The Theorem
Theorem 11.8 of Nielsen and Chuang collects the basic properties of the von Neumann entropy of a quantum state. On a system whose state space has dimension , how small and how large can the entropy be, and which states reach each bound? When two systems together are in a pure state, how do the entropies of the two systems on their own compare? And what is the entropy of a mixture of states with probabilities when the states live on mutually orthogonal subspaces, or when the th state is the product of a pure state of one system with a state of another, the orthonormal?
The library answers them in seven theorems on states of quantum systems, each bound and each case of equality a theorem of its own.
A Quantum System
A quantum system is a physical system whose state space is a finite dimensional complex Hilbert space.✦ The composite system of two systems and is the system whose state space is the tensor product of the state spaces of and .✦
A state of is a positive operator of unit trace on the state space of .✦ A pure state of is a unit vector in the state space of .✦ The pure state determines the operator on that space, the rank one projector onto .✦ A state of is pure when for some pure state of .✦
The maximally mixed state of a system whose state space is nonzero, of dimension , is , the identity on that space divided by .✦
Reduced States, Products and Mixtures
The reduced state of a state of is the state of given by its partial trace over , an operator on the state space of .✦ The reduced state of on is its partial trace over .✦
The product state of a state of and a state of is the tensor product of the two operators on the state space of , acting as on the first factor and as on the second.✦
The mixture of states of , indexed by a finite set, with probabilities , nonnegative and summing to one, is the convex combination .✦
The support of a state of is the range of , the subspace spanned by its eigenvectors with nonzero eigenvalues.✦
The Entropy
The Shannon entropy of probabilities is , with .✦ The von Neumann entropy of a state of is the real number (1)✦ the Shannon entropy of the eigenvalues of . Logarithms are natural throughout, and entropies are measured in nats.
The Claim
Theorem 1. Let and be quantum systems, the dimension of the state space of , and a state of . Then (2)✦ (3)✦ (4)✦ and, if the state space of is nonzero, (5)✦ If is a pure state of , its two reduced states have equal entropy: (6)✦ Let be probabilities indexed by a finite set. If are states of with pairwise orthogonal supports, then (7)✦ If the pure states of , indexed by the same finite set as the , form an orthonormal family and the are any states of , then (8)✦
What Is Checked
The statements are made over the constructions defined here: the states of a system and of a composite system, the projector of a pure state and the condition of purity, the entropy (1), the maximally mixed state, the mixture , the support, the product state and the two reduced states.
The library proves that these constructions are states: the projector onto a unit vector, the mixture , the product of two states, the reduced states and of a state of , and are each positive operators of unit trace.