I have a question about the two equations:

  1. Any matrix in $SU(2)$ could be parametrized as $$ R_{\hat{n}}(\theta) = \cos\left(\frac{\theta}{2}\right)I-i\sin\left(\frac{\theta}{2}\right)(\hat{n}\cdot\vec\sigma) $$
  2. And the $U_3$ gate in qiskit is defined as $$ U_3(\gamma,\beta,\delta) = \begin{bmatrix} \cos\left(\frac{\gamma}{2}\right) & -e^{i\delta} \sin\left(\frac{\gamma}{2}\right) \\ e^{i\beta} \sin\left(\frac{\gamma}{2}\right) & e^{i(\delta + \beta)} \cos\left(\frac{\gamma}{2}\right) \end{bmatrix} $$

For the first equation, I think the $\theta/2$ term is kind of relevant to the Dirac belt trick: the electron spin state will negate under a $2\pi$ rotation, so $4\pi$ will return to the original state. For the second equation, on the other hand, I think $\gamma/2$ is because on the Bloch sphere it looks like $|0\rangle$ and $|1\rangle$ are on the same line pointing toward the opposite directions, but they're orthogonal states. I'm wondering if there're any connections between the two cases when we divide $\theta$ or $\gamma$ by 2. Thanks!!

PS: In my understanding, $\theta$ refers to the angle of rotation along the axis $\hat n$ following the right-hand rule, and $\gamma$ here is the included angle of $\hat n$ and $z$-axis in the spherical coordinate.


2 Answers 2


Yes, the referenced terms divided by two are directly related. First note that $R_{\hat n}(\theta) \in SU(2)$ is isomorphic to $U_3 \in U(2)$. You can see this by rewriting $U_3$ as an arbitrary element of $SU(2)$ multiplied by the global phase $e^{i(\delta+\beta)/2} \in U(1)$.

In other words, $$SU(2) \ni U_3^\prime =U_3\times e^{-i(\delta+\beta)/2} = \begin{bmatrix} e^{-i(\delta+\beta)/2}\cos\left(\frac{\gamma}{2}\right) & -e^{i(\delta-\beta)/2} \sin\left(\frac{\gamma}{2}\right) \\ e^{-i(\delta-\beta)/2} \sin\left(\frac{\gamma}{2}\right) & e^{i(\delta + \beta)/2} \cos\left(\frac{\gamma}{2}\right) \end{bmatrix},$$ such that $U_3^\prime$ is simply an alternative parameterization of an arbitrary element of $SU(2)$, and $U_3^\prime \cong U_3$ up to global phase.

In reference to your comments on the Bloch sphere, the Bloch sphere is a Riemann sphere (i.e. the complex projective line), not a Euclidean 2-sphere. The former is a part of elliptic geometry, the latter of spherical geometry. Elliptic geometry is essentially spherical geometry with antipodal points identified. This is why orthogonal states appear as antipodal points in a Bloch sphere representation. As you suggested this is all related to spinors ("the Dirac trick"), but that rabbit hole goes very deep.


With $$ \vec\sigma = \begin{bmatrix}\begin{bmatrix}0 & -i\\ i &0 \end{bmatrix}\begin{bmatrix} 0 & -1 \\1 & 0\end{bmatrix}\begin{bmatrix}-1 & 0 \\ 0 & -1 \end{bmatrix} \end{bmatrix} $$ For the first equation I like to think that defining the formula with $\frac{\theta}{2}$ is a nice way to have a direct understanding of the effect on the Bloch sphere.
if $\hat{n}=[1\space 0\space 0]$ then $R_{\hat{n}}(\theta) =R_X(\theta)$ and $R_X(\theta)$ is a $\theta$ rotation around the X axis of the Bloch Sphere.
if $\hat{n}=[0\space 1\space 0]$ then $R_{\hat{n}}(\theta) =R_Y(\theta)$ and $R_Y(\theta)$ is a $\theta$ rotation around the Y axis of the Bloch Sphere.
if $\hat{n}=[0\space 0\space 1]$ then $R_{\hat{n}}(\theta) =R_Z(\theta)$ and $R_Z(\theta)$ is a $\theta$ rotation around the Z axis of the Bloch Sphere.

Essentially $R_{\hat{n}}(\theta)$ is a generalization of the rotation matrix hence the need of $\frac{\theta}{2}$. Another nice consequence is that for the Pauli gates $X=R_X(\pi)$, $Y=R_Y(\pi)$, $Z=R_Z(\pi)$

For the second equation the reasoning is essentially the same.

$U_3$ is a generalization of the $R_\phi$, $T$, $S$, $Z$ gates and defining it with $\frac{\gamma}{2}$ allows to have this nice identities that links to the axis rotation matrix.

$U_3(\gamma,0,\pi/2) = R_X(\gamma)$
$U_3(\gamma,0,0) = R_Y(\gamma)$
$U_3(0,0,\delta) = R_\phi(\delta)$

Remark : There is no relation between $U_3$ and $Z$


  • 1
    $\begingroup$ "Essentially $R_n(θ)$ is a generalization of the rotation matrix hence the need of $\theta/2$" - this doesn't really explain the factor of two; when one represents ordinary rotations just working in $\mathbb{R}^3$ no such factor appears. Its more due to the unique relationship between the rotation group $SO(3)$ that we use to visualize Bloch sphere rotations and the special unitary group $SU(2)$ that we use to manipulate two level systems. $\endgroup$
    – forky40
    Apr 4, 2021 at 19:56
  • $\begingroup$ @forky40 thanks for the comment. I think your remark is clearly aligned with what I wrote "For the first equation [...] θ/2 is a nice way to have a direct understanding of the effect on the Bloch sphere." $\endgroup$
    – PilouPili
    Apr 4, 2021 at 21:21
  • $\begingroup$ @forky40 Thanks for the comment:) How are the two groups $SO(3)$ and $SU(2)$ related in this sense? $\endgroup$
    – ZR-
    Apr 4, 2021 at 23:14
  • $\begingroup$ @ZR- I can't do the explanation justice in a comment but if you search for terms like "$SU(2)$ $SO(3)$ double cover" or "dirac belt trick" you should find tons of material about this relationship from either a group theory or physics (but still group theory) perspective $\endgroup$
    – forky40
    Apr 5, 2021 at 1:48
  • $\begingroup$ @forky40 Thanks! $\endgroup$
    – ZR-
    Apr 5, 2021 at 2:46

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.