> 117v ... 2 leds ... 2.4v .. is series. 2.2k 1 watt
117V-2.4V-2.4V is 112V.
112V across 2.2K is 51mA.
LEDs are usually 20mA max.
Also 112V, squared, divided by 2200 ohms is 5.7 Watts.
This "can" be made to work (with different values, maybe better plan).
But with 2.4V parts you really should be looking for a lower voltage. 6V heater supply is a lot closer than 117V.
> a LED isn't a diode ?
Yes, but the Reverse Breakdown is low, and not well specified.
Often it is speced as >5V. Now if the supply voltage is much greater, then the resistor will offer about the same current either way. 20mA on the 2.4V forward connection is as much heat as it will stand. If the reverse path gets 20mA (or 19mA), at double voltage (5V instead of 2.4V), it will run twice as hot as designed, and may melt.
We could use lower current. (That's also less heat in the resistor.) However the ">5V" is not a critical spec. Nobody cares. It could be 10V or 20V, depending how long they left the LEDs in the oven. If we flow 15mA at 20V, that's _6_ times the rated heat ini the LED.
Also half-wave LEDs flicker. If you move your head fast you see the dotted-line effect. This might even be good for a "skull". For a normal pilot-light, I find it annoying.
Now that we are on 6V power, the ~~9V peak is only 4V above the LED reverse breakdown. You should run the math, but it is pretty unlikely to flow big reverse current or heat.
For a 2-eye skull, I'd run the diodes opposite polarity. Not only balances the DC-from-AC, now the skull eye flicker alternates left-right....