At 100mW, the small heatsink in a Dorcy would be enough. These diodes dissipate just under 0.6W of heat at that power, and reds even less.
The laser should still be left to cool off for a minute, between measurements.
Otherwise, the fixed focus should be set to a parallel beam. Not diverging. It doesn't matter to a thermopile, and it doesn't matter to a photo-cell. And if it did matter, it would be even more important to have a parallel beam, otherwise the distance would be too big a factor. A spreading beam could also be too big for optical sensors..