DEVELOPMENT OF ANIMATED SIMULATION OF SEMICONDUCTOR ELECTRONIC DEVICES FOR CLASSROOM DEMONSTRATION

Constance J. Chang-Hasnain, James D. Plummer, Robert W. Dutton, and Zhiping Yu*
Electrical Engineering Department
Stanford University
Stanford, CA 94305-4085
Tel: (415) 723-0111; FAX (415) 725-2533

*on leave from Tsinghua University, Beijing, China

ABSTRACT

In this paper, we describe the development of animation videos based on computer simulation of semiconductor electronic devices for classroom demonstration of abstract concepts. The device physics taught in junior-level required courses in electrical engineering are considered and include: p-n diode, metal oxide semiconductor (MOS) capacitor, MOS field effect transistor (FET), and bipolar junction transistor (BJT). The concepts and methodology will be described, followed by several examples.

INTRODUCTION

Since the first demonstration of diodes and transistors, semiconductor electronic devices have continued to revolutionize our society through an enormous number of inventions and applications built upon them. Therefore, an introductory course on the basic concepts of semiconductor electronic devices is generally considered essential for all electrical engineering students, regardless of their final field of specialization. Nevertheless, the fundamental physics of these devices is of an abstract nature. Students with basic sophomore physics training often have difficulty imagining electrons and holes in semiconductors, let alone a p-n junction with electron and hole current flows. Although they have usually, by this time, been introduced to the fundamentals of quantum mechanics and of materials, a semiconductor crystal with conduction and valence bands is very difficult to grasp and visualize. It is based on this experience that the authors began the development of animation video demonstrations of basic electronic devices. The objective is to use animated simulation to provide a visual-aid for understanding complicated concepts and the underlying formula. In addition, the animations give students a clear grasp of many physical parameters and their relative importance. Furthermore, such animations will stimulate both interest for learning and deeper understanding.

METHODOLOGY

The methodology of creating the animated simulation movies is outlined in Figure 1. The devices of interest in this project include p-n junction diode, p-i-n junction diode, MOS capacitor, MOSFET and BJT. These devices were first simulated using typical device parameters at various operating biases. The device simulation program we used was TMA-MEDICI*, a commercial version of PISCES (Dutton and Yu 1993), on a cluster of UNIXTM-based DEC workstations. With each operating condition, a graph of data is generated such as current vs. voltage (I-V) and capacitance vs. voltage (C-V) characteristic curves, carrier and space charge distributions, electron and hole current distributions, electric field distributions, and energy band diagrams. For all the devices, the graphs can be made into simple two-dimensional (2D) plots, 2D contour vector plots, or three-dimensional (3D) plots. These graphs are captured and transferred onto an Apple workstation using a graphics terminal emulator (for example VersaTerm). A graphics application (for example Canvas or MacDraw) is used to add artistic perception of electrons, holes, and currents onto the band diagrams, and to touch up the graphs. The animation is then done using MacroMind Director, an animation package for Apple workstations. This program allows us to combine various objects called cast members into a complete screen frame. The cast members can be a PICT file, a combination of PICT files, or a sound source. In our case, the frames can be calculated plots generated by MEDICI for a specific device at various operation conditions, e.g. biases. A motion picture of this device as one changes the bias can thus be made by putting the frames in sequence. In many cases, electrons and holes with appropriate motions are included to exhibit current flows. Sound effects are also used to enhance the demonstrations.

Fig. 1. Steps in creating animation of semiconductor electronic devices. This method offers several advantages.

First, with the use of MacroMind Director, movies in a variety of formats, i.e. multimedia, can be generated. They can be exported to a video tape for classroom demonstrations, to a Quicktime** movie to be played on Apple computers, to presentation view graphs, or online movies. Secondly, there is a great flexibility in devices to be simulated and even simulation programs to be used. For example, simulators such as B2-SPICE+, PSPICE, etc. can be used to generate circuit simulation graphs, which can also be made into a movie. Finally, future improvements can be easily made since the production is independent on any particular piece of software or program.

EXAMPLES

Four sets of simulated results are shown here as examples. The first device is an ideal Si p-n junction diode with uniform doping densities of 1x1016 and 1x1015 cm-3 on the p and n sides, respectively. The energy band diagrams for three bias values are shown in Fig. 2. These graphs and many more with in-between bias values are then made into frames to compose a movie. The students can observe the effect of the bias on the built-in potential as well as the depletion region width. The electron and hole concentration distributions for various biases are shown in Fig. 3. In this case, each curve is used as one frame of a sequence. From this set of curves, it is easy to see when the applied bias is higher than 0.5 V, the diode is in the "high-level injection" regime and the approximations used for the "low-level injection" in textbooks are no longer valid. Also, this calculation was done for the "short base" case and the carrier densities drop quickly to zero at the contacts. Different movies can also be generated for long and medium base cases for comparison.

The second example is a MOS capacitor. The input file to MEDICI is shown in Fig. 4. The C-V plots at low and high frequencies are shown in Fig. 5, which are part of a short animation with frequency being the operating variable. The electron concentration as a function of distance away from the oxide layer is shown in Fig. 6 for various biases. Again in this case, each curve forms a frame.

