Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:quant-ph/0001014 · Physical Review A

Separability and Fourier representations of density matrices

A. O. Pittenger, Morton H. Rubin

Published Aug 18, 2000• 56 citations
arXivHTMLPDF

Abstract

Using the finite Fourier transform, we introduce a generalization of Pauli-spin matrices for d-dimensional spaces, and the resulting set of unitary matrices $S(d)$ is a basis for $d\ifmmode\times\else\texttimes\fi{}d$ matrices. If ${N=d}_{1}\ifmmode\times\else\texttimes\fi{}{d}_{2}\ifmmode\times\else\texttimes\fi{}\ensuremath{\cdot}\ensuremath{\cdot}\ensuremath{\cdot}\ifmmode\times\else\texttimes\fi{}{d}_{b}$ and ${H}^{[N]}=\ensuremath{\bigotimes}{H}^{[{d}_{k}]},$ we give a sufficient condition for separability of a density matrix $\ensuremath{\rho}$ relative to the ${H}^{[{d}_{k}]}$ in terms of the ${L}_{1}$ norm of the spin coefficients of $\ensuremath{\rho}.$ Since the spin representation depends on the form of the tensor product, the theory applies to both full and partial separability on a given space ${H}^{[N]}.$ It follows from this result that for a prescribed form of separability, there is always a neighborhood of the normalized identity in which every density matrix is separable. We also show that for every prime p and $n>1,$ the generalized Werner density matrix ${W}^{[{p}^{n}]}(s)$ is fully separable if and only if $s<~{(1+p}^{n\ensuremath{-}1}{)}^{\ensuremath{-}1}.$

DOI
10.1103/physreva.62.032313

Related papers

Complete separability and Fourier representations of n-qubit states

A. O. Pittenger, Morton H. Rubin

Necessary conditions for separability are most easily expressed in the computational basis, while sufficient conditions are most conveniently expressed in the spin basis. We use the Hadamard matrix to define the relationship between these two bases and to emphasize its interpretation as a Fourier transform. We then prove a general sufficient condition for complete separability in terms of the spin coefficients and give necessary and sufficient conditions for the complete separability of a class of generalized Werner densities. As a further application of the theory, we give necessary and sufficient conditions for full separability for a particular set of n-qubit states whose densities all satisfy the Peres condition.

arXiv:quant-ph/9912116•Physical Review A•Sep 1, 2000•20 citations

Geometry of density matrices. II. Reduced density matrices and N representability

John E. Harriman

The reduction of density matrices defined with respect to a finite basis set is considered. A basis set can be introduced into the vector space of Hermitian matrices acting on functions of the coordinates of $p$ particles for each value of $p$. An analogy with the construction of spin eigenfunctions is used to obtain a particular basis set called the reducing basis. Reduction of one of these basis elements is either one to one or maps into the origin. Any element in the preimage of a density matrix can thus be resolved into two components, one uniquely determined by the density matrix and the other arbitrary within a certin subset. A restatement of the $N$-representability problem is given, and two sufficient conditions and one necessary condition for $N$ representability are given in terms of distances and norms.

Physical Review A•Apr 1, 1978•84 citations

Separability of Density Matrices and Conditional Information Transmission

Robert R. Tucci

We give necessary and sufficient conditions under which a density matrix acting on a two-fold tensor product space is separable. Our conditions are given in terms of quantum conditional information transmission.

arXiv:quant-ph/0005119•May 28, 2000•17 citations

On a matrix equality involving partial transposition and its relation to the separability problem

Soni, Vaibhav, Rishone Deshwal, Garg, Aayush +2 more

In matrix theory, a well established relation $(AB)^{T}=B^{T}A^{T}$ holds for any two matrices $A$ and $B$ for which the product $AB$ is defined. Here $T$ denote the usual transposition. In this work, we explore the possibility of deriving the matrix equality $(AB)^Γ=A^ΓB^Γ$ for any $4 \times 4$ matrices $A$ and $B$, where $Γ$ denote the partial transposition. We found that, in general, $(AB)^Γ\neq A^ΓB^Γ$ holds for $4 \times 4$ matrices $A$ and $B$ but there exist particular set of $4 \times 4$ matrices for which $(AB)^Γ= A^ΓB^Γ$ holds. We have exploited this matrix equality to investigate the separability problem. Since it is possible to decompose the density matrices $ρ$ into two positive semi-definite matrices $A$ and $B$ so we are able to derive the separability condition for $ρ$ when $ρ^Γ=(AB)^Γ=A^ΓB^Γ$ holds. Due to the non-uniqueness property of the decomposition of the density matrix into two positive semi-definte matrices $A$ and $B$, there is a possibility to generalise the matrix equality for density matrices lives in higher dimension. These results may help in studying the separability problem for higher dimensional and multipartite system.

arXiv:2104.06117•Apr 13, 2021•0 citations

Operational criterion and constructive checks for the separability of low-rank density matrices

