One way to solve this kind of problem is to find the reduced mass of the rod along the contact normal.
Problem Setup
Consider the general situation below. A moving rigid body is impacted by a sphere. The motion of the rigid body is tracked by the location $\vec{r}_1$, velocity $\vec{v}_1$ of the center of mass and the rotational velocity $\vec{\omega}_1$. The impact occurs on some point A with location $\vec{r}_A$, and along the contact normal direction $\hat{n}$. The response depends on the vector $\vec{c} = \vec{r}_1 - \vec{r}_A$ describing the location of the center of mass relative to the contact point.

After the impact, there is a step in motion of the center of mass described by the vectors $\Delta \vec{v}_1$ and $\Delta \vec{\omega}_1$. The impact consists of an exchange of momentum with magnitude $J$ (called an impulse) along the contact normal $\hat{n}$ such that $$\begin{aligned} \Delta \vec{v}_1 & = \tfrac{1}{m_1} \hat{n} J \\ \Delta \vec{\omega}_1 & = \mathrm{I}_1^{-1} ( -\vec{c} \times \hat{n} J) \end{aligned} \tag{1}$$
General 3D solution
The kinematics of the contact point A after the impact defines the impulse magnitude $J$. Before impact, the speed of approach is
$$v_{\rm imp} = \hat{n} \cdot( \vec{v}_1 + \vec{c} \times \vec{\omega}_1 - \vec{v}_2 ) \tag{2}$$
where $\vec{v}_2$ is the velocity of the impacting body. And after the impact the speed of release is equal and opposite of the speed of approach for an elastic contact.
$$ v_{\rm rel} = - v_{\rm imp} $$
or the change in speed must be twice the impact speed
$$ \hat{n} \cdot(\Delta \vec{v}_1 + \vec{c} \times \Delta \vec{\omega}_1 - \Delta \vec{v}_2 ) = - 2 \hat{n} \cdot( \vec{v}_1 + \vec{c} \times \vec{\omega}_1 - \vec{v}_2 ) \tag{3}$$
Use (1) in (3) to find $J$ as
$$ J = \frac{2 \hat{n} \cdot( \vec{v}_1 + \vec{c} \times \vec{\omega}_1 - \vec{v}_2 )}{\frac{1}{m_1}-\hat{n}\cdot(\vec{c}\times \mathrm{I}_1^{-1} \vec{c}\times \vec{n}) + \frac{1}{m_2}} $$
or how I prefer to state it
$$ J = 2\; m_{\rm imp}\; v_{\rm imp} \tag{5} $$
with the reduced mass defined as $$m_{\rm imp} = \left( \tfrac{1}{m_1}-\hat{n}\cdot(\vec{c}\times \mathrm{I}_1^{-1} \vec{c}\times \vec{n}) + \tfrac{1}{m_2} \right)^{-1} \tag{6}$$
Notice that for two spheres the above reduces to $m_{\rm imp} = \left( \tfrac{1}{m_1} + \tfrac{1}{m_2} \right)^{-1}$ which is a well known formula.
Once $J$ is known, then the motion afterward is calculated from (1)
Planar Projection
Now to make the planar projection of the above with $\hat{n} = \pmatrix{n_x \\ n_y}$ and $\vec{c} = \pmatrix{c_x \\ c_y}$
and use it in (6) to get
$$ m_{\rm imp} = \left( \frac{1}{m_1} + \frac{(c_x n_y - c_y n_x)^2}{I_1} + \frac{1}{m_2} \right)^{-1} \tag{7}$$
And the change motion from (1) and (5) in 2D
$$ \begin{aligned}
\Delta \vec{v}_1 & = \pmatrix{ \tfrac{n_x}{m_1} J \\ \tfrac{n_y}{m_1} J } \\
\Delta \omega_1 & = \tfrac{n_x c_y - n_y c_x}{I_1} J
\end{aligned} \tag{8}$$
Specific Problem
The solution to the question is found by using the following geometry
$$ \begin{aligned}
I_1 & = \tfrac{m}{12} \ell^2 \\
\vec{c} & = \pmatrix{-\tfrac{\ell}{4} \\ 0} \\
\vec{n} & = \pmatrix{0 \\ 1}
\end{aligned} $$
which results in $m_{\rm imp} = \left( \frac{1}{m_1} + \frac{3}{4 m_1} + \frac{1}{m_2} \right)^{-1}$ and so to get $J$ from (5) and the motion from (1).
The choice of $\hat{n} = \pmatrix{0\\1}$ is because the contact normal has to be normal to both surfaces, and thus perpendicular to the rod. It would be a mistake to assume that $\hat{n}$ must be parallel to the impacting velocity $\vec{v}_2$ which as you can see in my diagram they are at an angle to each other.