Hi, Sarah.

You did not inform about what kind of spectrophotometer was used, single- or dual-channel. I assume, that you record DMSO relatively air (black line) on dual-channel device. Reflectance from cell responds for level 0.04-0.05A for all 350-800nm range.For DMSO cut-off point is 265-270 nm, it's intrinsic absorption of DMSO. Below this wavelength (200-265 nm) spectra are incorrect even though a compensating reference.

After that, you applied this spectrum as baseline for the next spectra. Your sample (red line) was DMSO solution without absorption in visible range, so you get zero at all range with noize below cut-off point.

Use compensation cell with DMSO in second channel to minimize impact of the solvent, increase concentration of your drug (2-3 times) and analize region from 270-350nm (although it depends on the drug). 

I assume that the red line is for the solvent (DMSO) and black line when you added your drug in DMSO and you have performed baseline correction with DMSO.

- As soon as you add the drug, the baseline has shifted. There is increased absorbance at all wavelengths. This could be due to increased scattering caused by the insolubility of the drug in DMSO. The drug may be in the form of an aggregate.

- Check the solubility of the drug in DMSO. If found insoluble change the solvent.

- Below 300 nm you have DMSO absorption (UV cutoff wavelength 268nm), drug absorption and some scattering contribution.

- Is your drug colored? If yes, then it should absorb in the visible region. It you don't see the bands then either the compound is insoluble or the concentration is too low. Just increase the concentration of the drug. If no, then reduce the concentration of the drug, it seems high.

@Sarah Glass - On your question of why your sample (DMSO + drug) is showing lower absorbance than the blank (DMSO): as you would know, the expected result is the opposite.  I'm assuming you're using different cells for the blank and the sample [please disregard the following, if you're already using the same and only one cuvette for both experiments].

To answer this, I think the easiest way is to measure the blank with one cell, and then empty that cell, and use it again for a DMSO + drug solution.  Take care to use the same quartz micro-cuvette in the same orientation as you did for the blank.  It could be that small scratches or slight differences in cuvette wall thicknesses are enough to change the baselines in the amounts you observed.  I'm assuming that you cleaned the outer surface of the cuvette with the appropriate cleaning agent before measurement.  I would suggest that for the above "check" experiment, that you do not clean the surface, and take care not to touch the side that the UV-vis light will pass through.  If your baseline is the same in this new experiment, then you know that your baseline difference comes from the difference in the cells you used.  If your baseline still looks different, then something else is at play that I do not understand.

Of course, since you're not seeing any signal from the drug itself, the natural next step, whatever the outcome of the "check" experiment above, is to increase the drug concentration (like RMG stated clearly already in the first answer above).