Trace element (TE) distributions for As, Cd, Cr, Mn, Ni and Pb have been studied around the recovery boiler of a softwood kraft mill [1].The fate of these elements is of interest in closed cycle mills, where their toxicity could be an issue. An extensive sampling program was conducted in the mill, which had a production rate of 1700 adt/d, and used TCF bleaching. The black liquor was fired at 80%, and the capacity of the recovery boiler (RB) was 3000 t BLS/d. The lime kiln had a capacity of 500 t CaO/d; the lime was dried from 80% to 100% dry solids before burning. About 2% of the CaO used in the causticizing was added as make-up lime.

The wood chips were the primary source of all the trace elements (more than 80%), with the exception of Ni. Ni entered the mill with the heavy oil used in the lime kiln, the CaO make-up and from the corrosion of equipment. Mn was the dominant element in all sampled streams (with a total input of about 200 kg/d), with the exception of the flue gas. Although Mn is an NPE, rather than a TE, it has been included in this study due to emission limits on its toxicity. Cr and Ni concentrations were 2 to 3 orders of magnitude lower that that of Mn. As, Cd and Pb concentrations were one order of magnitude lower than that of Cr and Ni.

For the non-volatile trace elements (Cr, Mn and Ni), the amount which circulated with the fly ash was 3% of the total input to the RB. For the volatile elements (As, Cd and Pb), the amount varied between 20-50%. The amount of TE in the recycled ESP dust varied considerably. For As, Cd and Pb, about 30-40% of these elements volatilize and go upwards in the RB. For Cr, Mn and Ni, the amount in the ESP dust is 2-5%. The results show that the TEs are distributed between the smelt and the upper furnace. The volatile Cd, Pb and As are enriched in the dust, while Cr, Mn and Ni are not. Of the total amount of trace elements entering the RB, 4% ended up in the flue gas. Of that 4%, 7% went to the stack, while the rest was circulated as fly ash, which was added to the black liquor. The flue gas emission contained 0.4% of the total TEs input. As, Cd , Ni and Pb were detected in the flue gas. The total emission was below the permitted emission levels, both in the gas phase and in the particulate matter.

TEs enter the causticizing cycle with the smelt and the burned CaO, and leave with the lime mud and the dregs. The GL and WL filters have the strongest effect on the reduction of Cd, Cr, Ni and Mn, while As and Pb reduction was quite small. About 15% of the TEs end up in the WL. The overall highest build-up in the recovery cycle, excluding the lime cycle, was for As and Ni. Enrichment factors of up to 2-5x were observed for As and Ni in the liquors. The highest concentration of TEs were found in the lime cycle, especially for Cr, Mn, Ni and Pb. This was because no major output from the lime cycle was available. Of the total TE input, 56% exited with the dregs and 42% with the pulp. Cr, Mn and Ni mainly exited with the dregs and pulp, while As, Cd and Pb mostly left the mill with the fly ash and the pulp.

Although the mass balances were variable throughout the mill, the values for Cd, Pb, Ni and Mn were relatively good, whereas the As and Cr balances were poor. Factors affecting the balances were the accuracies of the sampling method, the chemical analysis and the flow measurements.

Under reducing conditions, all six elements can form sulphides. Although pure sulphides have very high melting points, as minor components in the Na carbonate-sulphide-chloride smelt bed they can be in liquid form at lower temperatures. A thermodynamic (TD) model developed for the smelt was used to investigate the nature of the TEs in the lower furnace. In this model, the smelt bed is described as a mixture of Na₂S, Na₂CO₃ and NaCl. The RB is divided into reducing and oxidizing zones. The TD data for the six elements were added to the model, as well as those of the important gaseous species. The calculations were done for 1000 ^oC. The results show that the non volatile Cr, Mn and Ni end up mainly in the smelt, whereas As, Cd and Pb are volatilized. The volatilization appeared to be mostly as monoatomic gas, and only to a lesser extent as gaseous hydroxides, sulphides or chlorides. The results from the predictive model were in good agreement with the experimental data. They indicated that a considerable part of the TEs found in the ESP dust originates in devolatilization from the smelt bed. Another source is the in-flight volatilization from the BL char particles. The conditions inside the char particles are strongly reducing, favouring the volatilization of As, Pb and Cd. The extent of the volatilization depends on the particle size, swelling and flow distribution in the RB.

The TEs react further in the oxidizing zone of the RB. The experimental data show that most of them are found in the ESP dust, due to condensation reactions. The TEs can form sulphate or chloride, depending on the concentration of SO₂ and Cl in the upper furnace. Temperature, local SO₂/Cl conditions, and kinetic studies must be considered in the development of a predictive model. The chemistry of the individual elements, as well as their reactions with Na, S, K and K must be included. The presence of NPEs such as Ca, Mg and P could also affect the TEs reactions.

1.Backman, R. et al, “Trace Element Distribution in and Around the Recovery Boiler”, International Chem. Rec. Conf., 2004