I never seen this notation, but I assume $[[A]]_{\alpha}$ stands for the value of the formula $A$ in the valuation $\alpha$, am I correct?
Since $F$ and $G$ have no atomic propositions in common, assume $p_{1}, ... , p_{n}$ are the atomic propositions in $F$ and $q_{1}, ... , q_{m}$ the atomic propositions in $G$, which satisfy $\{p_{1}, ... , p_{n}\}\cap\{q_{1}, ... , q_{m}\}=\emptyset$.
Now, if $F$ and $G$ are both non-tautologies, as Mauro ALLEGRANZA suggests, take valuations $\alpha$ and $\beta$ such that $[[F]]_{\alpha}=0$ and $[[G]]_{\beta}=0$. And now comes the important part: define the valuation $\nu$ such that
$$[[p_{1}]]_{\nu}=[[p_{1}]]_{\alpha}, ... , [[p_{n}]]_{\nu}=[[p_{n}]]_{\alpha}\quad\text{and}\quad [[q_{1}]]_{\nu}=q_{1}]]_{\beta}, ... , [[q_{m}]]_{\nu}=[[q_{m}]]_{\beta},$$
what can be done since all $p_{i}$ and $q_{j}$ are atomic propositions. Since $p_{1}$ through $p_{n}$ are the atomic propositions in $F$, and $\nu$ and $\alpha$ agree in these statements, $[[F]]_{\nu}=[[F]]_{\alpha}=0$; analogously, $[[G]]_{\nu}=[[G]]_{\beta}=0$, meaning that
$$[[F\vee G]]_{\nu}=[[F]]_{\nu}\vee[[G]]_{\nu}=0\vee 0=0,$$
what would imply that $F\vee G$ is NOT a tautology, against our hypothesis. By reduction to absurd, we must derive that either $F$ or $G$ is a tautology.