# Prove that all $2 \times 2$ orthogonal matrices can be expressed as rotation or reflection

Let A be some $$2 \times 2$$ matrix with real entries. Prove that $$A^TA$$ = $$I$$ if and only if $$A$$ is the rotation matrix or the reflection matrix.

My Progress: It can be shown that if $$A$$ is either the rotation or reflection matrix, then $$A^T A = I$$ holds by matrix multiplication. Where I get suck is showing that if $$A$$ is a $$2 \times 2$$ orthogonal matrix, then $$A$$ must either be the Rotation or Reflection Matrix. I suppose that since $$A$$ is orthogonal, it is distance preserving - and the only $$2 \times 2$$ matrices that preserve distance are the Rotation and Reflection Matrices, but this isn't really a proof.

• Observe that $A$ also preserves orthogonality of vectors: if $x^T y = 0$, then $(Ax)^T(Ay) = x^T (A^T A) y = x^TIy = x^Ty = 0$. Similarly, $A$ maps unit vectors to unit vectors. Now consider how this constrains $Ae_1$ relative to $Ae_2$, where $e_1, e_2$ is the standard basis of $\mathbb R^2$.
– user169852
Commented Sep 20, 2018 at 17:03

With not so many variables running around, we may verify this claim algebraically. Write $A = \begin{pmatrix} a & b \\ c & d \end{pmatrix}$. Then $A^TA = \begin{pmatrix} a^2+c^2 & ab+cd \\ ab+cd & b^2+d^2 \end{pmatrix}$. So $\begin{pmatrix} a \\ c\end{pmatrix}$ and $\begin{pmatrix} b \\ d\end{pmatrix}$ are unit vectors so that $\begin{pmatrix} a \\ c\end{pmatrix}\cdot \begin{pmatrix} b \\ d\end{pmatrix} = 0$. We can check that this implies $b = \pm c$ while $d = \mp a$. So $A = \begin{pmatrix} a & c \\ c & -a \end{pmatrix}$ or $A = \begin{pmatrix} a & -c \\ c & a \end{pmatrix}$, which are the forms of a reflection composed with a rotation and a rotation, respectively.
• To clarify why $b=\pm c$: notice that $ab=-cd$ from orthogonality, so $a^2b^2=c^2d^2$; also notice that $a^2=1-c^2$ and $d^2=1-b^2$ from unit magnitude. So $(1-c^2)b^2=c^2(1-b^2)$, so $b^2=c^2$ and $b=\pm c$. We can use analogous methods to derive $a^2=d^2$, then use $ab=-cd$ to show $d=\mp a$. Commented May 9, 2020 at 1:04
Write $$a_{11}$$ as $$\cos \theta$$ so $$a_{12} = \pm \sin \theta$$ because the first row must have norm 1. Likewise, $$a_{21} = \pm \sin \theta$$ as well. It follows that $$a_{22} = \pm \cos \theta$$; with the sign chosen so as to make the rows (or equivalently, columns) orthogonal. This idenitifies all the possible forms of the matrix.