Physics
Laboratory

Information
Technology
Laboratory

Electronics and
Electrical
Engineering
Laboratory

Project
Home Page

Objective

The objective of this project is to develop an extensible quantum information test-bed and the scalable component technology essential to the practical realization of a quantum communication network. The test-bed will demonstrate quantum communication and quantum cryptographic key distribution with high data rate. This test-bed, once developed, will provide a measurement and standards infrastructure that will be open to the DARPA QuIST community and will enable wide-ranging experiments on both the physical- and network-layer aspects of a quantum communication system. The infrastructure will be used to provide calibration, testing, and development facilities for the QuIST community.

This project is one part of the broader Quantum Information Program at the National Institute of Standards and Technology (NIST). The DARPA QuIST effort focuses on:

• the quantum information test-bed,
• the development of a single photons and entangled photons on demand,
• the development of a single photon detector for calibration purposes,
• the development of metrics and calibration capabilities for both hardware and software,
• the development of hybrid quantum-classical authentication protocols,
• the development of high-level models of a quantum processor to optimally implement error correction and avoidance and to provide quantum architectures that will allow for structural scalability beyond the physical layer.

Optical communications equipment for the NIST QKD system. The telescope for the classical channel is operational at the top of the picture. The telescopes for the quantum channel are being assembled at the lower left.

Approach

The quantum information test-bed has been designed to provide an upgradeable, high-speed (~1 Gigabit) network for Quantum Key Distribution and Quantum Teleportation. Initially it will consists of two free space nodes using attenuated lasers implementing the BB84 quantum key protocol. Unlike existing experimental quantum communication links, our quantum communications channel and a free space classical channel are based on adopting existing high-speed optical wireless technology. This approach will allow us to significantly improve the rate of quantum key distribution as on demand single photon sources are developed and as the efficiency and repetition rate of single photon detectors improves. The software for quantum communication, which always requires a classical communication channel, is being designed to interface with exiting classical communication network protocols including TCP/IP and IPSEC. Although the software is initially being designed at a high layer it will eventually be pulled down into the low-level communication protocols and NIST will work with others to develop a general security standard for quantum keys. A hybrid approach to authentication will be described below.

The single photon on demand source will be based on parametric down-converters and single photon detectors to herald single photon states on demand while controlling the probability of having more than one photon. This method can also be used to produce entangled photon pairs on demand, and by controlling the input pump level and the number of parametric down-converters, one can independently design a system that has as many photons in a single photon pulse train as desired. A separate internally funded NIST effort has recently demonstrated a single photon turnstile based on quantum dots and future extensions of this work are underway.

The approach to single photon detectors is based on superconducting transition edge sensors. By using anti-reflection coatings, these detectors have shown efficiencies of ~ 90% at 1550 nm with a 0.15eV FWHM energy resolution, a count rate greater than 10 kHz, and the ability to distinguish individual photons. We believe that with special coatings and a better fiber interface can ultimately improve efficiency to 99%. These detectors can clearly distinguish between one and two photons and do not have a dark count problem. Initially we will use a parametric down-converter to measure the absolute quantum efficiency of this detector. Experiments with the NIST developed single photon turnstile are planned for FY04. Ultimately we will hope to use this detector to test and calibrate single photon sources developed either internally at NIST or by the QuIST community.

Controlling access to large network resources is one of the most common security problems. Any pair of parties in a network should be able to communicate, but must be authorized to do so. The fundamental problem is how to authenticate resources to each other while minimizing the number of cryptographic keys that must be distributed and maintained, given the potential for pairs of communicating resources. This project will develop an authentication solution based on a combination of quantum cryptography and a conventional secret key system. Two significant advantages of this approach over conventional authentication protocols are 1) timestamps and exact clock synchronization between parties are not needed; and 2) that even the trusted server cannot know the contents of the authentication ticket.

Our approach to developing an optimal approach to quantum error correction and error avoidance will be based on physical layer models, gate level models, and the development of quantum simulators. No practical quantum computation system can be developed unless the problem of decoherence is addressed in the most basic system design. The models being developed via collaboration of computer scientists and physicists are being designed to achieve fault tolerance in components demonstrated by the NIST Physics Laboratory. Solutions may be based on optimal use of decoherence-free subspaces and quantum error correcting codes and will be applied to specific configurations suggested by specific hardware components. To the extent that modified configurations can lead to significant improvement in performance, the models may lead to the design of new trap/qubit configurations. NIST is also pursuing broader architectural concepts to provide scalability beyond the physical layer. These concepts are aimed at insuring efficient gate operations between any two arbitrary qubits and at providing sufficient internal parallelism to guarantee scalability at the logical qubit layer. Such models are needed to calculate actual error thresholds for specific quantum architectures.

View of the two endpoints of the QKD free space optical communications link at the NIST site in Gaithersburg, Maryland.

Recent Accomplishments

The design, development, and beta testing of the quantum information test-bed is now complete and we now routinely produce 1.0 megabits per second (Mbps) of raw sifted key per second with a sifted key bit error rates of around 1.1% for our 730 m free space system. See paper in Optics Express. Currently our reconciliation and privacy amplification implementation currently limits our overall throughput to no more than 0.14 Mbps of cryptographically secure key as result of systems overheads in the reconciliation and privacy amplification implementation leading to buffer overflows and system timeouts.

Some of our specific accomplishments include the following.