Fig. 2 The energy band diagrams for an ideal Si p-n junction diode at three bias values. The doping densities for the p and n sides are Na = 1 x 1016 cm-3 and Nd = 5 x 1015 cm-3, respectively.

Fig. 3 Electron and hole concentration distributions for the same diode at various biases. The doping densities for the p and n sides are Na = 5 x 1015 cm-3 and Nd = 5 x 1015cm-3, respectively.

Fig. 4 The input file to MEDICI for a typical MOS capacitor.

Fig. 5 The C-V plots at low and high frequencies for the same MOS capacitor.

Fig. 6 Electron concentration of the MOS capacitor as a function of distance away from the oxide layer at various bias voltages.

The n-channel MOSFET simulation is shown in Fig. 7 and 8. Fig. 7 shows the current contour plots for a fixed gate voltage (Vg = 2 V) and various source-drain voltages Vd. The electron concentration contour plots are exhibited in Fig. 8 at a fixed Vd and various Vg. The 2D contour and 3D representations provide excellent illustrations of what actually takes place in a 3-terminal device. Without a movie of such graphs, it takes tremendous effort for the lecturers to explain the device behavior properly.

Fig. 7 Current contour plots for an n-channel MOSFET at a fixed gate voltage (Vg = 2 V) and various source-drain voltages Vd.

Fig. 8 Electron concentration contour plots for the same MOSFET at a fixed Vg = 2V and various Vd.

Fig. 9 Energy band diagram of an n-p-n transistor at various collector voltages Vc and a fixed base voltage Vb.

The final example is an n-p-n transistor. The energy band diagram of the transistor at various collector voltage Vc and a fixed base voltage Vb is shown in Fig. 9., and vice versa in Fig. 10. The carriers can be added on to diagrams like these to provide motion and illustration of current flows.

Fig. 10 Energy band diagram of the same transistor at various base voltages Vb and a fixed base voltage Vc.

EXPECTED IMPACT AND LIMITATIONS

The use of animated simulations is expected to have a great impact in classroom teaching of abstract and mathematical materials. It should help to make abstract ideas more concrete and physical. In addition, the movies will help students to visualize time variations and spatial distributions of various device parameters. In the next phase, we will start using the animation as a teaching tool in the junior-level electronic devices courses. We envision a large number of students, not only those who are introduced to semiconductor electronics for the first time but also graduate students and practicing engineers in the field, will benefit from the animation videos once they are more widely used in other courses.

One major advantage of developing the animated movies on an Apple workstation is that Quicktime movies can be made available to students on floppy disks. With the high accessibility to Apple computers on campus and in the dorms, the students can play the movies at their pace conveniently and repeatedly.

A main limitation of our methodology at present is the fact that the animation cannot be fully interactive since the graphs are pre-generated with a set of pre-determined device parameters and operating conditions. This limitation is mainly dictated by the high level of computation power required by PISCES. The device simulation thus has to be done on a more powerful computer workstation. On the other hand, the animation application is far more flexible and functional with Apple computers. To overcome this limitation, we plan to make many example movies for a variety of parameters and operating conditions so that the students can attain a similar effect as interactive animation.

ACKNOWLEDGMENT

We acknowledge the diligent and enthusiastic work of Michael Kwong, Danny Lee, and Keith Toh, undergraduate students in the Electrical Engineering Department at Stanford University, who have spent their entire summer working on all the details of this project. We express our thanks to members of the NSF-supported National Center for Computational Electronics (NCCE) for their help on device simulations. Finally, we thank the generous support of the EE Department Chairs, Professors Joe Goodman and Gene Franklin, on this project.

REFERENCES

Dutton, R. W. and Z. Yu. 1993. Technology CAD - Computer Simulation of IC Processes and Devices. Kluwer Academic Publishers, Norwell, MA.

BIOGRAPHY

Constance Chang-Hasnain received her PhD degree in electrical engineering from the University of California at Berkeley in 1987. From 1987 to April 1992, she was a Member of Technical Staff at Bellcore, Red Bank, NJ. She is now an Assistant Professor and Reid and Polly Anderson Faculty Chair in the Electrical Engineering Department at Stanford University. Her current research include vertical cavity surface emitting lasers, semiconductor ring lasers and Al-free InGaAs/InGaP lasers. She has published over 45 technical journal articles, been awarded two patents, and has contributed one book chapter.

Prof. Chang-Hasnain was awarded with the Sakrison Memorial Prize for the most outstanding doctoral dissertation from the EECS Department at UC Berkeley. She was named the 1991 Outstanding Young Electrical Engineer by Eta Kappa Nu. She was awarded with the 1992 National Young Investigator Award from the National Science Foundation, the 1992 Packard Fellowship from the David and Lucile Packard Foundation, and the 1993 Young Alumnus of the Year Award from UC Davis. She is a senior member of IEEE and an elected member of Board of Governors of the IEEE Lasers and Electro-Optics Society.

* MEDICI is a two-dimensional semiconductor device simulator developed by Technology Modeling Associates, Inc.

** Apple's standard multimedia file format for storing sound and video images. MacroMind uses its own format to save animations, but it can be saved into Quicktime format.

+ An analog circuit design and simulation software developed by Beige Bag Software.