5.2.8 Ring Memory

   We are proposing an idea which is a generalization of the bistable quantum cell for cellular automata by C. Lent et al. [75]. On the circuit level it is also similar to the MTJ memory, because it is a multiple tunnel junction connected to a ring, which is why we call it a ring memory cell (see Fig. 5.10).

  

Figure 5.10: Circuit diagram of the ring memory cell.

An even number n (in our case n=6) of tunnel junctions are connected to a ring, and n/2 electrons are inserted into the ring. These electrons will repel each other and can form two stable configurations (see Fig. 5.10 left side). Applying voltage pulses on $V_{\text{in,set}}$ or $V_{\text{in,reset}}$ will switch the state of the ring to either one of the stable configurations. The capacitors C₀ should be small compared to the capacitances of the tunnel junctions, so that the electrons have a large influence on their neighbors and keep their distance.

A more complex memory cell, but with considerably improved characteristics is a combination of a ring memory and a MTJ memory, as shown in Fig. 5.11.

  

Figure 5.11: Circuit diagram of a ring trap memory combination.

The idea is to load and empty a ring memory over a MTJ. Further the two stable states of the ring memory are not used to distinguish between logic zero and one, but to move trapped charges around by applying a sequence of pulses to $V_{\text{read}}$, which produce Coulomb oscillations in an adjacent SET transistor. If no electrons are trapped in the ring no current oscillations will be induced in the SET transistor. This memory cell is random background charge independent, since only the presence and absence of Coulomb oscillations distinguishes between logic zero and one. The basic idea of using Coulomb oscillations to achieve random background charge independence was taken from the memory explained next in Section 5.2.9. However our ring-trap memory has a non-destructive read cycle, which comes at the cost of a higher circuit complexity.