zero cross switching triac
The way you've drawn it, the relay does short the triac's A1/A2 and therefore it can not fire anymore even if you drive the gate constantly. As I understand triac-output optocouplers, Rt1 should be superfluous, BTW.
If you want to add a zero cross start to the circuit you'd need to have the triac in series with the relay contact/load (and switch the relay first, then the triac).
The sparks you noticed in your experiment without switching the relay are probably due to your load not being resistive. A 250V (or 140V if you use 110V mains) VDR in parallel with your load should suppress those.