Reuse of manganese sulfate as raw material by recovery from pesticide's wastewater using nanofiltration and electro-electrodialysis: process simulation and analysis from actual data

Growing environmental concern about the presence of pesticides and other contaminants has implied greater pressure on environmental authorities to make industrial wastewater disposal more restrictive (Molina et al. 2014). Pesticide manufacturers use different wastewater treatments to reduce, eliminate or destroy the active ingredients of pesticides and by-products in wastewater. Treatments for this wastewater include a pretreatment step for removing additives and the main treatment to remove active ingredients. Some of these methods have clear disadvantages, such as low efficiency of pollutants removal, high energy consumption, high cost, and large sludge disposal, among others (Đorđević et al. 2017; Ayare & Gogate 2019).

The fungicide produced by the company whose industrial wastewater was chosen for this study is a product of the coordination of zinc ions with manganese ethylene-1,2-bis-dithiocarbamate polymer and is an important member of the ethylene bis-dithiocarbamates (EBDC). These are considered to have low acute toxicity, a broad spectrum of activity, low production cost and short environmental persistence, which is why they are the most common fungicide worldwide (López-Fernández et al. 2015, 2016).

Some of the required chemicals for the fungicide formulation are ethylene diamine, carbon disulfide, sodium hydroxide, manganese sulfate (22–24% by weight solution), zinc sulfate, water, and others. The main residual water streams which are the result of this process correspond to the mother liquor and the gas scrubbing purge, both with a high content of sodium sulfate (Na₂SO₄), manganese hydroxide (Mn(OH)₂), zinc hydroxide (Zn(OH)₂), and suspended solids. This residual stream is currently treated by conventional filtration and sedimentation.

Several factors, including process optimization, commitments to meet the Sustainable Development Goals, and mainly more restrictive limits of pollutant discharge, make it essential for the fungicide company to evaluate and implement alternatives for reusing or recycling of its wastes. Consequently, a new wastewater treatment plant (WWTP) is being currently constructed.

In recent years, different alternatives for the recovery and revaluation of manganese and zinc from different types of waste have been studied. For example, the selectively recovery of these metals from waste batteries and other electronic devices has been achieved with a low content of impurities, mainly through leaching processes plus precipitation (Sayilgan et al. 2010; Innocenzi & Veglio 2012; Biswas et al. 2016; Chen et al. 2017) or, recently, assisted by membrane technology (Innocenzi et al. 2018). And even the reuse of these metals as raw material for batteries has been suggested (Sitarz-Palczak et al. 2019). Furthermore, for the mining industry wastes and low-grade materials, selective manganese recovery and revaluation as raw material for high-quality sulfate manganese is reported (Ekmekyapar et al. 2012; Romero & Azañero 2014; Lian et al. 2018).

Other techniques for the separation of these metallic species are ion exchange, adsorption, oxidation/filtration, and electrolytic extraction. But their main drawbacks are the technical and commercial difficulties due to low recovery, low current efficiency, limited industrial implementation, the need for a rigorous purification system and the use of harmful additives, which affect the recovered metal quality. Accordingly, the precipitation is the simplest and the most conventional with regard to Mn and Zn recovery (Lu et al. 2014, 2016; Fattahi et al. 2016; Patil et al. 2016).

On the other hand, the studied wastewater is characterized by a high content of Na₂SO₄, recovery of which is considered uneconomic because of its low market price, where about a third of it is produced worldwide as waste or by-product in other industrial processes (Nowak et al. 2014; Pisarska et al. 2017a).

In this scenario, membrane technologies are effective and attractive alternatives in order to recovery and also revaluate Na₂SO₄, as well as Mn and Zn. Technologies such as nanofiltration (NF), which is relatively new and gaining attention in industry, for this particular research present clear advantages over other advanced water treatment systems, such as lower energy consumption and lower investment cost, while providing a high rejection of multivalent ions such as sulfates (for selective recovery), not to mention the high quality of treated wastewater (Bruggen et al. 2004; Bora & Dutta 2014; D'Costa 2015; Kamali et al. 2019). Some of the NF applications related to this investigation are the combination with operations such as crystallization, where 99% of sulfates removal has been achieved (Curcio et al. 2010); rejection ranges between 75% and 96% of sulfates and only 2–11% of monovalent ions in desalination brines (Pérez-González et al. 2015); the selective removal of sulfates from saline wastewater, thus opening up the opportunity to reuse both water and recovered species (Yan et al. 2016); and the selective separation of dyes and sulfates for the recycling of both species in the production of dyes (Liang et al. 2019).

Among other techniques that allow a selective separation of chemical species in different industrial effluents, electromembrane processes have been studied, such as membrane electrolysis, electro-electrodialysis (EED) or bipolar membrane electro-dialysis (BMED), which are widely used for the removal of ions or charged molecules in wastewater, obtaining an acid and a base from the salt. This could be considered as an advantageous situation to use these processes as integrated systems, where the acid and hydroxide obtained could be used in the synthesis process in which a residual solution of Na₂SO₄ is formed (Yang et al. 2014; Pisarska et al. 2017a, 2017b; Wu et al. 2019). Of these, the efficiency of the salt division increases with the concentration of residual salt, so it is convenient to have an adequate pretreatment prior to the electromembrane (Cichy et al. 2017). In general, these industrial wastewaters contain organic constituents, raw materials and synthesis products of secondary reactions, so the presence of impurities can limit their reconstitution. However, according to Pisarska and colleagues, the impurities in the waste solution did not affect the efficiency of the electro-membrane process, although it did imply an increase in the energy consumption (Pisarska et al. 2017a). Among other studies recently carried out on electromembrane technology for the treatment and reuse of wastewater there are: EED tested in fermentation processes for the recovery and recycling of organic acids with high recovery yields and low energy consumption (Handojo et al. 2019); phosphoric acid recovery in wastewater from an aniline production plant with EED, with low energy requirements, low waste generation and chemical consumption (Duan et al. 2018); separation of sodium formate and sodium thiosulfate in industrial effluents with EED (Aravind 2018); elimination of phenolic molecules in the form of phenoxide ions and simultaneously the desalination of wastewater by EED from oil processing (Wu et al. 2019); zinc sulfate recovery from electroplating industry waste by an improved electrodialysis system (Babilas & Dydo 2018); generation of sodium hydroxide from spent caustic soda through BMED and EED (Wei et al. 2013); and sodium naphthenate recovery into naphthenic acid and sodium alkalis from oil processing wastewater by BMED (Achoh et al. 2019).