Paweł Horodecki, Maciej Lewenstein, Guifré Vidal +1 more

We consider low-rank density operators $\ensuremath{\varrho}$ supported on a $M\ifmmode\times\else\texttimes\fi{}N$ Hilbert space for arbitrary M and N $(M<~N),$ and with a positive partial transpose (PPT) ${\ensuremath{\varrho}}^{{T}_{A}}>~0.$ For rank $r(\ensuremath{\varrho})<~N$ we prove that having a PPT is necessary and sufficient for $\ensuremath{\varrho}$ to be separable; in this case we also provide its minimal decomposition in terms of pure product states. It follows from this result that there is no rank-3 bound entangled states having a PPT. We also present a necessary and sufficient condition for the separability of generic density matrices for which the sum of the ranks of $\ensuremath{\varrho}$ and ${\ensuremath{\varrho}}^{{T}_{A}}$ satisfies $r(\ensuremath{\varrho})+r({\ensuremath{\varrho}}^{{T}_{A}})<~2MN\ensuremath{-}M\ensuremath{-}N+2.$ This separability condition has the form of a constructive check, thus also providing a pure product state decomposition for separable states, and it works in those cases where a system of couple polynomial equations has a finite number of solutions, as expected in most cases.

arXiv:quant-ph/0002089•Physical Review A•Aug 17, 2000•151 citations

New considerations on the separability of very noisy mixed states and implications for NMR quantum computing

J.D. Bulnes, Roberto S. Sarthour, Eduardo Ribeiro de Azevêdo +5 more

We revise the problem first addressed by Braunstein and co-workers (Phys. Rev. Lett. {\bf 83} (5) (1999) 1054) concerning the separability of very noisy mixed states represented by general density matrices with the form $ρ_ε= (1-ε)M_d+ερ_1$. From a detailed numerical analysis, it is shown that: (1) there exist infinite values in the interval taken for the density matrix expansion coefficients, $-1\le c_{α_1,...,α_N}\le 1$, which give rise to {\em non-physical density matrices}, with trace equal to 1, but at least one {\em negative} eigenvalue; (2) there exist entangled matrices outside the predicted entanglement region, and (3) there exist separable matrices inside the same region. It is also shown that the lower and upper bounds of $ε$ depend on the coefficients of the expansion of $ρ_1$ in the Pauli basis. If $ρ_{1}$ is hermitian with trace equal to 1, but is allowed to have {\em negative} eigenvalues, it is shown that $ρ_ε$ can be entangled, even for two qubits.

