Quantum computer is a machine that makes use of quantum systems instead of classical transistors to perform computations. While a classical computational unit (classical bit) only contains binary information, 0 and 1, a quantum bit (qubit) can store any arbitrary state between 0 and 1.
In addition, a quantum computer allows entanglement of multiple qubits—a property that has no classical computation analog. With quantum entanglement, the information on any two qubits can be directly tied together, and any alteration to one qubit has a direct consequence to the other, without the need for a second classical control.
Finally, the quantum computation can make use of the interference among qubits to perform calculation, where the different superposition states can interact with each other in a way that is not possible classically. The three properties allow for an exponential speed up in computational power for a large class of problems and open up the door to a wide variety of quantum algorithms.
Quantum computers that are being developed today can be separated into two major categories based on the complexity and computational power.
Several technical challenges that need to be satisfied to create a working quantum computer. The widely accepted criteria for quantum computation was proposed by David P. DiVincenzo with the following requirements:
These criteria are not limited to any physical system and several quantum systems have been proposed and demonstrated, such as single atoms, single ions, photons, and artificial atoms. We list some of the examples below:
Qubits are made of the ground and excited states of ions and couplings are mediated by coulomb interactions. Since all ions are identical, the qubits are very stable. The coherence time can be several minutes, while the spin lifetime is virtually infinite. Since the stability is directly related to the isolation of the qubit from the external environment, external control is relatively slow (100 microseconds) compared to other platforms.
Qubits are made of Josephson junction devices to mimic an artificial atom and couplings are mediated by superconducting resonators. It is believed to be one of the most promising platform since the qubits are defined by nanofabrication and the techniques can be harvested from the classical semiconductor industry. Control gates are achieved via microwave pulses on the superconducting resonator within few nanoseconds. The direct trade-off from the fast control is the coherence time is still limited to below a millisecond.
Electrons and nuclear spins are trapped in quantum dots or donors and couplings are mediated by spin-orbit or exchange interactions. The host material can be chosen to have very low noise, such as isotopically enriched silicon-28. Although the fabrication of the qubits is extremely difficult, making scalability a challenge, tremendous experimental progress has been made during the past 20 years. Notable examples include interfacing spin qubits in quantum dots and donors—hot, dense, and coherent and a silicon quantum processor with robust long-distance qubit couplings.
Electron and nuclear spins associated with a defect in solid state is used as a qubit and couplings are mediated by hyperfine interactions. The system is most widely known for its long coherence time and the ability to operate at room temperature and in ambient conditions. The current limitation is scalability, since the impurities are atomic defects and deterministically placing these impurities on an atomic scale remains a challenge. Notable experimental results include quantum control over single spins in diamond and architecture for a room temperature solid-state quantum information processor.
Single optical photons are used as qubits and their couplings are based on optical interference, making interactions almost instantaneous. However, the system is still limited by the fidelity of optical control electronics and the storage and preservation of quantum information of the entangled photons. Notable experiment includes large-scale silicon quantum photonics implementing arbitrary two-qubit processing.