MANGANESE RECOVERING FROM ALKALINE SPENT BATTERIES BY AMMONIUM PEROXODISULFATE

The process of manganese removal from spent battery leaching solutions, with ammonium peroxodisulfate, prior to recovery of zinc by electrolysis is presented. The experiments were carried out according to a 2 3 full factorial design considering ammonium peroxodisulfate concentration, temperature and pH as factors investigated. The analysis of variance (ANOVA) was carried out on the precipitation yields of Mn and Zn after 30 min, 1h, 2h and 3h of reaction. Optimal conditions for obtaining Mn as MnO2 have been 20 % (NH4)2S2O8, 90 °C and pH 6. Data from XRF and AAS during the reaction at different time are presented to analyse the kinetic behaviour of the system. The MnO2 precipitated and solutions were characterized by XRF and XRD. The solid samples were kept at 800 °C for 1 h to produce chemical manganese dioxide (CMD) and were characterized by cyclic voltammetry for their electrochemical activity. The overall results denoted that chemical oxidation of manganese from spent batteries leaching solutions with ammonium peroxodisulfate is a suitable method for manganese recovery as MnO2 prior zinc recovery by electrolysis, also to produce of a suitable product (CMD), in other words, it could be used in a process for recycling spent batteries.


Introduction
The recycling of spent batteries (SB) is imposed by environmental regulations/legislation [1].Even though the costs of recycling SB could be quite high, a new perspective might arise as wastes to be considered secondary raw materials.In this case their price would be cheaper than ores.At present time, most of research work of recovery and recycling of SB is focused on the recovery and recycling of metal and oxides [2][3][4], due to their economic and environmental impact.The process of recycling helps to eliminate wastes and allows obtaining of new materials with specific properties.By volume and economic value, SB are undoubtedly wastes worth to be recycled for the obtained materials and the conservation of raw materials in the interest of the sustainable development.Two major approaches have been considered for SB recycling, pyrometallurgical and hydrometallurgical methods [2].The hydrometallurgical processes are characterised by different pretreatment stages (physical treatment processes of the spent batteries), followed by leaching and separation of different metals by electrolysis, liquid-liquid extraction, or selective precipitation.Usually, Zn is recovered from spent batteries leach solutions by electrolysis, but the excessive manganese concentration can cause problems.A major area of research is based on the MnO 2 production (using methods as soon chemical, sol gel, electrodeposition and hydrothermalmethod) and improvement of its electrochemical performance [5][6][7][8][9].MnO 2 is economic and commercial important with applications in different fields such as battery industry, water treatment plants, steel industry and chemicals [9].The studies presented in the literature described the obtaining of elemental metals, oxides, alloys, salts from spent batteries.Valuable materials can be reuse or synthetize from leaching solutions of SB.Therefore, manganese oxides were obtained from SB by electrochemical methods which are clean, versatile and effective remediation alternative.A mixture of Mn 2 O 3 and Mn 3 O 4 with 70 % grade of Mn was obtained from spent alkaline and Zn -C batteries by acid leaching and electrowinning [10].High current efficiency (90 %) has been obtained in the process of simultaneous recovery of Zn and Mn (as MnO 2 on lead anode) from alkaline SBs [11].
The aim of this study is to recover manganese as MnO 2 (CMD) from spent battery leaching liquors obtained from an industrial plant using ammonium peroxodisulfate, without altering the concentration of zinc in solutions that can be recovered by precipitation or electro winning.

Battery preatreatment
Spent alkaline and zinccarbon batteries were collected and dismantled in an industrial area located in North of Italy.The black powder (BP) was separated from plastic, paper, metallic scraps and nonmetallic scraps and milled by a ball mill.In this study the fine fraction of black powder, less than 2 mm.This powder was washed with water to remove chlorine, ammonium and potassium, in order to reduce the consumption of sulphuric acids in the subsequent leaching step.The acid leaching was performed with 2M H 2 SO 4 , at 60 °C and 3 hrs.Original and leached powder were analysed by Xray fluorescence.All these treatments were carried out in an industrial scale plant.The tests were performed in 200 mL flasks introduced in a thermostatically controlled water bath equipped with a mechanical stirrer (Dubnoff, ISCO).For each test, 25 mL of spent battery leaching solution was brought to desirable pH and then the volume of 40 % ammonium peroxodisulfate for the suitable concentration was added, according to the experimental plan.

