nLab
density matrix

Context

Measure and probability theory

Physics

physics, mathematical physics, philosophy of physics

Surveys, textbooks and lecture notes


theory (physics), model (physics)

experiment, measurement, computable physics

Contents

Idea

In quantum mechanics a density matrix is a linear endomorphism of a Hilbert space of pute quantum states that represents a statistical ensemble? of quantum states, hence mixed states.

The space of density matrices inside all suitable endomorphisms is called the Bloch region.

Notation

In the following definition, we use Dirac’s “bra-ket” notation for vectors where a state vector, describing a pure state of a quantum system, is represented by a “ket” which is a column vector,

ψ α=( x n1 x n x n+1 ). {|\psi_{\alpha}\rangle} = \left( \begin{aligned} \vdots \\ x_{n-1} \\ x_{n} \\ x_{n+1} \\ \vdots \end{aligned} \right) .

The Hermitian adjoint, ψ α=(ψ α) {\langle\psi_{\alpha}|} = ({|\psi_{\alpha}\rangle})^{\dagger}, is called a “bra” (hence “bra(c)ket”) and is a row vector.

The density operator of an ensemble

Suppose we have a quantum state QQ that arises from some random process such that the state ψ α{|\psi_{\alpha}\rangle} has a probability p α[0,1]p_{\alpha} \in [0,1] (we often speak of having ‘prepared’ the state with the associated probability). The possible states ψ α{|\psi_{\alpha}\rangle} need not be orthogonal and we thus call such a collection of states a mixed state. More specifically, a mixed state is often described as an ensemble? of quantum systems.

Suppose we now measure some observable A\mathbf{A} on the system as a whole, i.e. on the ensemble. The expectation value of A\mathbf{A} over the state ψ α{|\psi_{\alpha}\rangle} is A α=ψ αAψ α\langle{A}\rangle_{\alpha} = \langle\psi_{\alpha} | \mathbf{A} |\psi_{\alpha}\rangle. Over the entire ensemble, this becomes

A= αp αA α= αp αtrψ αψ αA. \langle{A}\rangle = \sum_{\alpha} p_{\alpha} \langle{A}\rangle_{\alpha} = \sum_{\alpha} p_{\alpha} tr {|\psi_{\alpha}\rangle} {\langle\psi_{\alpha}|} \mathbf{A} .

Given the above, we define the density operator to be

ρ= αρ αψ αψ α. \mathbf{\rho} = \sum_{\alpha} \rho_\alpha {|\psi_{\alpha}\rangle}{\langle\psi_{\alpha}|} .

We call the matrix representation of the density operator, relative to a given basis, the density matrix.

Characterisation

An operator ρ\rho is the density operator associated to some ensemble if and only if it is a positive operator with trace 1. (Nielsen and Chuang Theorem 2.5, p. 101)

Coherence

Diagonal density matrices with at least two non-zero terms on the diagonal represent mixed states. Density matrices that posses non-zero off-diagonal terms represent superposition states. Such states are referred to as coherent and the off-diagonal entries are called the coherences. Any physical process that has the effect of suppressing the coherences is known as decoherence.

Limitations

Note that a density operator, as the representation of the state of a quantum system, is less restrictive than a state vector which specifies the wavefunction. On the other hand, two different state vectors can give rise to the same density operator. However, in that case, the two vectors are the same up to a phase, so arguably the density operator still describes the physical state unambiguously.

More controversially, two entirely different probabilisitic combinations of state vectors can give rise to the same density operator. Roger Penrose, for one, has argued that this means that that the density operator does not describe mixed states unambiguously. But one can also argue the reverse: that mixed states with the same operator really are the same physical state, since they are observationally indistinguishable.

References

  • Nielsen, M, and Chuang, I. Quantum Computation and Quantum Information, Cambridge University Press, Cambridge, 2000.

  • Schumacher, B. and Westmoreland, M. Q-PSI: Quantum Processes, Systems, and Information, Cambridge University Press, Cambridge, 2010.

  • Greg Kuperberg, section 1.4 of A concise introduction to quantum probability, quantum mechanics, and quantum computation, 2005 (pdf)

Revised on March 20, 2014 02:12:17 by Robin Adams? (145.116.190.180)