Earlier, we claimed that one of the most important features of the density operator formalism is its ability to describe the quantum state of a subsystem of a composite system.
Let us now show you how this works.
Given a quantum state of the composite system AB described by some density operator ρAB, we obtain reduced density operators ρA and ρB of the subsystems A and B (respectively) by the partial trace:
ρABρAB⟼ρA=partialtraceoverBtrBρAB⟼ρB=partialtraceoverAtrAρAB
We will revisit the notion of partial trace quite a few times, but for now we simply define the partial trace over B (or A) first on a tensor product of two operators A⊗B as
trB(A⊗B)trA(A⊗B)=A(trB)=(trA)B,
and then extend to any operator on HA⊗HB by linearity.
Here is a simple example.
Suppose a composite system AB is in a pure entangled state ∣ψAB⟩.
We can always write this as
∣ψAB⟩=i∑ci∣ai⟩⊗∣bi⟩,
where ∣ai⟩ and ∣bj⟩ are two orthonormal bases (e.g. the Schmidt bases, from Exercise 5.14.13), and where ∑i∣ci∣2=1 (due to the normalisation).
The corresponding density operator of the composite system is the projector ρAB=∣ψAB⟩⟨ψAB∣, which we can write as
ρAB=∣ψAB⟩⟨ψAB∣=i,j∑cicj⋆∣ai⟩⟨aj∣⊗∣bi⟩⟨bj∣
Let us compute the reduced density operator ρA by taking the partial trace over B:
ρA=trBρAB=trB∣ψAB⟩⟨ψAB∣=trBi,j∑cicj⋆∣ai⟩⟨aj∣⊗∣bi⟩⟨bj∣=i,j∑cicj⋆∣ai⟩⟨aj∣(tr∣bi⟩⟨bj∣)=i,j∑cicj⋆∣ai⟩⟨aj∣δij⟨bi∣bj⟩=i∑∣ci∣2∣ai⟩⟨ai∣.
So, in the ∣ai⟩ basis, the reduced density matrix ρA is diagonal, with entries pi=∣ci∣2.
Similarly, if we take the partial trace over A, then we get ρB=∑i∣ci∣2∣bi⟩⟨bi∣.
In particular, if dimHA=dimHB=d, then the maximally mixed state
∣ψAB⟩=d1i∑d∣ai⟩∣bi⟩,
in the (d×d)-dimensional Hilbert space HA⊗HB is such that the reduced density operators ρA and ρB are also the maximally mixed states of their respective subsystems: ρA=ρB=d11.
It follows that the quantum states of individual qubits in any of the Bell states are maximally mixed: their density matrix is 211.
A bipartite state such as
21(∣00⟩+∣11⟩)
guarantees perfect correlations when each qubit is measured in the standard basis: the two outcomes are “0 and 0” or “1 and 1” (which are equally likely), and we will never observe e.g. “0 and 1”, but the outcome of either single-qubit subsystem is completely random.