Crystallization

Ancient silver coins are sometimes extremely brittle. This brittleness is found in coins which are corroded as well as in coins which show no sign of external corrosion. Unfortunately, almost any ancient silver coin might suffer from undetectable crystallization. Fortunately, the vast majority of ancient silver coins are not crystallized or especially fragile. When a crystallized coin is dropped on a hard surface or handled roughly, it may break. While the exterior of the coin appears to be normal silver, the interior is white and does not appear metallic.

Although crystallization is the popular term used to describe this fragile condition, the term is a misnomer.  Granularization or embrittlement are perhaps better terms (but not customary).  Embrittlement of silver has been studied for a long time. It is the cause of some concern in museums holding archaeological silver and in archaeology itself. Embrittlement seems to be linked to inter-crystalline corrosion (see Ravitch, Lehmann, Organ, and Werner). Inter-crystalline corrosion can be exacerbated by the alloying elements present in the silver. Copper and lead are commonly encountered in brittle silver (Lehmann, Bhowmik, Toda, Thompson) but bismuth has also been detected (Rematullah). Discontinuous preservation of copper at the edges of the silver grains can also lead to embrittlement. Lead can make silver brittle even without corrosion (Toda).

References:

Thompson|, F. & A. Chatterjee. "The age-embrittlement of silver coins" in Studies in Conservation Vol. 1 No. 3, 1954, pp. 115-126.

Ancient silver objects are often found to be in extremely brittle condition. This brittleness can be observed in objects which are corroded as well as on those which show little or no sign of external corrosion. The brittleness of apparently uncorroded silver objects represents an interesting metallographic problem since the silver must have been ductile at the time the object was manufactured. The embrittlement implies a drastic change in the metallographic structure. The research laboratories of the Musie d'art et d'histoire in Geneva and the Metropolitan Museum of Art in New York City are collaborating on a project to study changes in the microstructure of silver-rich silver-copper alloys from long exposure to ambient temperatures. After preliminary work on a scanning electron microscope, microhardness tests, and examination of metallurgical cross sections of silver samples dating from 500 B.C. to A.D. 1000, research is now centered on copper precipitation from the silver-copper alloy. The binary-phase diagram for the silver-copper system shows that up to 8% of the copper will remain in solution at the eutectic temperature but that the silver can hold only one-tenth percent copper at room temperature. The precipitation of copper from the super-saturated solid solution occurs rapidly at temperatures between 150 and 450C, but very slowly below 100C. C. S. Smith suggested that a small but visible amount of copper could precipitate even at room temperature over many centuries. This type of precipitation is called "discontinuous" or "cellular." Precipitaion behavior of modern silver-copper alloys is discussed and compared with the observed microstructures of ancient silver samples. The possibilities and limitations of a new method of authentication by measuring the interlamellar distance between the copper-rich precipitates is treated.

Kallfass, M., P. Juergen & J. Hermann. "Investigations on the embrittlement of an antique Roman silver bowl" in Prakt. Metallogr (0032-678X) Vol. 22 No. 7, 1985, pp. 317-323.