I may have misunderstood your intentions, but based on your second sentence, the first thing I would do in your shoes would be to get my head around the idea of serial dilutions. Making a PPB solution in a single step (unless you're willing to make enormous quantities) is not the way to go and will result either in poor accuracy and repeatability or ridiculous volumes of analyte solution.
What am I on about? Let's imagine (like many labs) you have a standard 4dp analytical balance with 0.1mg readability, in that case, the minimum weight you can measure (and obtain less than 0.1% error) is 82mg and that's assuming absolutely ideal conditions in a ground floor weighing room with the balance set up perfectly on a weighing table! If we round this up to 100mg, then to make a 20PPB solution, you'd need to dissolve 100mg of your compound in nigh on 5000 litres of solvent - hardly practical!
Instead what we could do is make up a stock solution of say 100ug/ml (by dissolving 100mg of our analyte in 1L of solvent). By withdrawing 10ml of this and dissolving in 990ml of solvent we have a diluted stock solution which is at 1000ng/ml from which we can make up our calibration standards by further (1 in 2, 1 in 5 etc.) dilution. Doing it this way saves us 4998 litres of solvent! 2L of solvent is still a lot, so we could bring this down by making up a smaller volume of more concentrated initial stock solution (say 100ml) and adding additional dilution steps, remembering of course that each step contributes to the overall error.
What difference does using ethyl acetate (rather than water) as your solvent make to these calculations? Well that's up to you... Parts per notation is really annoying as it is often unclear as to what the scientist is on about as parts per notation can be weight:weight, volume:volume mole:mole and plenty of analytical chemists use weight:volume (in essence ignoring the difference in density between water and other solvents), so you're going to have to figure out which is appropriate for you and make the necessary mass to volume calculations where necessary! This is why I personally prefer to use weight volume (i.e. ng/ml) nomenclature in my work.
Out of curiosity what are you running and are you using SIM or Scan?