Ph.D. Thesis - Physics
Ph.D. Thesis - Physics
Ph.D. Thesis - Physics
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
problems, the others have a strong effect upon our considerations. Estimates of decoher-<br />
ence rates must be calculated for each trap design, and possible control errors must also<br />
be considered. Furthermore, although we will focus on the simulation of spin frustration<br />
as the ultimate goal, we note that the traps studied in this part may be useful for other<br />
simulations as well, such as Bose-Hubbard physics.<br />
The main considerations for a trap design are interaction rates, controls, and decoherence<br />
rates. We discuss each of these below. The design must:<br />
1. Provide a regular array of stationary qubits in at least two spatial dimensions.<br />
The array of ions desired here is an extension to two dimensions of the linear ion<br />
crystals used in most quantum information experiments. To do quantum simulations<br />
of problems that are unique to configurations of spins in two or more dimensions,<br />
such as spin frustration, we need a trap that will produce a two-dimensional array<br />
of ions. The word regular means here that the ions are stored in some configuration<br />
such that the ion-ion distance d between each pair of nearest neighbors is identical.<br />
This is important because the simulated coupling rate J depends strongly on the<br />
inter-ion distance d (varying as d −3 in the β ≪ 1 limit as discussed above). Quantum<br />
simulation is also possible in an array for which this distance is not constant, but<br />
this adds a constraint to the types of Hamiltonians that can be implemented (the<br />
J-coupling will be site-dependent). When applying a global effective Hamiltonian,<br />
this site dependence may translate into a control error, limiting the precision of the<br />
simulation, but not necessarily rendering it useless. Furthermore, such systematic<br />
control errors may be compensated if sufficient controls are available.<br />
2. Enable sufficient control over each qubit to implement the desired simulation.<br />
A trap design that fulfills the first condition does not guarantee that quantum sim-<br />
ulation may be effectively done. Sufficient controls are required to implement the<br />
desired effective Hamiltonian. These may include rotations of individual qubits, state-<br />
dependent forces arising from an optical or magnetic force, and global or ion-specific<br />
measurements. In addition, achieving some desired J-coupling rate requires a suf-<br />
ficient interaction rate between individual ions. Recall that the parameter β is the<br />
fractional transfer of the motional energy of one ion to another per secular period;<br />
although the dipolar approximation holds for β ≪ 1, the resulting J-coupling rate<br />
may prove too small to be observable within the decoherence time of the system.<br />
Therefore, in considering the controls required for quantum simulation, one must take<br />
into account the pertinent quantities β and J, and calculate each for each trap design.<br />
3. Support a low enough decoherence rate to perform meaningful simulations given cer-<br />
tain coupling rates.<br />
Internal state decoherence depends on the choice of qubit states, ambient field fluctu-<br />
ations, errors in the classical controls, and other factors that are not directly related<br />
100