Since you plan to apply carbon dating, I assume you are dealing with rather unconsolidated sediments that you can wash and sieve with distilled (or tap) water. If the washed residue is not loaded with organic debris - foram residues rarely are - the specimens can be easily picked. One simple rule: avoid chemicals in the processing when there is no real need for using them. For the disintegration of shales and marls we generally use a simple soda solution, but for unconsolidated clays tap water is mostly sufficient.
Never use diluted hydrogenperoxide (H2O2) for processing organic-rich and pyrite-rich sediments as sulfuric acid will be formed. This will partially dissolve and probably to some extent leach the tests. See also the paper by Kennedy & Coe (2014 - Journal of Micropalaeontology) on this problem.
May I suggest you to contact my erstwhile student Dr. Amit K. Ghosh of Birbal Sahni Institute of Palaeobotany, Lucknow who has been working in this field for many years. Amit may guide you in right way. You may contact him in his e.mail ([email protected]/ [email protected]) citing my reference.
Thank you all for your suggestions. I read paper by Kennedy & Coe (2014 - Journal of Micropalaeontology) and by rosenthal,Chapter Washing and Sieving Techniques Used in Micropalaeontology
I found out that oxidative cleaning can change the Mg/Ca ratio.
If you want to separate for C14 dating, isotopes and elemental ratios, then the most reliable method is to pick individual clean and intact specimen. Since, broken shells may give you biased results due to contamination.
You may use Colgan or Sodium hexameta phosphate (10% solution) for organic removal as well as H2O2 (5% solution) treatment can also be done for clay removal. But make sure not to leave sediment the chemicals in your sample for longer time, it may harm the foraminiferal shells. For the detailed foraminiferal processing, you may refer my Palaeo3 paper (2017).
In the Mg/Ca methodology, reductive cleaning may change the shell chemistry. You may also refer to Yu et al., 2007 paper for that.