Silicon Atomic Quantum Dots Enable Beyond-CMOS Electronics - Field-Coupled Nanocomputing: Paradigms, Progress, and Perspectives - page 52

contact, input/output, encapsulation and packaging issues among others which we
have charted courses for will not be discussed in this document. Likewise
approaches to eventual very rapid parallel fabrication processes are being devel-
oped but are beyond the present scope of discussion.

9 ASiQDs for Quantum Computing

In the QCA mode of operation and in analog electronic strategies the resulting
device is not quantum in that it does not depend upon quantum coherence or
superposition. However, it is anticipated that ASiQD-based qubits for eventual
quantum computing applications can be made - both charge-based and electron
spin-based qubits are possible.
The key attribute of the ASiQD-based charge qubit is that the rate of tun-
neling is very large [ 23 , 24 ]. It is estimated that the tunnel rate can be as
much as 106 times larger than the rate of decohering events. Practically this
means that there could be sucient time to undertake a coherent operation
before a disruption occurs. Though this is an attractive situation, it is also
challenging as it means that precisely phase-controlled operations on the qubit
must be done with fraction tunnel period time control which is not possible
with any conventional electronics. Some form of optical control is suggested
therefore. A new strategy for characterization of the charge qubit has been
presented [ 24 ].
The spin-type qubit would minimally use one ASiQD per qubit. Unlike in all
of the above discussion where it was presumed the isolated DB would have a 1 −

charge, the DB would be prepared in a neutral one-electron state. Application
of a magnetic field differentiates the up and down spin states to create a suitable
Zeeman-split two-level system. This is similar to the P atom dopant approach.
There, the P atom is studied at very low temperature where it does not ionize
and so is not in fact a dopant. It retains its one extra electron. The resulting
paramagnetic centre has desirable electron spin properties as proven by ESR
measurements of very large numbers of P atoms.
An architecture for spin-based quantum computing using P dopants was
first put forward by Kane [ 40 ]. In that scheme, the nuclear spins of P atoms and
the spins of bound electrons serve as qubits, which interact via hyperfine- and
exchange-interactions. In pursuit of that goal, great effort has been devoted over
the last decade to place single P dopant atoms into a silicon lattice either by a
chemical process that achieves nm position control but not perfect atom scale
control, or by ion implantation which is highly uncontrolled and not scalable
but which has led to the most impressive results so far [ 41 - 43 ]. The fabrication
challenges facing P dopant approaches to quantum computation are significant,
and it is known that quantum computation cannot work even in principle until
these are overcome [ 44 ]. A later proposal by Loss and DiVincenzo [ 45 ] makes
possible an all-electron-spin approach to quantum computation, which could do
away with the need for the nuclear spins of P donors. Paramagnetic ASiQDs are
a suitable platform for this architecture or some variant of it, and immediately