Non Silver Impregnation
Silver is the most common metal used for impregnation of tissue structures, but it is by no means the only one. Although it is relatively uncommon, gold can also be used for materials such as amyloid or the demonstration of astrocytes with Cajal's gold sublimate (gold and mercuric chloride).
Not everyone sees methods depositing metals other than silver or gold as being impregnations. To a degree, it depends on what constitutes an "impregnation" and what we define it to be. I interpret the term quite widely and would include almost any method in which a metal is in any way selectively attached to any tissue structure, then made visible in some fashion. This would include such methods as osmium tetroxide for triglycerides, copper for fatty acids and both the colloidal iron methods for acid mucopolysaccharides and Schmorl's ferricyanide reduction method for melanin and enterochromaffin. All these techniques deposit a metal then demonstrate the metal physically or chemically.
While metal ions are also used for mordant dyeing they are fundamentally either carriers for the dye or simply combine with it and alter its colour. The main focus is still on the dye, so they are seen as a variation in dyeing. In an impregnation it is the metal itself, or one of its insoluble salts, which forms the final visible component, i.e. the focus is on the metal.
Gold impregnations are not understood any more clearly than silver impregnations. They were devised empirically, and remain so. Usually, experience and familiarity with the techniques are required.
Osmication of lipids is quite well understood. Either the osmium tetroxide dissolves in the lipids or chemically reacts with them, depending on the type of lipid. During dehydration the osmium dissolved in the lipid is reduced to another osmium compound, possibly a lower oxide. During treatment with fat solvents such as xylene, the lipid is then dissolved away leaving a black osmium deposit where the lipid used to be. Those lipids which chemically react with the osmium tetroxide are fixed in the classical sense, and are unaffected by the processing chemicals. They remain in the processed tissue but are blackened by the reactions with osmium tetroxide which fixed them. Both processes are taking place at the same time, so in either case the presence of lipids is shown by black deposits of osmium.
The copper method for fatty acids is a chemical reaction. The copper forms a salt with the fatty acid and may then be demonstrated with methods for copper.
The colloidal iron technique for acid mucopolysaccharides is similar, the iron reacts chemically with the "acid" component of the acid mucopolysaccharides, and is then demonstrated with Perls' Prussian blue, a standard method for ferric iron. In Schmorl's method the solution is a mixture of potassium ferricyanide and ferric chloride, which do not react with each other. The reducing substance reduces the ferricyanide to ferrocyanide. This immediately reacts with the ferric chloride to form Prussian blue, which is insoluble and immediately precipitates at the site of formation, where the reducing agent was located.