Treatment of waste pickling baths from a batch hot-dip galvanizing plant

The Galva Power Group - UCL patented process for the treatment of waste pickling baths from a batch hot-dip galvanization plant was studied. The process provides a ferric chloride solution used to treat urban waste waters. The process involves four steps: Step 1 is the neutralization of the acid by means of a Fe(OH)3 sludge. Step 2 is a reduction of ferric iron to ferrous iron by means of iron scraps. The removal of heavy metals (Zn, Pb, Sn,...) as sulfides is the third step and step 4 is concerned with the oxidation of the FeCl2 solution into FeCl3. This thesis focuses on steps 2 and 3. As far as step 2 is concerned, it has been seen that a residual organic inhibitor in the pickling bath prevents this step from occurring efficiently. Two ways were studied to remove the inhibitor. Tests have compared eight different inhibitors. The first way is to oxidize and precipitate the inhibitor with ferric chloride. Some inhibitors were destabilized due to ferric iron. Once flocks were removed, the bath has been put in contact with iron scrap. The contact with scraps in a high stoichiometric excess yielded very high reduction efficiency: 91,8 % for Adacid 337, 79,3 % for Ferropas 7578, 89,9 % for Luttermix and 90,9 for PM Inhibitor. The ferric iron concentration passed from 14,2, 13,9, 15,4 and 14,1 g/l to 1,17, 2,88, 1,55 and 1,28 g/l respectively in thirty minutes without agitation. However Ferropas 7578, Luttermix and PM inhibitor are the preferred ones because they did not give rise to the formation of flocks in the presence of 10 g/l Fe3+. The second way is the adsorption of the inhibitor onto activated carbon. It has been found than more than 90% of Ferropas 7578 (6ml/l) is adsorbed with only 5 g/l of activated carbon. Its final carbon content is below 10 ppm. Adacid (6ml/l) was also removed with high efficiency as its concentration has been reduced from 640 to 56 ppm but onto 20 g/l of activated carbon. Special inhibitor concentration passed from 501,3 ppm to 200,7 by adsorption onto 5 g/l which represents 60 % efficiency. The economic appraisal shows that Ferropas 7578 would be the cheapest inhibitor to remove (4558 euros/yr), followed by Special Inhi. (8328 euros/yr). The third step of the process involves the sulfide precipitation of heavy metals selectively with respect to iron. Upscaling tests were run on a eight liters batch reactor and it was confirmed that the precipitant must be added very slowly in one hour with injection nozzles close to the impeller to minimize the parasitic reactions. The best result was obtained with a violent agitation: 750 rpm when the initial ferric iron concentration was much lower than 1 g/l. This leads to less than 16 % of iron in the cake. It is possible to use the redox potential to control the ferric iron concentration before the precipitation. Tests have shown that a redox potential below 100 mV is required to get less than 1 g/l Fe3+ (at pH 1 and 1,5). Furthermore, experiments have shown that pH 3 is beneficial as a zinc elimination percentage of 95 % was attained which corresponds to 0,17 g Zn2+/l in the solution. Lead and copper are also eliminated with high percentages, over 97,4 and 89 % respectively. Tin and nickel are partly removed by co-precipitation: over 50%and 33%respectively. Manganese is not eliminated. The kinetic investigations have revealed that zinc precipitation seems to follow a 3/2 order kinetics during the first half hour and a second order kinetics during the maturation. Finally, upscaling calculations have found that a larger (0,5 m3) but geometrically identical reactor would require 2,7 kW to stir the liquid in such conditions