QCA Devices May Take Over When CMOS Is Done

John Baliga, Associate Editor -- Semiconductor International, 10/1/1999

Researchers at Bell Labs (Murray Hill, N.J.) recently stated that oxide gate dielectrics probably will have to be replaced by 2012, and if they aren't, a new type of device would have to be found. One possible alternative to CMOS technology is quantum cellular automata (QCA) technology.

The study of cellular automata goes back to the early 1980s, and the study of quantum mechanics of course goes back much further. In recent years the two have been combined to develop computing devices (logic) with an eye towards molecular-sized cells.

A cellular automaton is a model in which a system is divided into cells. At di screte time steps, a change in a cell's state is a function of its own state and those of neighboring cells.¹ These models are used to study self-organizing behavior, which is in the same realm of study as fractals and chaos. Why would they be of any use in a device for computing definite ones and zeros?

Fig 1 Coulomb repulsion forces the electrons to opposite corners of the cell, making a bistable device.2

Fig 2  Collections of the cells form a wire (a), a fan out (b), an inverter (c) and a majority gate (d) among others.2

Instead of modeling a system with them, a system can be made to take advantage of their rules, and if the state changing rules are quantum mechanical, there is potential to scale the cells to near atomic size. In fact, building a system with near-atomic-size cells probably will require self-assembly and familiarity with cellular automata modeling.

Researchers at Notre Dame University, headed by professor Craig Lent, have been studying QCA devices made of aluminum quantum dots with aluminum oxide (AlO_x) tunneling barriers. Though they are not using semiconductors in their work, their results can apply to any system that involves quantum dots, such as arsenic clusters embedded in gallium arsenide.

The basic cell structure is shown in Fig. 1. It contains four quantum dots with four tunneling junctions around the edge. When two excess electrons are added, coulomb forces keep them on opposite corners of the cell, making two possible states for the cell.

Fig. 2 shows the building blocks for constructing logic devices out of these cells. When these cells are placed in a line, they form a wire (2a). A state change at the input cell propagates down the wire. When a 90° turn is placed in the line, it still acts as a wire, and a T-shape functions as a fan out (2b). Diagonal placement (45°) makes an inverter (2c). A three-input device called a majority gate, where the output takes the state of the majority of the inputs, is easily realized using these cells (2d).

One very compelling reason for studying QCAs is that state changes can be made and propagated without passing current. The electrons stay in their cells. The challenge that goes with this is that the output state must be measured without altering that state of the output cell, and the Notre Dame group has developed a method for doing it.

More information on QCAs is available at the group's QCA Webpage, at http://www.nd.edu/~qcahome/. Included are Java applets that demonstrate currentless state change propagation with these cells.   

References

1. S. Wolfram, 'Cellular Automata and Complexity: Collected Papers,' Addison-Wesley Publishing Co., 1994.
2. G.L. Snider et al, 'Quantum-dot cellular automata: Review and recent experiments,' J. Appl. Phys. 85, 4283 (1999).