Quantum Memory for Photonic Qubits
This invention, in one aspect, is a method for performing quantum logic operations using linear optical elements with much smaller errors than in earlier methods. As shown in Fig. 1, the proposed quantum logic devices have two logical inputs, each of which is a quantum bit (qubit) represented by a single photon. The value of a qubit is determined by which of two possible paths each photon occupies. The logic devices also have two output qubits, represented in the same way by two photons. In addition to the two inputs and two outputs, N additional photons, referred to as ancilla, are also input to the device; the ancilla carry no information and are destroyed in the operation of the device. A key feature of the invention is that the ancilla photons are generated in an entangled or correlated state chosen to minimize the error rate in the subsequent operation of the logic device. The logic operation is performed by quantum teleportation of the input qubits to form the output qubits. The main advantage of this procedure is that the average error rate is inversely proportional to the square of N. Thus, when the number of ancilla is large (as it must be), the invention gives a much lower error rate than in previous methods thereby making it possible to achieve the low error threshold required for the operation of a quantum computer.
In classical (conventional) computing, the bits of information have definite values of “0” or “1” and can be easily stored in a variety of computer memory devices. In quantum computing, however, the quantum bits (called “qubits”) can exist in complex superposition states of 0 and 1, and the requirements for storing them in a “quantum memory” are much more stringent. The invention, in a second aspect, as shown in Fig. 2, is a method for implementing such a quantum memory device for polarization-encoded photonic qubits that utilizes a simple free-space optical storage loop, and the coherence of the photonic qubits is maintained during switching operations by using a high-speed electro-optic polarizing Sagnac interferometer switch. The photons stored in the loop can be switched out after any number of round trips, providing a cyclical quantum memory (CQM) that can be synchronized with the cycle time of an optical quantum computer. The switching techniques of the invention allow the use of polarization-encoded qubits rather than path-encoded qubits providing for more robust and desirable quantum operations. There are currently a number of proposals for providing a single-photon from the spontaneous emission of an isolated two-state quantum system. In principle, these approaches offer the possibility of a single-photon “on demand”, but, while the probability of single photon emission can be high in these approaches, there is no method for ensuring that a photon has actually been emitted. In contrast, a third aspect of the invention is based on the production of pairs of photons through low-power pulsed-pump parametric down-conversion, and the detection of one of the photons of a pair can signal the presence of the other photon with near certainty along with its direction and optical mode. As shown in Fig. 3, a photon pair is emitted from a pulsed parametric down-conversion source. Classical information describing the detection of one of the photons is used to activate a high speed electro-optic switch that re-routes the other photon into a storage loop. The stored photon is then known to be circulating in the loop, and can be switched out at a later time chosen by the user, providing a single photon for potential use in a variety of quantum information processing applications. Although the stored single photon is only available at periodic time intervals, those times can be chosen to match the cycle time of an optical quantum computer and, therefore, a pseudo-demand source of this kind is just as effective as a source in which the single photon can be produced on-demand at arbitrary times.
Patent Status: U.S. patent(s) 7019875; 7355769 issued.CONTACT: