the sample is ready for experimental observation is to remove the sample from the glass slide. This is completed by heating the sample on a hot plate until the wax has melted. Once this happens the sample is then cleaned with the same three stage solvent cleaning method earlier describe to remove any remaining wax. Finally it should be dried with nitrogen gas, with care taken as the sample could be blown away due to it not being on the glass slide. Once the all stages have been completed the sample is ready for analysis at the probe station.

DEVICE CHARACTERISATION

Now the slide is ready, the first test can commence. The sample is placed in the probe station and a Source measurement unit (SMU) is connected to the probe station and a computer so the results can be recorded. As the SMU acts as both the source and meter, this is connected to one probe whilst the base is connected to the other. Now the probe is carefully aligned onto one of the small circles that has been created on the sample, with great care being taken due to the sample only having a very thin layer which is easily penetrable. Next, a voltage sweep is created from -3V to 3V with a low current limit so ensure the device is not broken. Once the data has been recorded the probe is moved onto a different circle and the process is repeated.

The behaviour of the Schottky diode can be modelled by the following equation:

If the gradient of the voltage versus Log(J) is taken, the equation can be rearranged to find n. n is the factor that is used to determine the non-ideality of the diode.

The values for the current density and n can be found in table 1.

The next experiment is to measure Capacitance verses Voltage. The measurements will be taken in a similar way to the IV measurements however, a LCR meter will be used instead. This is due to the LCR being able to measure capacitance and phase angle. For these measurements, the sweep will start at 0 and be reduced until the phase angle is roughly 75°. This is due to the need of being in reverse bias.

The following formula can be used to find the dopant density of the semiconductor.

The dopant density of the sample was found to be ***. The voltage barrier of each diode can be found by taking the gradient of the graphs.

DISCUSSION AND CONCLUSIONS

Reviewing the IV graphs it is clear that as the current increases, the linearity of the current density starts to become unstable. This can be explained due to the possibility of defects in the sample. In most industry practices, technicians are not present during the manufacturing fabrication stage and is solely completed by machines. Even though great care has been taken to ensure that the samples used in this exercise have been unaffected by unwanted particles, there is clearly still evidence of impurities throughout the sample.

When the data from the CV graphs is reviewed, the observation that as the diameter of each diode increases, the barriers height decreases can be made. Having a higher barrier height is important as this will create a rectifying contact which is beneficial as it only lets the current flow one way. Therefore the conclusion can be made that having diodes with lower diameters have more advantages when creating Schottky diodes.

This lab has demonstrated how a Schottky diode has been made and what the different characterises are. Great care has been taken to ensure little impurities affect the sample however there is still room for improvement and shows the essential need to uphold the safety requirements. Due to the low power needed and the fast switching abilities Schottky diodes are able to be used much more universally than standard PN diodes.