We present a study of dynamical decoupling schemes for the suppression of phase errors from various noise
environments using ions in a Penning trap as a model ensemble of qubits. By injecting frequency noise we
demonstrate that in an ohmic noise spectrum with a sharp, high-frequency cutoff the recently proposed UDD
decoupling sequence gives noise suppression superior to the traditional CPMG technique. Under only the influence
of ambient magnetic field fluctuations with a 1/ω4 power spectrum, we find little benefit from using the
UDD sequence, consistent with theoretical predictions for dynamical decoupling performance in the presence of
noise spectra with soft cutoffs. Finally, we implement an optimization algorithm using measurement feedback,
demonstrating that local optimization of dynamical decoupling can further lead to significant gains in error
suppression over known sequences.
A source of single photons allows secure quantum key distribution, in addition, to being a critical resource for linear optics quantum computing. We describe our progress on deterministically creating single photons from spontaneous parametric downconversion, an extension of the Pittman, Jacobs and Franson scheme [Phys. Rev A, v66, 042303 (2002)]. Their idea was to conditionally prepare single photons by measuring one member of a spontaneously emitted photon pair and storing the remaining conditionally prepared photon until a predetermined time, when it would be "deterministically" released from storage. Our approach attempts to improve upon this by recycling the pump pulse in order to decrease the possibility of multiple-pair generation, while maintaining a high probability of producing a single pair. Many of the challenges we discuss are central to other quantum information technologies, including the need for low-loss optical storage, switching and detection, and fast feed-forward control.
The process of up-conversion is shown to enable superior single-photon detectors in the infrared, compared to InGaAs and germanium Avalanche Photodiodes (APDs) (normally used for IR single photon detection). After up-converting an infrared photon to a visible one in a non-linear crystal-Periodically Poled Lithium Niobate (PPLN)-we use a silicon APD to efficiently detect the frequency up-shifted IR photon. We have demonstrated this process at the "high-intensity" level and at the single-photon level, where the up-converted state is effectively a superposition of the single photon Fock state and the vacuum state. We achieve an 80% conversion efficiency over the width of the pulse, and show that this process is coherent, a necessary ingredient for many applications.
A number of optical technologies remain to be developed and
optimized for various applications in quantum information processing,
especially quantum communication. We will give an overview of our
approach to some of these, including periodic heralded single-photon sources based on spontaneous parametric down-conversion, ultrabright sources of tunable entangled photons, near unit efficiency single- and multi-photon detectors based on an atomic vapor interaction, quantum state transducers based on high efficiency frequency up-conversion, and low-loss optical quantum memories.
We propose a method of single photon detection of infrared (IR) photons at potentially higher efficiencies and lower noise than allowed by traditional IR band Avalanche Photodiodes (APD). By up-converting the photon from IR, e.g., 1550 nm, to a visible wavelength in a nonlinear crystal, we can utilize the much higher efficiency of visible wavelength APDs. We have used a nonlinear crystal -- Periodically Poled Lithium Niobate (PPLN) -- and a pulsed 1064-nm Nd:YAG laser to perform the up-conversion to a 631-nm photon. When properly quasi-phase-matched, PPLN provides a large enough second order nonlinear susceptibility that near unit conversion efficiency of the IR photon into the visible should be possible. We have been able to observe peak conversion efficiencies as high as 80%, and have demonstrated scaling down to the single photon level while maintaining a background of 3 x 10-4 dark counts/count. Since the PPLN only acts on one polarization of the single photon, we also propose a 2-crystal extension of this scheme whereby orthogonal polarizations may be up-converted coherently, thereby enabling complete quantum state transduction.