Preparation of MnO
During each test, 1 mL of solution was withdrawn after 15min, 30 min, 1 h, 2 h and 3 h to measure the concentration of Mn and Zn by AAS(AAS, spectrometer SpectrAA 200, Varian), after centrifugation.Each sample was diluted 1:10 with solution of nitric acid (pH 2) to avoid precipitation of metals.The pH was measured at the end of each reaction.The filtered residual solid was washed with 150 mL of solution of nitric acid (pH 2), dried at 105 °C for 24 h and weighted by an analytical balance (Mettler PE 600).The Yates' algorithm was used to evaluate whether effects and interactions among the investigated factors are significant with the respect to the experimental error.The solution and powders obtain were analysed by X-ray fluorescence analyses (XRF, Spectro Xepos).
The solids obtained were analyzed by Xray diffraction analyse (XRD Philips Analytical).The commercial EMD from 3 different battery producers and 4 of the solid samples obtained by chemical oxidation with ammonium peroxodisulfate of the spent batteries leaching solution (the precipitates obtained for sample 4 and sample 8, before and after they were kept in the furnace at 800 °C) were characterised by cyclic voltammetry to assess their electrochemical activity.homogeneous black paste.The paste thus obtained was immersed into a homemade disk electrode (disk area of 0.071 cm 2 ), with glass walls and graphite bar as current collector.A DXC240 computer controlled Bipotentiostat was used for all electrochemical measurements.An electrochemical glass cell (V = 100 mL) equipped with a MnO 2 based working electrode, a saturated Ag/AgCl reference electrode and a platinum wire counter electrode were used.Cyclic voltammetry (CV) studies were performed between 1 V -(-0.2V) in 1 M Na 2 SO 4 electrolyte solution at different scan rates.In order to determine the contribution of graphite to the overall charge capacity of the electrode, a cyclic voltammetry, CV, experiment was performed on a blank electrode containing no MnO 2 .The charge storage capacity of this electrode was found to be negligible (≈ 10 −4 mA*h/cm 2 ).The capacity values (mAh/cm 2 ) were calculated from the area of the cathodic part of CV, corresponding to the different MnO 2 samples [7] and were reported to the electrode disk area.
The stability of the electrode material was achieved by multipulse technique for loading and unloading.Galvanostatic chargedischarge experiments were performed in a similar setup cell as described above and consisted in maintaining the charge/discharge current at 0.14 mA/cm 2 for 3 hours.The curves obtained were represented by the potential evolution over time.

Results and discussion
The focus of this study was the selective separation of manganese as MnO 2 from solutions, without loss of Zn.Table 2 reports the characterization of the initial solid and liquid samples obtained from leaching solution of spent batteries in the industrial plant.Table 3 and 4 shown the data of the XRF (semiquantitative composition of element after precipitation of Mn) for the solid samples (precipitates) obtained in different treatments and for the final solutions after the reaction between the spent batteries leaching solutions and ammonium peroxodisulphate.
It can be noticed that most of the manganese content was found in the precipitates obtained after the filtration of solutions resulted in the reaction between the spent batteries leaching solutions and ammonium peroxodisulphate.Most of Zn content remained in solutions.This confirms the fact that Mn was selective precipitated from the solution as MnO 2 , based on the chemical reaction (1).Liquid samples collected at different time (15 min, 30 min, 1 h, 2 h and 3 h), during the precipitation process, were also analysed by AAS.The best results were obtained in the treatments 4 and 8, where the Mn precipitation was the highest (~ 90 %) and Zn precipitation was the lowest (25 % for sample 4 and 5 % for sample 8).These results were obtained after 3 hours of reaction at 90 °C.The difference between these two samples was the pH; pH 4 for sample 4 and pH 6 for sample 8. Optimal conditions to obtain MnO 2 are: 20 % (NH 4 ) 2 S 2 O 8 , pH 6, 3 hours of reaction and 90 °C.The Anova results for Mn extraction yield after at 30 min, 1 h, 2 h and 3 h are presented in Fig.
1 [12].As it can be seen, after 30 min of reaction only temperature has a significant positive effect on Mn precipitation ( Fig. 1a ).After 1 h of reaction all three factors: ammonium  peroxodisulfate concentration, temperature and pH have a positive effect on Mn precipitation yield (Fig. 1b).Interaction between factor A and B is also positive: increasing the ammonium peroxodisulfate concentration and temperature the precipitation of manganese is higher with respect to the single effect.The other effects although significant from the statistical point of view are not relevant.Similar significant effects of factors and interaction between them can be noticed after 2 h of reaction (Fig. 1c).
There are positive effects on manganese precipitation of all three factors (A, B, C) and the interaction between ammonium peroxodisulfate concentration and temperature (AB) after 3 h of reaction (Fig. 1d).The positive effects of ammonium peroxodisulfate concentration, temperature and their interaction have been identified.The positive effect of pH could be related to the precipitation of Mn 2+ and the possible reaction of Mn 2+ with dissolved oxygen as indicated in reactions ( 2) and (3). ) Therefore, the Mn precipitation is positively influenced by increasing of pH.
To conclude, all the investigated factors have strong positive effects on the Mn precipitation process.At the same time Zn precipitation is not significant affected by any factor or their interaction, after 3 h of reaction (data not shown here).
The XRD performed on the most significant solid sample obtained during Mn precipitation is shown in Fig. 2. As it can be seen, the ammonia manganese oxide is the main component of the precipitates obtained.

