SUMMARY
In unprotected melting conditions (torch melting) the more or less oxidizing conditions of the flame can influence the final composition of the material. Oxidizing or reducing flame conditions can result in the reaction or non-reaction of certain elements with oxygen present in the air (e.g. Zinc). These elements, being also responsible for the colour tone of the material, can also, if they are in concentrations different from those expected, lead to substantial differences in surface colour
PROBLEM
The problem arises as a color variation on objects produced according to a well-standardized (therefore repeatable) cycle, after a torch fusion phase with hydrogen and oxygen in an uncontrolled atmosphere. Subsequently, the piece was cooled and then rolled.
The semi-finished products thus created showed a different color from those coming from vacuum casting. The color detected is brighter (see photo 1), more yellow and therefore lighter, and it was found that the problem was more evident during polishing. This last factor led to think that it was not a normal oxidation, as the color persisted even in the subsequent processing phases.
SOLUTION
A first consideration was therefore to discard the hypothesis that these were phenomena of zinc evaporation as the piece should have been "redder" and not "yellower". Furthermore, the results of chemical analyses carried out at an external laboratory showed that the overall composition of the samples with the altered colour could be considered identical to that of regular samples: in fact, the slight differences that were found cannot justify colours perceived as different by the human eye.
A surface microanalysis, carried out on both "altered" and "normal" colored samples, showed that on the surface of the sample with a light yellow color (altered) there was a very high content of Au and Ag while there were low values of Zn and especially of Cu compared to the surface of the one with standard color. In our opinion, this was enough to explain the difference in color found.
A microscopy at the SEM is reported, relating to the surfaces analyzed (photo 2).
The contents of the various elements, although comparable with the values determined by the analyses carried out externally, are not identical because different analysis techniques have been used.
A very important consideration is that, even with microanalysis, the differences in composition that have been found on the surfaces of the two pieces of different colors disappear if the altered color sample is sanded (removing a layer of thickness of around 100 um): the composition in the sanded area is the same as the 3N color object (and consequently also the color). See photos 3 and 4.
Our belief is that the color variation was due not to an evaporation phenomenon, but rather to surface oxidation during the melting phase with the torch: oxygen preferentially combines with Cu and Zn, which are the more electronegative elements, since Au and Ag are more noble, and the depth of oxidation is such that it could not be removed during the polishing phase.
To understand the phenomenon, consider what happens during "pickling" (when the piece is descaled in acid): preferential decomposition of Cu and Zn occurs on the surface, leaving the alloy rich in Au and Ag; the result is that a different composition is present only in the surface area. It is hypothesized that this layer may have a thickness of around 50–60 µm.
From this consideration, it is also clear why the external laboratory that performed the assay did not detect a substantial difference in composition between the two samples of different color.
It should nevertheless be noted that it is essential to control and maintain constant the conditions of the melting process (melting temperature, hydrogen-oxygen ratio, whether borax is used or, in any case, a protective system during melting), the acids used for the "pickling" phase (concentration), and the duration of exposure to the acid itself.