• Classical and quantum physical link layers have been fully integrated and we have observed the generation of sifted quantum key at rates in excess of 3.5 Mbps over the 730 m free-space link. This is the highest sifted-key rate ever achieved by a single-photon QKD system and marks a significant advance in the physical layer implementation of a QKD system.
• Attenuated laser sources and drivers were redesigned to achieve higher pulse compression (800 ps data pulse to 200 ps optical pulse) and higher extinction ratio (greater than 200:1). This enables higher quantum channel transmission rates at reduced error rate.
• Single-photon detector timing resolution has been optimized for use with our high transmission rates on the quantum channel. However the detector timing resolution, or jitter, currently limits us to spacing our quantum channel transmissions by 4 clock cycles thereby reducing our maximum rep-rate on the quantum channel from 1.25 GHz to 312.5 MHz. Techniques to achieve higher detector timing resolution have been identified.
• Transmit board (Alice) and receive board (Bob) are fully operational under Linux with direct memory access.
• The maximum rate at which the boards can sift the quantum channel data was increased by a factor of 30 by implementing all sifting functions in the FPGAs, with communications taking place over the primary classical channel (the timing channel). The boards can now sift data generated by the quantum channel at rates up to 32 Mbps.
• Initial integration of our hybrid CASCADE/Forward Error Correction (FEC) reconciliation algorithm with the fully integrated classical and quantum channels has been achieved. See preprint.
• We have produced error-corrected and privacy amplified key at rates up to 140 kbps over the free-space link. The maximum expected rate of cryptographically secure bits is roughly 2 Mbps for a 1.1% sifted key bit error rates. This reduction from the 3.5 Mbps of raw sifted key is a result of operating system overhead in the reconciliation and privacy amplification implementation that causes our continuous system to drop over 90% of the raw sifted key packets to prevent buffer overflows and system timeouts.
• Modifications to our implementation of our hybrid CASCADE/FEC reconciliation algorithm are underway and have already increased the maximum throughput to 300 kbps. We have identified a variety of improvements, and expect more gains in the rate of reconciliation as work proceeds. Ultimate resolution of this may require the design of a second high-speed dedicated board.
• A high-level application demonstrating secure one-time pad video transmission has been developed and tested. This application will be fully implemented on the system when the reconciliation and privacy amplification issues have been addressed.

An integrated PDC source running with output photons near 1550 nm with fiber coupling and integrated fiber optical switching has been designed and the components have been procured and most have now been delivered. The high speed switching circuit to control this system is designed and completion is expected shortly.

Single photon detection remains a limiting factor for a PDC based single photon source, for quantum communication, and for single photon quantum computing schemes. Papers resulting from a workshop held in the spring of 2003 have just appeared in a special issue of the Journal of Modern Optics.

The final area of our program is in the quantum architectures and quantum protocols area. The accomplishments here are:

• We have developed a new universal quantum circuit capable of implementing any unitary operator. It has a top-down structure which concentrates the circuit on the less significant qubits, and the parameters for a given unitary may be computed using standard matrix analysis software. The number of controlled-nots is only half the number of matrix coefficients. Moreover, a theoretical lower bound shows that this new universal circuit may be improved by at most a factor of two. The circuit adapts well to architectures in which only nearest-neighbor interactions are possible, i.e. spin chains. See quant-ph/0406176
• A generalization to Deutsch's Algorithm detecting concentrated maps has been developed. See preprint.
• A constrained QR factorization enabling quantum circuit design has been developed. Any matrix of dimension m by n can be reduced to upper triangular form by multiplying by a sequence of mn-n(n+1)/2 appropriately chosen rotation matrices. In this work we address the question of whether such a factorization exists when the set of allowed rotation planes is restricted. Applications to the design of 3-qubit circuits as well as to the design of quantum circuits for qudits have been studied. See preprint.
• The mathematical foundations for studying one-type of multi-qubit entanglement have been established. We investigated a function on states of many qubits, the n-concurrence, which satisfies the necessary and sufficient properties of an entanglement measure including invariance under local unitary operations. The n-concurrence is a measure of the overlap of a state with its time reversed counterpart and as such is maximal for GHZ type states. By decomposing unitaries on n qubits using a singular value decomposition with respect to time reversal symmetry, we were able to quantify how entangling are unitary dynamics on qubits. A significant result is that almost all unitaries on a large number of n (even) qubits can produce states with maximal n-concurrence. See paper in Journal of Mathematical Physics. A second important result is a theorem stating that the ground state of a time reversal symmetric Hamiltonian have maximal n-concurrence if they are non-degenerate.
• We have developed new models for scalable quantum architectures that may be appropriate to neutral atom quantum computing. See paper in Physical Review A. Development and refinement of these ideas is ongoing.

Current Plan

Optimize, analyze, and document the NIST QKD system. Collect data on bit error rates and total key generated on the NIST QKD system under various weather conditions and optimize the physical layer performance and throughput. Modify system as required for improved performance and obtain, if possible, full 1.25GHz quantum transmission rate by removing additional physical limitations.

Optimize device drivers, sifting procedures, and the reconciliation algorithm for NIST based system. Port the Java based reconciliation and privacy amplification procedure to C/C++. Determine whether a hardware based implementation of the reconciliation and privacy amplification implementation is required to prevent bottle-necks.

Complete development and test a switched telecom single photon source based on parametric down conversion (PDC) including fiber coupling. Assemble the system and test in the lab.

Begin development of other conceptual and physical architectural approaches to scalable quantum computation that use resources and relax constraints on physical systems. Initiate a study to determine device independent benchmarks that can quantitatively evaluate and compare different physical systems. Apply these schemes to NIST based quantum computation schemes to determine improved error strategies. Refine the error threshold criteria at a physical level against numerous resource tradeoffs appropriate for to the physical system.

Sponsors

This work is sponsored by the by the DARPA Quantum Information Science and Technology (QuIST) program. Additional funding has been provided by the NIST Physics, Information Technology, and Electronic and Electrical Engineering Laboratories.

This page's URL is http://math.nist.gov/quantum/overview.html.
Last update : 16 July 2004 by RFB.