Three things will prevent an SCR, once triggered, from latching on: insufficient holding current, insufficient voltage across the SCR, and a defective SCR. Since you have already tried increasing the holding current with a 2.2 kΩ resistor without success, I would decrease this resistance further to 1 kΩ to perhaps as low as 100 Ω to see if you can get the SCR to latch. But first I would substitute a brand new SCR and test for latching with just an LED and a current-limiting resistor in the anode circuit... your combination #3 in your post #11.
The first two scenarios, combinations #1 and #2, don't test the SCR. They simply verify the ptm (push-to-make?) switch in parallel with the SCR will light the LED or sound the buzzer.
In the other two scenarios, combinations #3 and #4, the SCR appears to be behaving as a MOSFET or BJT would. I am not aware of a failure mechanism that would cause an SCR to behave this way, but who knows? Try a new SCR. You are also over-driving the gate terminal when the beam is blocked, essentially connecting the gate to 9 V DC through a 10 kΩ resistor, applying roughly 900 μA to a gate rated to turn on at 200 μA with a forward voltage drop of 0.8 V I doubt this would be responsible for damaging the SCR, but again, who knows? Try a new SCR. Maybe drive the gate from a voltage divider to decrease the amount of forward voltage applied to turn the gate on. I am shooting in the dark here. Try a new SCR.