Compilation of selective pulse network on liquid-state nuclear magnetic resonance system

In creating a large-scale quantum information processor, the ability to construct control pulses for implementing an arbitrary quantum circuit in a scalable manner is an important requirement. For liquid-state nuclear magnetic resonance (NMR) quantum computing, a circuit is generally realized through a sequence of selective soft pulses, in which various control imperfections exist and are to be corrected. In this work, we present a comprehensive analysis of the errors arisen in a selective pulse network by using the zeroth and first order average Hamiltonian theory. Effective correction rules are derived for adjusting important pulse parameters such as irradiation frequencies, rotational angles and transmission phases of the selective pulses to increase the control fidelity. Simulations show that applying our compilation procedure for a given circuit is efficient and can greatly reduce the error accumulation.