ABSTRACT
As we just described above, the hyperfine interaction with randomly ori-
ented nuclear spins is a fundamental decoherence mechanism for electron
spin in a semiconductor quantum dot. However, the nuclear spins them-
selves could be used to create long-lived quantum memory for quantum
bits. Nuclear spins can possess very long coherence times because of their
weak environmental coupling, but single nuclear spins are very difficult to
manipulate and measure [8]. A recent proposal by Taylor et al is moti-
vated by the possibility to use the nuclear spins as a quantum memory and
aims to combine the strengths of electron-spin (or charge) manipulation
with the long-term memory provided by the nuclear spin system [16]. The
target is to achieve a nuclear polarisation degree that is high enough to
transfer a coherent spin state from a single electron to the nuclear spin
system. This is achieved through the effective control of the spin-exchange
part of hyperfine contact interaction (via an external magnetic field pulse
for instance). After the transfer is completed, the resulting superpositions
could be stored for very long times and mapped back into the electron spin
degrees of freedom on demand. It would thus be possible to reliably map
the quantum state of a spin qubit onto long-lived collective nuclear-spin
states. This proposal is quite similar to the use of atomic ensembles as
quantum information carriers [72]. The experimental implementation of
this technique will require on the one hand, a strong strong nuclear-spin
polarization in the vicinity of confined electrons, and on the other hand
long carrier spin life and coherence times, see section 9.4. We have thus
performed a detailed investigation of the nuclear spin polarization in self-
organized InAs/GaAs quantum dots. Achieving a very large nuclear spin
polarization would (i) limit the fundamental decoherence mechanism for
electron spin in a quantum dot [28, 73] (see section 9.4) and (ii) opens the
route towards the implementation of quantum memories based on collective
nuclear spin states.