Electrochemical characterization
Electrochemical activity of manganese oxides can be assessed by voltammetry which can provide additional information for material characterization from the shape, the number and the position of the different peak resembling the redox processes [9].The samples obtained from chemical oxidation, and used in these experiments, were the precipitates obtained for sample 4 (S1) and 8 (S2) and kept in the furnace at 800 °C (see Table 1).The electrochemical reversibility (ER) is one of the predominant factors influencing the power capabilities of electrochemical system.To evaluate ER of the sample the potential window of CV was changed in the following three domains: (a) from 1 to -0.2 V; (b) from 0.8 to 0 V; (c) from 0.6 to 0.2 V (Fig. 3).The symmetric i -E responses were observed indicating that the sample exhibited an electrochemical reversibility in the mentioned potential ranges and sweeprate.The effect of scan rate is presented in Fig. 4. As the scan rate increases, the CV profile deviates from the ideal capacitive behaviour.This is mainly because the redox reactions depend on the insertiondeinsertion of the alkali ion or protons from the electrolyte as explained by [13].At slower scan rates, the diffusion of ions from the electrolyte can gain access to almost all available pores on the electrode surface, leading to a complete insertion reaction.Capacity values (mAh/cm 2 ) have been evaluated using a similar method as Pagnanelli et al. [9].The specific values calculated from the first cycle for each material were 0.03 mAh /cm 2 for S1, 0.6 mAh /cm 2 for S2, showing a higher performance of the S2 material in comparison with the commercial ones.Dynamic trends of capacity were also calculated for the following cycles and reported in Fig. 6.
The cycling capability of the S2 was studied galvanostatically at a constant current of 0.14 mA/cm 2 and between 0 and 1 V.Typical chargedischarge curves for the S2 sample is shown in Fig. 7.The charge and discharge curves are basically symmetric and present a good stability in time.The factors investigated by ANOVA were ammonium peroxodisulfate concentration, temperature and pH.The statistical analysis shown that ammonium peroxodisulfate concentration and temperature had a great positive significant (> 95 %) influence on the product obtained (CMD).Highest Mn precipitation (aprox.90 %) was achieved after 3 h of reaction.Xray spectra of the solid samples with the highest amount of Mn have shown that the CMD composition is mainly ammonia manganese oxide.The conditions of precipitation seem to influence the properties of product (CMD).

Fig. 2
Fig. 2 XRD of the precipitate obtained during test number 4

Fig. 6 Fig. 7 4 Conclusion
Fig. 6 Capacity of EMD and CMDs calculated for the different voltammetric cycles

Table 1
Experimental conditions of 23full factorial design with central point and the final pH of solutions

Table 2
XRF data of some elements in BP < 2mm and leach liquor solutions

Table 3
Semiquantitative composition of solid samples, by XRF, after Mn precipitation

Table 4
Semiquantitative composition of leach liquor, by XRF, after Mn precipitation

Table 5
Mn precipitation yield versus time in the investigated experimental conditions

Table 6
Zn precipitation yield versus time in the investigated experimental conditions