Consider a measurement operator ("observable") $\hat O$ which has ("a spectrum of") only two distinct eigenvectors; formally

$$\hat O |\bullet\rangle := r_{\bullet}~|\bullet\rangle, \qquad \hat O |\circ\rangle := r_{\circ}~|\circ\rangle, \qquad \langle \bullet |\circ\rangle = 0, $$

where the eigenvalues $r_{\bullet}$ and $r_{\circ}$ are (suitable) real numbers.

Is there a time-dependent unitary operator $\hat U[~t~]$ such that

$$ \frac{|\bullet\rangle + |\circ\rangle}{\sqrt 2} = \hat U[~t_{\text{blend}}~] ~ |\bullet\rangle $$ ?

If so, can this operator be expressed in the form $\hat U[~t~] := \text{Exp}[~\frac{-i~t}{\hbar}~\hat H[~t~]~]$ ?

If so, can the corresponding operator $\hat H[~t~]$ be expressed in terms of the given operator $\hat O$, or its eigenvectors or eigenvalues, and, as necessary, the coordinate value $t_{\text{blend}}$ and the coordinate variable $t$?

(along with correcting the above "evolution equation" stated within the bra-ket formalism):

Within the density matrix formalism it could be asked correspondingly whether there is a time-dependent unitary operator $\hat U[~t~]$ such that

$$ \sqrt{\frac{1}{2}} \left( \array{ |\bullet\rangle\langle\bullet | & |\bullet\rangle\langle\circ | \cr |\circ\rangle\langle\bullet | & |\circ\rangle\langle\circ | } \right) = \hat U[~t_{\text{blend}}~] ~ \left( \begin{array}{cc} |\bullet\rangle\langle\bullet | & 0 \\ 0 & |\circ\rangle\langle\circ | \end{array} \right) ~ \hat U[~t_{\text{blend}}~]^{\dagger}.$$


1 Answer 1


OK firstly operators act on kets as vectors, so for some unitary operator $\hat{U}$, the action of $\hat{U}$ sends $|\psi\rangle\rightarrow \hat{U}|\psi\rangle$ not $\hat{U}|\psi\rangle\hat{U}^\dagger$, which is normally not a well defined expression.

For the rest of your question, since we are only dealing with a two dimensional space (or at least only a two dimensional subspace of some larger space) it is clearest to use matrix notation. So \begin{equation} \hat{U}(t) = \left(\begin{array}{cc}a&b\\c&d\end{array}\right)\quad |\bullet\rangle = \left(\begin{array}{c}1\\0\end{array}\right)\quad |\circ\rangle = \left(\begin{array}{c}0\\1\end{array}\right) \end{equation}

In this notation your blending condition becomes \begin{equation} \frac{1}{\sqrt{2}}\left(\begin{array}{c}1\\1 \end{array}\right) = \left(\begin{array}{cc}a&b\\c&d\end{array}\right)\left(\begin{array}{c}1\\0\end{array}\right) = \left(\begin{array}{c}a\\c\end{array}\right) \end{equation} The requirement that $\hat{U}$ be unitary then fixes $b$ and $d$ up to a sign, so it is easy to work out that\begin{equation} \hat{U} = \frac{1}{\sqrt{2}}\left(\begin{array}{cc}1& \pm 1\\1&\mp 1\end{array}\right) \end{equation}

If we put the minus sign in the top right (the case where it is in the bottom right differs only by a change of basis) then $\hat{U}$ has a standard representation in terms of Pauli matrices\begin{equation} \hat{U} = \exp\left(-\imath\frac{\pi}{4}\sigma^{\,y}\right) \end{equation}

  • $\begingroup$ By Symmetry: "OK firstly operators act on kets as vectors [...] $|\psi\rangle \rightarrow \hat U |\psi\rangle$" -- Aw; thanks for clarifying (+1). So, what I wrote at least doesn't work with "bra-ket" notation ... (I may need a while for deciding whether and how to correct/edit my OP regarding this "formality".) "[...] clearest to use matrix notation." -- Great. However: I'm missing some explicit $t$ dependence in your expressions for operator $\hat U[~t~]$. (This still seems a substantive part of my question which you answer didn't address.) $\endgroup$
    – user12262
    Commented Apr 20, 2015 at 5:14
  • 1
    $\begingroup$ @user12262 Just replace $\pi/4$ in the expression for $\hat{U}$ by anything of the form $\Omega t$, where $\Omega$ has dimensions of frequency and $t$ is time. Then you get your required unitary operator by setting $\Omega t_\mathrm{blend} = \pi/4$. $\endgroup$ Commented Apr 20, 2015 at 9:59
  • $\begingroup$ @Mark Mitchison: "Just replace $\pi/4$ in the expression for $\hat U$ by anything of the form $\Omega~t$ [with] $\Omega~t_{\text{blend}} = \pi/4$. Can you please prove that: $$\hat U[~t_{\text{blend}}~]=\frac{1}{\sqrt 2}\left(\array{1&-1\cr 1&1}\right) =?= \text{Exp}[~i~\frac{\pi}{4}~\hat \sigma_y~] = \text{Exp}[~i~\frac{\pi}{4}~{\left(\array{0&-i \cr i&0}\right)}~]=\text{Exp}[~\frac{\pi}{4}~\left(\array{ 0 & 1 \cr -1 & 0}\right)~]$$ ? (Meanwhile I ponder this ...) $\endgroup$
    – user12262
    Commented Apr 20, 2015 at 18:25
  • 1
    $\begingroup$ @user12262 firstly I realised I missed a minus sign in my expression for $\hat{U}$. The expression for $\hat{U}$ follows straightforwardly from the power series definition of the matrix exponential and the fact that $\hat{\sigma}_y^2 = \hat{I}$ $\endgroup$ Commented Apr 20, 2015 at 18:36
  • 1
    $\begingroup$ @user12262 You have the identity, which you can prove using the arguments in By Symmetry's comment: $$ e^{i \theta\mathbf{x}\cdot\mathbf{\sigma}} = \cos(\theta) + i\hat{\mathbf{x}}\cdot \mathbf{\sigma} \sin(\theta), $$ where $\sigma$ here is a vector of Pauli operators $\sigma^{x,y,z}$, and $\hat{\mathbf{x}}$ is a unit vector parallel to $\mathbf{x}$. $\endgroup$ Commented Apr 20, 2015 at 21:05

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.