Best guess:
The problem with liquids is probably the high IR absorptivity of water. So aqueous solutions will probably give a low signal. Try loading the small sample holder with a little (just barely covering the middle indentation) of the solution you are trying to analyze and see what the sensor makes of it? (You’ll have to clean out the holder carefully between samples, so it’s going to be a bit inconvenient, but it would be a start…)
Start with analyzing the same liquid a couple of times and see whether the results are consistent (whether the curves have a similar shape) before you analyze different liquids.
Oh, and let us know what you find out!
/Otto