A class of two-qubit states called X-states are increasingly being used to discuss entanglement and other quantum correlations in the field of quantum information. Maximally entangled Bell states and "Werner" states are subsets of them. Apart from being so named because their density matrix looks like the letter X, there is not as yet any characterization of them. The su(2) X su(2) X u(1) subalgebra of the full su(4) algebra of two qubits is pointed out as the underlying invariance of this class of states. X-states are a seven-parameter family associated with this subalgebra of seven operators. This recognition provides a route to preparing such states and also a convenient algebraic procedure for analytically calculating their properties. At the same time, it points to other groups of seven-parameter states that, while not at first sight appearing similar, are also invariant under the same subalgebra. And it opens the way to analyzing invariant states of other subalgebras in bipartite systems.
Main idea: Classifying “X-states” as the subalgebra \(su(2) \times su(2) \times u(1)\) of \(su(4)\)
An “X-state” is a \(4\times 4\) Hermitian matrix that’s zero except on both diagonals. Described by 7 real parameters. Bell state included.
Can be represented by elements on the Fano plane (triangle with circle inside). It has one fully commuting element, and each of other 6 is part of two sets of anti-commuting elements, plus 1 “conjugate” (commuting) element.
This representation makes it easier to do calculations on X-states (for things like “quantum discord” and “concurrence”)
Making an X-state: \(\rho + X_1 \rho X_1^\dagger\) where \(X_1\) is the fully commuting element.
(Each of 15 operators in \(su(4)\) commutes with 6 and anticommutes with the other 8)
There are 15 possible \(su(2) \times su(2) \times u(1)\) subalgebras. Some represent different things.
Can be viewed as a stabilizer
If your H is an X-state, and you’re in an X-state, you’ll stay in an X-state!
If you sandwich with another Pauli operator, It may be become an X-state in a new representation
Other thoughts:
Can be viewed as two independent qubits plus a phase shift \(u(1)\)
How is this related to two Bloch spheres?
Can model dissipation as an X-state
What do the non-equivalent X-states look like? Are any of them useful?
In \(su(2) \times su(2)\) what restrictions on \(g\) make it a pure state? (as an analogy to \(I/2 + \vec{a} \cdot \vec{\sigma} /2\) is pure iff \(\lvert \vec{a} \rvert = 1\)) Is there a way to categorize it?