Paradoxical solubilities of cobalt complexes in the Scott presumptive test for cocaine

[BabcockHall]

"An elucidation of the reaction mechanisms involved has been published [56] . lt assigns the formula (alkaloid H)2 [Co(SCN)4] to the relatively water-insoluble blue complexes formed in neutral-to-basic solutions, and the formula [Co(alkaloid)2] (SCN)2 to the more water-soluble, brownish-red to pink complexes formed in acid-to-neutral solutions. Solubilities in water and chloroform are given for complexes formed with 29 alkaloids and nine metal thiocyanates." Schlesinger

Presumably (alkaloid H) is the protonated form of cocaine or other compounds.  Experimentally, there are blue crystals initially formed in the reaction, then the solution turns pink upon addition of HCl and that addition of chloroform leads to a blue organic layer in the Scott test.  What I don't understand is why the cationic species (the conjugate acid) is in the organic layer and the neutral species (conjugate base) is in the water layer.  Yet all of the reference material (one textbook and a couple of articles) have stated some variation of the passage above.  I am in the process of acquiring additional references, including (56) above.  Does anyone have any thoughts?

Edited January 21, 2020 by BabcockHall

[BabcockHall]

Oguri K et al. (1995) Japanese J. Toxic. Environ. Health 41(4):274-279 were not able to isolate a cocaine-containing complex by silica gel chromatography.  They suggested that the complex is too labile.  However, they determined the stoichiometry between cocaine and cobalt to be 2:1 in one complex.  They proposed a structure involving two bidentate cocaine ligands and two isothiocyanate (Figure 2 in their paper).  My only reservation about their proposed structure is that the isothiocyanate may coordinate through nitrogen, not sulfur.  I am still working on the acid-base aspects of this question.

Edited January 21, 2020 by BabcockHall