You already have nice solutions, but one might wonder where such a problem comes from. This answer provides a setting of differential equations. Once it is put in this setting, the result of the integral follows from a general theorem.
When solving the Sturm-Liouville eigenvalue problem
$$
-u''(t)=\lambda u(t),\quad u(0)=0,\ 2u'(1)-u(1)=0,
$$
one first finds that the general solution to the differential equation is
$$
u(t)=A\cos\sqrt{\lambda} t+B\sin\sqrt{\lambda}t.
$$
Imposing the first boundary condition, the cosine term disappears, and one get
$$
u(t)=B\sin\sqrt{\lambda}t.
$$
Now, the second boundary condition implies that
$$
0=2u'(1)-u(1)=2B\sqrt{\lambda}\cos\sqrt{\lambda}-B\sin\sqrt{\lambda}.
$$
Looking for non-trivial solutions, $B\neq 0$, and we find that $\lambda$ must satisfy
$$
\tan\sqrt{\lambda}=2\sqrt{\lambda}
$$
This equation has infinitely many positive solutions, denote them by $\lambda_k$, $k=1$, $2$, $\ldots$. These are called eigenvalues of the Sturm-Liouville problem above. The corresponding functions $u_k(t)=\sin\sqrt{\lambda_k}t$ are called eigenfunctions.
There is a very nice theorem that fits in this setting:
Theorem Eigenfunctions corresponding to different eigenvalues are
orthogonal.
In our case, orthogonal means that
$$
\int_0^1 u_j(t) u_k(t)\,dt=0\quad\text{if }j\neq k.
$$
With $\alpha=\sqrt{\lambda_j}$ and $\beta=\sqrt{\lambda_k}$, it follows that your integral is zero.