I'm not sure if this applies to this situation, but pigments, including many fluorescent "dyes", are highly insoluble in water (around 1 part per trillion). They are particulate in aqueous solution and require some turbulance to remain in the water column. The dyes in question could be tested for solubility by centrifugation of the diluted sample (something around 1000g x 5 minutes). Edit: Or just dilute in coolant in a tall jar and let it sit a few days to see if any flourescenece begins to band at the bottom.
As far as the general loss of flourescence some people have seen, this could also simply be photobleaching (loss of activity of the flourophore). A stabilizing agent such as n-propyl gallate may be required.
http://www.sigmaaldrich.com/cgi-bin/...ProdNo=W294705
n-propyl gallate is what I have used in fluorescence microscopy, but there are many anti-fading reagents, some of which may be better suited to home use. I suspect the original sample of dye may contain an anti-fading reagent, but the dilution when added to cooling systems may bring the concentration below a feasible level.
PS: I hope I'm not restating well-known data. I've only recently started reading this forum in hopes of quieting my HTPC.