I have a solution of water and ethanol containing my hydrophobic target molecule that dissolves well in ethanol (5mg/ml) and poorly in water (0.00005 mg/ml). If these solvents were immiscible, I'd feel comfortable knowing how much of the molecule is in each phase (expecting to find 100,000 times more in the water than the ethanol). However, I am not confident that I can simply use these numbers to predict the amount of my target molecule that is soluble in this miscible mixture. If I spun down 10 mls of a 50:50 mix of water and ethanol containing 20 mg of my target molecule – where should I expect to find most of my target molecule?
ethanol and water are completely miscible, that mean any quantity of ethanol mix very well with water. You have a substance soluble in ethanol and present a low solubility in water. I suggest you, if possible to do experiments on solubility because the behaviour os solubility in a mixture of solvents can be even higher. see for instance
https://ac.els-cdn.com/S1877705816310153/1-s2.0-S1877705816310153-main.pdf?_tid=868fe17a-01e8-11e8-97bd-00000aab0f02&acdnat=1516896064_f7d59161b3ec6d18f2bbc88fb549425e
Thanks, Abdullah, Silas and Steven, for your replies.
Yes, an experimental approach is the best way forward but the substance I am working with is expensive so I'd like to avoid that route or at least narrow down the range to test in order to minimise the amount needed for the tests.
I've dug deeper into the problem and am convinced that Steven is right that “There is not a simple rule of thumb for solubility in miscible solvent mixtures.”
For others who might find this post and have the same question, I recommend “Review of the cosolvency models for predicting solubility of drugs in water-cosolvent mixtures” by Abolghasem Jouyban in J Pharm Pharmaceut Sci (www. cspsCanada.org) 11 (1): 32-58, 2008)
I decided, for my purpose, to use the log-linear model of Yalkowsky (explained in the review) to approximate the solubility – not as a substitute for the experimental approach, but as a place to start my experimental design (so as to minimise time and expense). This might not be the best model to use but it is the simplest. That said, if you are doing a lot of calculations, it can be cumbersome, so I have also attached an Excel file that others are free to use (at their own risk).
Best regards,
Jamie
- Badges
- Science topic