arXiv:quant-ph/0404020•Apr 3, 2004•2 citations
Using the finite Fourier transform, we introduce a generalization of Pauli-spin matrices for d-dimensional spaces, and the resulting set of unitary matrices S(d)S(d)S(d) is a basis for d\ifmmode×\else\texttimes\fidd\ifmmode\times\else\texttimes\fi{}dd\ifmmode×\else\texttimes\fid matrices. If N=d1\ifmmode×\else\texttimes\fid2\ifmmode×\else\texttimes\fi\ensuremath⋅\ensuremath⋅\ensuremath⋅\ifmmode×\else\texttimes\fidb{N=d}_{1}\ifmmode\times\else\texttimes\fi{}{d}_{2}\ifmmode\times\else\texttimes\fi{}\ensuremath{\cdot}\ensuremath{\cdot}\ensuremath{\cdot}\ifmmode\times\else\texttimes\fi{}{d}_{b}N=d1​\ifmmode×\else\texttimes\fid2​\ifmmode×\else\texttimes\fi\ensuremath⋅\ensuremath⋅\ensuremath⋅\ifmmode×\else\texttimes\fidb​ and H[N]=\ensuremath⨂H[dk],{H}^{[N]}=\ensuremath{\bigotimes}{H}^{[{d}_{k}]},H[N]=\ensuremath⨂H[dk​], we give a sufficient condition for separability of a density matrix \ensuremathρ\ensuremath{\rho}\ensuremathρ relative to the H[dk]{H}^{[{d}_{k}]}H[dk​] in terms of the L1{L}_{1}L1​ norm of the spin coefficients of \ensuremathρ.\ensuremath{\rho}.\ensuremathρ. Since the spin representation depends on the form of the tensor product, the theory applies to both full and partial separability on a given space H[N].{H}^{[N]}.H[N]. It follows from this result that for a prescribed form of separability, there is always a neighborhood of the normalized identity in which every density matrix is separable. We also show that for every prime p and n>1,n>1,n>1, the generalized Werner density matrix W[pn](s){W}^{[{p}^{n}]}(s)W[pn](s) is fully separable if and only if s< (1+pn\ensuremath−1)\ensuremath−1.s<~{(1+p}^{n\ensuremath{-}1}{)}^{\ensuremath{-}1}.s< (1+pn\ensuremath−1)\ensuremath−1.
Necessary conditions for separability are most easily expressed in the computational basis, while sufficient conditions are most conveniently expressed in the spin basis. We use the Hadamard matrix to define the relationship between these two bases and to emphasize its interpretation as a Fourier transform. We then prove a general sufficient condition for complete separability in terms of the spin coefficients and give necessary and sufficient conditions for the complete separability of a class of generalized Werner densities. As a further application of the theory, we give necessary and sufficient conditions for full separability for a particular set of n-qubit states whose densities all satisfy the Peres condition.
The reduction of density matrices defined with respect to a finite basis set is considered. A basis set can be introduced into the vector space of Hermitian matrices acting on functions of the coordinates of ppp particles for each value of ppp. An analogy with the construction of spin eigenfunctions is used to obtain a particular basis set called the reducing basis. Reduction of one of these basis elements is either one to one or maps into the origin. Any element in the preimage of a density matrix can thus be resolved into two components, one uniquely determined by the density matrix and the other arbitrary within a certin subset. A restatement of the NNN-representability problem is given, and two sufficient conditions and one necessary condition for NNN representability are given in terms of distances and norms.
In matrix theory, a well established relation (AB)T=BTAT(AB)^{T}=B^{T}A^{T}(AB)T=BTAT holds for any two matrices AAA and BBB for which the product ABABAB is defined. Here TTT denote the usual transposition. In this work, we explore the possibility of deriving the matrix equality (AB)Γ=AΓBΓ(AB)^Γ=A^ΓB^Γ(AB)Γ=AΓBΓ for any 4×44 \times 44×4 matrices AAA and BBB, where ΓΓΓ denote the partial transposition. We found that, in general, (AB)Γ≠AΓBΓ(AB)^Γ\neq A^ΓB^Γ(AB)Γ=AΓBΓ holds for 4×44 \times 44×4 matrices AAA and BBB but there exist particular set of 4×44 \times 44×4 matrices for which (AB)Γ=AΓBΓ(AB)^Γ= A^ΓB^Γ(AB)Γ=AΓBΓ holds. We have exploited this matrix equality to investigate the separability problem. Since it is possible to decompose the density matrices ρρρ into two positive semi-definite matrices AAA and BBB so we are able to derive the separability condition for ρρρ when ρΓ=(AB)Γ=AΓBΓρ^Γ=(AB)^Γ=A^ΓB^ΓρΓ=(AB)Γ=AΓBΓ holds. Due to the non-uniqueness property of the decomposition of the density matrix into two positive semi-definte matrices AAA and BBB, there is a possibility to generalise the matrix equality for density matrices lives in higher dimension. These results may help in studying the separability problem for higher dimensional and multipartite system.
We consider low-rank density operators $\ensuremath{\varrho}$ supported on a $M\ifmmode\times\else\texttimes\fi{}N$ Hilbert space for arbitrary M and N (M< N),(M<~N),(M< N), and with a positive partial transpose (PPT) ${\ensuremath{\varrho}}^{{T}_{A}}>~0.$ For rank $r(\ensuremath{\varrho})<~N$ we prove that having a PPT is necessary and sufficient for $\ensuremath{\varrho}$ to be separable; in this case we also provide its minimal decomposition in terms of pure product states. It follows from this result that there is no rank-3 bound entangled states having a PPT. We also present a necessary and sufficient condition for the separability of generic density matrices for which the sum of the ranks of $\ensuremath{\varrho}$ and ${\ensuremath{\varrho}}^{{T}_{A}}$ satisfies $r(\ensuremath{\varrho})+r({\ensuremath{\varrho}}^{{T}_{A}})<~2MN\ensuremath{-}M\ensuremath{-}N+2.$ This separability condition has the form of a constructive check, thus also providing a pure product state decomposition for separable states, and it works in those cases where a system of couple polynomial equations has a finite number of solutions, as expected in most cases.
We revise the problem first addressed by Braunstein and co-workers (Phys. Rev. Lett. {\bf 83} (5) (1999) 1054) concerning the separability of very noisy mixed states represented by general density matrices with the form ρε=(1−ε)Md+ερ1ρ_ε= (1-ε)M_d+ερ_1ρε​=(1−ε)Md​+ερ1​. From a detailed numerical analysis, it is shown that: (1) there exist infinite values in the interval taken for the density matrix expansion coefficients, −1≤cα1,...,αN≤1-1\le c_{α_1,...,α_N}\le 1−1≤cα1​,...,αN​​≤1, which give rise to {\em non-physical density matrices}, with trace equal to 1, but at least one {\em negative} eigenvalue; (2) there exist entangled matrices outside the predicted entanglement region, and (3) there exist separable matrices inside the same region. It is also shown that the lower and upper bounds of εεε depend on the coefficients of the expansion of ρ1ρ_1ρ1​ in the Pauli basis. If ρ1ρ_{1}ρ1​ is hermitian with trace equal to 1, but is allowed to have {\em negative} eigenvalues, it is shown that ρερ_ερε​ can be entangled, even for two qubits.