The Stinespring dilation theorem states that any CPTP map $\Lambda$ on a system with Hilbert space $\mathcal{H}$ can be represented as $$\Lambda[\rho]=tr_\mathcal{A}(U^\dagger (\rho\otimes |\phi\rangle\langle \phi|)U)$$ where $\mathcal{A}$ is an ancilla, $|\phi\rangle\in\mathcal{A}$ is an arbitrary pure state, and $U$ is a unitary on the joint system $\mathcal{H}\otimes\mathcal{A}$. Importantly, this works for any choice of the state $|\phi\rangle\in\mathcal{A}$ -- that is, neither does $|\phi\rangle\in\mathcal{A}$ have to depend on $\Lambda$, nor is a specific choice of $|\phi\rangle\in\mathcal{A}$ required.

Would this representation still be valid if, instead of attaching an ancilla in the pure state $|\phi\rangle\langle\phi|$, we instead attached an ancilla in some arbitrary mixed state $\sigma$? That is, given a state $\sigma$ of $\mathcal{A}$, I would like to know whether for every CPTP channel $\Lambda$, there exists a unitary $U$ on $\mathcal{H}\otimes\mathcal{A}$ such that $$\Lambda[\rho]=tr_\mathcal{A}(U^\dagger (\rho\otimes \sigma)U).$$

  • $\begingroup$ To be clear: are you asking whether, for any map you specify, there exists a choice of $U$ and $\sigma$ that implements the map, or whether for any map and $\sigma$ that you specify, there exists a $U$ that implements the map? $\endgroup$
    – DaftWullie
    Mar 1, 2021 at 11:15
  • $\begingroup$ Clearly, the mixed state cannot be arbitrary, see Mateus' answer and quantumcomputing.stackexchange.com/questions/12951/… . What makes sense is to ask whether, given a mixed-state dilation, you can find a pure one on the same ancilla system. I think this is generally false, as the Stinespring dilation may need a maximal ancilla system while the mixed-state version can get away with a lower-dimensional one using a higher-rank mixed state. $\endgroup$ Mar 1, 2021 at 12:28
  • $\begingroup$ Note that if you have found a mixed-state dilation, the mixed state is arbitrary, as long as the rank is fixed (same argument as for pure states!). $\endgroup$ Mar 1, 2021 at 12:30
  • 3
    $\begingroup$ Please mark cross-posts to qc.se or other se sites, for the very least. Otherwise your creating double work, which is rather inconsiderate. $\endgroup$ Mar 1, 2021 at 21:22
  • 1
    $\begingroup$ See physics.stackexchange.com/questions/617686/… $\endgroup$ Mar 1, 2021 at 21:23

2 Answers 2


No, that doesn't work. It's fine to use an arbitrary pure state because the unitary $U$ can always be used to take it to any pure state you want. This argument doesn't work for a mixed state, as unitaries cannot take mixed states to pure states.

As a concrete example, consider the CPTP map $$ \Lambda(\rho) = |0\rangle\langle 0| \operatorname{tr}(\rho), $$ and let $\sigma = I/d_{\mathcal A}$, the maximally mixed state on $\mathcal A$. If you now apply this CPTP map to $I/d_{\mathcal H}$, the maximally mixed state on $\mathcal H$, you should get $|0\rangle \langle 0|$, but $$ \operatorname{tr}_\mathcal{A}(U^\dagger (\rho\otimes \sigma)U) = \operatorname{tr}_\mathcal{A}(U^\dagger (I/d_{\mathcal H}\otimes I/d_{\mathcal A})U) = I/d_{\mathcal H}$$ for any unitary $U$.

  • 1
    $\begingroup$ "as unitaries cannot take mixed states to pure states" -- that's no reason. $\endgroup$ Mar 1, 2021 at 21:28
  • $\begingroup$ Why not? If you could take $\sigma$ to a pure state, you would be back at the usual Stinespring dilation with an arbitrary pure ancilla. $\endgroup$ Mar 2, 2021 at 9:41
  • $\begingroup$ True, but you use it as an argument for the converse direction. $\endgroup$ Mar 2, 2021 at 10:22
  • 2
    $\begingroup$ I don't. I'm just saying that the argument why you can get away with an arbitrary pure state doesn't hold for an arbitrary mixed state. You're just being pedantic. $\endgroup$ Mar 2, 2021 at 10:46
  • $\begingroup$ I see. It seems the "This doesn't hold" was ambiguous. $\endgroup$ Mar 2, 2021 at 11:06

No. If you take $\sigma=\tfrac1{d_A}\mathbb I$, you will have that for $\rho=\tfrac1{d_S}\mathbb I$, $$ \mathrm{tr}_A(U^\dagger(\rho\otimes\sigma)U)=\tfrac1{d_S}\mathbb I\ , $$ and thus, it will not be possible to implement any channel for which $$ \Lambda(\tfrac1{d_S}\mathbb I)\ne \tfrac1{d_S}\mathbb I\ , $$ such as any channel mapping all inputs to a constant output other than the identity.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.