In the soda roasting process, vanadium and tungsten are converted into the water-soluble compounds NaVO3 and Na2WO4 as shown in reaction 1 and 2. However, in the spent SCR catalysts, in addition to vanadium and tungsten, acting as catalysts, there is SiO2 contained in the glass fiber matrix which improves the mechanical properties of the catalyst, therefore Na2SiO3 is generated (reaction 3) during the roasting process. Na2SiO3 is leached alongside V and W inhibiting the complete leaching of the title metals. As a result, the amount of vanadium and tungsten leached decreases due to the presence of silicon soluble compounds. Therefore, an experiment was performed to reduce the leaching rate of SiO2 while reaching the optimum leaching conditions for vanadium and tungsten.
V2O5 + 2NaOH → 2NaVO3 + H2O (1)
WO3 + 2NaOH → Na2WO4 + H2O (2)
SiO2 + 2NaOH → 2Na2SiO3 + H2O (3)
NaOH phase and temperature effect.
In the soda roasting process, the reaction temperature is a very important factor in the conversion of V2O5 and WO3 to NaVO3 and Na2WO4 from the feedstock. An experiment was conducted to determine the effect of the phase and reaction temperature of NaOH used as a roasting agent in the amount of vanadium, tungsten and silicon present in the leaching solution. 15g of NaOH were added as a solid to the SCR spent catalyst sample and roasted at 773 ~ 1173 K for 2 hours while 15g NaOH were dissolved in a sufficient amount of water and subjected to the same roasting temperatures. After roasting, the residue leached in 100 mL of distilled water at 298 K for 3 hours. As shown in Fig. 1(a), the leaching efficiency of vanadium and tungsten augmented as the roasting temperature increased. In the case of dissolved NaOH, the leaching rate was very low at 773 K, but it increased significantly at 873 K, and it was found to reach the maximum at 973 K. On the other hand, when the roasting agent is solid NaOH, the reaction starts after melting resulting in a lower leaching efficiency of the title metals. Moreover, dissolved NaOH has superior mass transfer activity with vanadium and tungsten in the waste catalyst and the ionized sodium ions react more evenly with the waste catalyst particles than NaOH in the molten state. Fig 1 (a) shows that the leaching ability of vanadium and tungsten is maximized when the roasting agent is NaOH solution, while Fig 1 (b) shows that at the same roasting temperature (973K) the concentration of silicon leached in the solution is almost half in the case of NaOH solution used as a roasting agent.
Effect of sodium hydroxide concentration.
An experiment was conducted by adding the same mass of SCR spent catalyst with an identical volume of NaOH solution at different concentrations ranging from 10 to 50% to compare the leaching efficiency of vanadium and tungsten. The mixture of NaOH aqueous solution and feedstock was roasted in a muffle furnace for 2 hours at 973K, and after roasting; the sample was added to distilled water and leached at 278 K for 3 hours. The amount of vanadium and tungsten leached according to each experimental condition is shown in Fig. 2 (a). As the concentration of NaOH augmented, the leaching efficiency of vanadium and tungsten was greater. When the concentration was more than 10%, the leaching rate increased significantly, and when the concentration was 40%, the leaching efficiency of tungsten reached a maximum, while vanadium leaching increase was insignificant. Therefore 40% NaOH is considered as the optimum concentration.
Effects of roasting time.
To compare the leaching efficacy of vanadium and tungsten according to the roasting time, roasting was performed for 30 to 180 minutes, followed by leaching. As shown in Fig. 2 (b), leaching efficiency of vanadium increases as the roasting time increases, and the tungsten leaching reaches a maximum after 120 minutes. Due to the low concentration of vanadium in the original sample, 120min is established as the ideal roasting time.
Effect of pulp density for leaching.
Various experiments were used to examine how the pulp density affects the leaching efficiency of V and W. Leaching experiments were performed at 298 K and 300 rpm for a sufficient time to have complete leaching (120minutes) at 10-50% pulp density. As shown in Fig. 6, the pulp density does not have a great influence in the leaching efficiency; still, the highest leaching efficiency was at 30% pulp density. While, above 40%, the leaching efficiency decreases. It is expected that if the amount of spent catalyst used is higher, the concentration of vanadium and tungsten leached will be higher too, however, as noticed in Fig. 3 (a) the leaching efficiency does not improve with the increase of the pulp density. Thus, for an ideal mass transfer 30% pulp density was set as the optimum condition.
Effect of leaching time.
