Every try to remove the heat from the light affects the total light output and would change the spectrum.
i can rewin ~80% of the produced heat power only from the backside of the led platine.
That means 60% (0,8 * 75%) of the lamp power(Vf*If) is inside the heatexchanger.
40% is flux & heat @ the frontside of the chip(LES).
I was thinking about cooling the LES
but yes - this lens would reduce light flux may be more - than it can help to increase heat production. We can`t change the high heat resistance and (s)low heat conductivity of silicone layers, but we can reach much lower temps. of these phosphored silicones also at the backside, if we suck big quantities of heat into the water. The transparency of silicone get worse with higher temp and reaches a critical point at ~150°C.
Pic.: If i switch on to drive the 50W led @ max. current - my powermeter shows 50W going down to 47W-46W if the case temp. reach ~60°C without water.
When i drive it watercooled with a low watertemp like 15°C - my powermeter shows 53,5W
The drivers are cooled as well and work @ lower Tc (also with an expected - slightly better efficiency ?)