Undergraduate Research in Engineering at Rice
Christoph Winkler, Senior Materials Science and Engineering student
Sample Production and Optimization of the Sputtering System (Fall 2001)
with Dr. Suzanne Stemmer
With construction of the sputtering system completed, emphasis was placed on both the production of samples and optimization of key properties. Initially, I helped familiarize Dr. Stemmer’s new graduate student with the operation of the sputtering system. After this short introduction period, we began to produce strontium titanate (STO) thin films on silicon wafers as planned. STO is a ceramic oxide thin film with a very high dielectric constant—among other sought after properties—that may replace silicon dioxide (SiO2) as the gate oxide of choice in metal-oxide semiconductors (MOSFETs). The importance of replacing silicon dioxide looms larger everyday; as engineers continue to shrink the size of microprocessors and other electronics built upon MOSFETs, several fundamental limitations become increasingly detrimental to the performance of such chips. One such limitation is that as SiO2, when acting as a gate oxide, becomes increasingly thin, the amount of current leaking through the material increases dramatically. Such leakage currents are naturally undesirable because they produce unwanted heat and increase the power requirements of chips. The higher dielectric constant of STO would allow for a thicker gate oxide which would decrease the amount of leakage current, and hence reduce power requirements and waste heat.
In this particular project, we are interested in using STO as a gate dielectric for MOSFETs using organic semiconductors, in collaboration with Prof. Natelson and his students in the Physics Department. For this application, some of compatibility problems STO shows with Si (low barrier height and chemical instability) are not relevant.
The goal of our research was to optimize the structure and chemical composition of STO thin films. As mentioned, we started by sputtering onto silicon wafers which are cheap and easily acquired. However, silicon wafers are always covered with an amorphous layer of silicon oxide, due to oxidation when exposed to atmosphere. By sputtering on top of this SiO2 layer, we limited our ability to analyze the structure of the STO film because the amorphous oxide layer disturbed expitaxial growth. Thus, we first focused on optimizing the sputtering parameters to produce stoichiometric films, or films that have the same chemical composition as the STO target. There is little equipment at Rice which allows for accurate quantitative chemical analyses for thin films; we attempted to use the energy dispersive x-ray fluorescence spectrometer (EDX), but accurate results require thin film samples of known composition which were not available to us. To obtain accurate results, we moved to wavelength dispersive spectroscopy (WDS) which is similar to EDX, but slower and much more accurate. An added bonus is that the WDS system we used at the University of Houston was staffed by a knowledgeable team that assisted and trained. Using WDS, we were able to optimize the sputtering parameters to produce stoichiometric films. Additionally, we were able to analyze how a change in the sputtering power affected the chemical composition of the thin film, for example.
We constructed a tube furnace for annealing experiments. After sputtering the STO films, we put them in a high temperature (1200oC) furnace and pass oxygen through. This allows for the film to ‘reorder’ itself, reducing oxygen vacancies, stresses, and other defects common to oxide thin films. Using WDS we could analyze how the annealing process changed the chemical composition of the films.
Another powerful technique I was fortunate to learn was atomic force microscopy (AFM). This technique allows for very high resolution surface imaging, which indicates how smooth or rough our films are. AFM can also illustrate grain formation which indicates a relative level of crystallinity in the films. We used AFM more for analyzing other sputtered films in projects that I am not directly involved in, but was fortunate enough to have been taught the technique.
We are still involved in cooperation with Dr. Natelson’s group. Instead of supplying them with alumina thin films, they decided to switch to STO films due to the higher dielectric constant. They build organic transistor devices on top of our STO films and measure transistor curves such as those shown in figure 1; though note that this is for an alumina film. Unfortunately, the devices built upon STO have proved to be too ‘leaky’ for practical purposes. One possible indication of this is that the quality of the films is poor; there is the possibility that our films are full of defects such as oxygen vacancies which would change the STO from an insulating dielectric into a conductor, resulting in extremely leaky films. We are in the process of tweaking the sputtering and annealing processes to produce higher quality films.
Figure 1 Transistor curves for an organic transistor built upon 2000Å of Al2O3. Though not visible, there is very little leakage current, indicating a high quality sputtered film.
Further plans include the continual optimization of STO thin film properties, with an emphasis on films that are less leaky and highly structured. Additionally, I will be working on a project on the effect of titanium ‘doped’ STO thin films for which I was awarded a $1000 research grant by the Materials Research Society.
Department of Mechanical Engineering and Materials Science
Back to Undergraduate Research page
|
About the School | Academics & Research | Departments | What's New Alumni | Calendar | Contacts | SE Web | Duncan Hall Conference Room Reservations |
 
|
||
