SummaryFoaming has often been observed in experimental continuous liquid-fed, joule-heated, ceramic melters that produce simulated nuclear waste-containing borosilicate glass. Such foaming, which results in significantly decreased melter throughput rates, has been attributed to the release of 9 gas resulting from the decomposition of higher valency metal oxides to lower valency oxides. It has been claimed that this foaming can best be controlled by adding reducing agents to the melter feed. In particular, the use of formic acid has been proposed to 1) acidify the aqueous waste stream, thereby dissolving some of the hydrous oxides and improving feed rheology, and 2) act as a reducing agent to control foaming. This document, prepared at Pacific Northwest Laboratory, reviews studies of nuclear waste glass foaming and the effects of reducing agents, particularly formic acid, on such foaming. It reviews the redox chemistry of many of the potential oxidizing and reducing species in or added to the glass-melter feed with the objective of predicting some of the chemistry that occurs in the cold cap and the glass melt as they might affect the foaming phenomenon. The review is not exhaustive since the literature search was limited by the complexity of the subject, and unknown effects on the chemistry, including various unknown catalytic effects, will certainly be important in such complex mixtures.Based on this review, little is actually known about the causes of or best remedies for foaming in the nuclear waste glass melters. The actual temperature regime (cold cap(a) plus melt) in the melter that is most important to the actual melter foaming is unknown. Workers have made widely different implicit assumptions in this regard. No evidence was found that adequate testing of proposed foaming remedies with well designed control experiments has been conducted in the glass melters.There is significant evidence based on run experience that, in the absence of added reductants, acidic feeds do not foam nearly as badly as alkaline feeds. Although alkaline feeds promote formation of higher oxidation state metal species in the cold cap, it is far from clear that easier formation of higher oxidation state metals causes the apparent greater foaming with alkaline feeds. The problem is that no run data were found in which only acidity or alkalinity were the significant variables.Based on the results of this review, there is strong question whether the varying degree of melter foaming is caused by different extents of decomposition of higher oxidation state metal oxides releasing O,, and whether this foaming can, therefore, be best controlled by adding reductants. This concept, while somewhat reasonable, must be proven with carefully designed experiments having virtually identical feeds with and without reductants in adequate quantity. These experiments must also compare other factors such as feed acidity-alkalinity with other variables held constant. Available laboratory data on the effect of reductants do not, for one reason or another...