The target concentration for our X-band EPR to obtain a decent signal to noise ratio is usually [400-2000 [(μmol spins)/L] (or ~200 μM sample). Minimum concentration could be 100 μM [(μmol spins)/L] of the desired signal. Higher concentrations can be beneficial, however, concentrations that are too high can potentially lead to artificial line broadening and enhanced relaxation from spin-spin interactions. This is not as important for typical protein samples as the surrounding protein usually acts as a buffer isolating the individual spins.
The last important aspect to consider is the solvent. Organic non polar solvents can typically be used with standard sized X-Band tubes at all temperatures. Unfortunately, samples in aqueous solvents present some difficulties at temperatures above 0° C. The large electric dipole of water interacts strongly with the E field of the standing microwave formed within the resonator. EPR microwave resonators are designed to limit the density of this field at the sample, but it is not possible to completely isolate the sample. The interaction between the two adversely affects the performance of the instrument and the ability to tune the resonator. Therefore, aqueous samples require the use of a thin sample holder called a flat cell or a Q-Band tube if the signal is strong enough. You can also consult with Heike Hofstetter about using capillary for aqueous solvents.