The effect of the leaching time from 10 to 180 minutes during the leaching process in the leaching efficiency of vanadium and tungsten was explored. After soda roasting at 973 K for 120 minutes, 100 mL of distilled water was used for leaching from 10 to 180 minutes at 298 K, 300 rpm, and 30% pulp density. As shown in Fig.3 (b), the leaching time does not contribute significantly to the leaching efficiency for vanadium and tungsten. Several researchers have studied the ideal conditions for roasting and leaching of spent SCR catalyst, however time leaching time has always been long (at least 1h) to maximize the leaching of vanadium and tungsten 19,20,23,24. For instance, Moon et al obtained an almost complete leaching of vanadium and tungsten in a 1h interval 25. Wu et al did an alkaline leaching without a preprocessing (soda roasting) of the catalyst which possesses the disadvantage of a large consumption of alkali in the process 19. The leaching system studied in this investigation reaches a maximum in 30 min, which makes not only savings in the economic part of the system but also in the efficiency point of view due to the reduced leaching time. It is deduced that vanadium and tungsten were converted into a form that facilitated leaching during the soda roasting process using NaOH aqueous solution.
Effect of temperature for leaching.
In the leaching process, the reaction temperature affects the mass transfer activity of the water-soluble target metal and acts as an important factor to determine the amount of metals leached. The effect of leaching temperature on leach efficacy of vanadium and tungsten was studied in a range between 298 to 343 K. As shown in Fig. 3 (c), the leaching efficiency for both title elements does not show a significant response to the temperature increase in the leaching process, which lead to the conclusion that to save resources 298K will be used as the optimal leaching time. Moreover, it is believed that as the reaction temperature increases, the activities of substances other than vanadium and tungsten in the spent catalyst increase, therefore a lower temperature (298K) is preferred to avoid an increase in the leaching of silicon or any other impurity.
Morphology of the samples at different metal recovery stages.
SEM-EDS images were taken for the raw spent SCR catalyst, after roasting and after leaching to compare the morphology of the residues at different stages. Figure 4 shows the SEM images at different magnifications for the original spent SCR catalyst, the spent SCR catalyst roasted with dissolved NaOH and the residue after leaching. Fig 4 (a1), (b1), (c1) show the spent catalyst sample before roasting, after roasting and after leaching at low magnification, respectively. It can be observed that in the case of the raw spent catalyst (a1) there is a clear presence of fiber glass rods (silica, alumina and calcium oxide) which are the main concern for this research due to the minimization of silica leaching as a primary research goal. As observed in the posterior EDS figures it is clear that the granules surrounding the glass fiber rods are composed mainly of titanium oxide (anatase form). In Fig 4 (b1) and (c1) the rods are not clear in shape at low magnification due to the crushing with mortar and pestle done to the roasted residue.
When the magnification is increased 1000 times, it can be observed that in Fig 4 (a2) and Fig 4 (b2) there is some agglomeration on the rod type structures while in Fig 4 (c2) the rod structures show a smooth texture. In the largest magnification Fig 4 (a3), (b3) and (c3), it can be observed in the original spent catalyst the absence of pores, which is characteristic of a deactivated catalyst. Moreover, in Fig 4 (b3) there is agglomeration over the rod-type structure, which can be related to the formation of sodium compounds on the fiberglass matrix. Finally in Fig 4 (c3), a smooth texture is observable characteristic of fiber glass surface, which leads to the deduction that vanadium and tungsten sodium compounds were leached from the surface leaving the fiber glass(composed of calcium, aluminum and silicon) in its original shape.
In Fig 5 corresponding to the EDS analysis of the raw spent SCR catalyst it can be appreciated that there is a clear silicon and calcium rod structure in the middle covered mainly in tungsten. Moreover, vanadium and titanium are spread along the sample, the presence of titanium around the rod type structures and spread is in accordance with the main component of SCR catalyst that is TiO2 (anatase phase)
Fig 6 shows the catalyst EDS analysis after roasting, an agglomeration of sodium particles can be observed covering the fiber glass rod which indicate the formation of sodium compounds such as sodium vanadate and tungstate. However, due to the large presence of sodium all over the rod-like structure, it can be deducted that Na2SiO3 might have been formed. Moreover, the percentage of sodium in the sample has increased in accordance to the reaction with the metals present.
It is observable in the EDS patterns in Fig 7 that after leaching the density of sodium particles present in the sample decreases, while the percentage of vanadium almost keeps constant (due to the small initial presence of this element) but tungsten decreases significatively. Moreover, the percentage of silicon compounds variates, which is in accordance to the results obtained in the roasting and leaching parameters analyzed previously. In addition, the morphology of the sample keeps the rod-type structure with small agglomeration of particles on top indicative of vanadium and tungsten not leached during the process.