In fact, a real qubit decays and that decay limits the qubits. A real qubit is never perfectly isolated from thermal and phase noise environment and also has a limited phase coherence lifetime on the order of 10 microseconds. There are therefore many hurdles to overcome before any practical qubits of quantum computing become a reality. Much like the early days of the practical 0 and 1 bits of semiconductor logic, Science has a long ways to go in order to realize a useful practical qubit that includes not only 0 and 1, but also quantum phase, theta.

The superconducting Josephson junction is a fundamental quantum oscillator that involves electron (Cooper) pairs tunneling through an insulator layer between two superconductors at very low temperature. Instead of the electrons and holes that determine semiconductor 0 and 1 bits, a Cooper pair is inherently a qubit. For a current of 40 nA, about 1e9 electron pairs result and from a 13 microV, a frequency of 6.6 GHz at 0.015 K. The Cooper pair current results from the specific geometry and materials of the junction as well as the applied voltage but the frequency is always just proportional to the applied voltage. In fact, this junction is a quantum oscillator at that frequency where each excited state includes one additional Cooper pair of electrons at a slightly lower frequency due to anharmonicity.

The basic qubit of a quantum computer incorporates not only the 0 and 1 of a classical bit, but also a quantum oscillation between 0 and 1 of the Cooper pair across a junction. A very common qubit is a particular Josephson junction called a transmon that incorporates a shunt capacitor to make the quantum oscillator more stable. The transmon that oscillates at around 6.6 GHz and so its qubits undergo this same quantum oscillation.

Another common qubit is the squid, which involves a loop with two Josephson junction. In any case, a qubit is the excitation of just one Cooper pair, 0 -> 1, across a junction at about 200 MHz lower frequency due to anharmonicity. The quantum anharmonicity also means that the 1 -> 2 transition is 200 MHz less that then 0 -> 1 transition. In fact, useable qubits need to have such isolated transitions and so the anharmonicity is what makes the transmon and the squid useful qubits as the figure shows.

However, there is an additional splitting of each level due to the phase or direction of the electron pair across the junction and that splitting reflects the spin or rotation of the qubit as the figure below shows. Much like electron spin emerges from the complementary rotations of electron charge loop oscillation, the complementary rotations of superconducting loop oscillations in the transmon and squid are then a kind of qubit spin.

*n*. Gate bias increases charge dispersion up to 60 MHz as the figure shows.

_{g}*T*, is therefore usually about 10 microsec, which is long enough for reading and still short enough for resetting, which all involve 10 nsec switches. The dephasing time,

_{1}*T*, is due to the entanglement among other qubit states that is necessary for effective computation. This dephasing time is important for quantum entanglement outcomes and is therefore limited by

_{2}*T*. However, it is then difficult to differentiate dephasing from pure decay.

_{1}