There exist other advanced water treatment technologies for the separation of ions such as organic extraction, ion sieve adsorption, and biological ion channels. However, most of these methods are expensive and energy intensive. Therefore, new hybrid technologies have been recently developed in order to achieve low energy requirement (Ge et al. 2016; Uysal et al. 2017; Mao et al. 2019).

There are several commercial products allowing comprehensive design, simulation, and optimization of membrane or electromembrane unit operations, such as Aspen Plus^®. Nevertheless, none of them includes them directly. Therefore, custom unit operations have to be added (Bobák et al. 2014). Some studies recently carrying out simulations of these unit operations are: EED in circuit of the iodine–sulfur cycle (Misra et al. 2013); BMED to treat wastewater from a fluid catalytic cracking refinery (Tian et al. 2019); and NF–crystallization to treat high Na₂SO₄ concentration wastewater (Du et al. 2018).

In this paper, the main objective is the selective recovery of manganese, zinc and sulfates, to subsequently regenerate manganese and zinc sulfate as a solution from 22 to 24% of concentration by weight, which is used as raw material in the original manufacturing process of the fungicide. To achieve this, conventional and non-conventional water treatment technologies are studied, in particular NF and EED.

Physicochemical characteristics and flow rate of the wastewater are presented in Table 1 as provided by the company.

In order to recover and reuse manganese, zinc, and sulfates from wastewater, the following alternatives are proposed and simulated using Aspen Plus.

As shown in Figure 1, this alternative involves the following steps:

• I.

  Removal of suspended solids.

• II.

  Selective chemical precipitation with sodium hydroxide (50% by weight), through two stages of precipitation reaching different pH levels.

• III.

  Sludge (precipitated) treatment with sulfuric acid for the production of manganese and zinc sulfate.

The recovery of sulfates is not intended.

This alternative is based on the future WWTP planned by the company which incorporates a chemical oxidation and a biological treatment in order to achieve the desired water quality for disposing of it. Therefore, this alternative's effluent should be treated in the mentioned unit operations. Because of confidentiality of this information and the approach given to this investigation, chemical oxidation and biological treatment is not considered in the process simulations. Nevertheless, they are considered in the costs and water quality estimation.

• I.

  Removal of suspended solids.

• II.

  Chemical precipitation with sodium hydroxide (50% by weight), through one stage.

• III.

  Sludge (precipitated) treatment with sulfuric acid for the production of manganese.

The recovery of sulfates is not intended. Effluent should be treated as mentioned in alternative 1.

• I.

  Removal of suspended solids.

• II.

  Chemical precipitation with sodium hydroxide (50% by weight), by a single precipitation stage.

• III.

  NF (sulfate recovery).

• IV.

  Sludge (precipitated) treatment with NF concentrate.

• I.

  Removal of suspended solids.

• II.

  Chemical precipitation with sodium hydroxide (50% by weight), by a single precipitation stage.

• III.

  EED (Selective sulfate and sodium recovery).

• IV.

  Sludge treatment with EED anionic concentrate.

• I.

  Removal of suspended solids.

• II.

  NF (manganese and sulfate recovery).

• III.

  Crystallization.

• I.

  Removal of suspended solids.

• II.

  EED (selective sulfate and sodium plus manganese recovery).

• III.

  NF (manganese recovery).

• IV.

  Crystallization.

The integration of NF and electromembranes has been assessed in different studies in order to improve the efficiency and selectivity. In these studies NF occurs first and is intended to raise salt concentration for revaluating in electromembranes (Lee et al. 2006; Zhang et al. 2017), as a pretreatment for electromembranes (Geluwe et al. 2011; Liu et al. 2016; Cichy et al. 2017), or to mitigate NF concentration problems (Geraldes & Pinho 1995; Sakar et al. 2016). For this alternative, NF is proposed after EDD in order to separate divalent ions (principally Mn) from monovalent ions in the cationic concentration and to be ready for a final crystallization process (Havelka et al. 2019).

All the alternatives were modeled using Aspen Plus 7.3.

The asymmetric ELECNRTL thermodynamic model was used. The stream class worked was MIXCISLD, where the subclass CISOLID was made for the fungicide and other by-products of the formulation which were set as user-defined solids, where it is assumed that there is no decomposition nor solubilization of them. The rest of the species were modeled as conventional components with available thermophysical data; all of this is presented in Table 2. Hydrated precipitated components were not contemplated because drying processes were not considered.

Blocks used in each alternative are presented in Table 3. The filtering operations were based on the composition of each subclass of the stream, as well as the type of each component. Likewise, the liquid phase was considered as moisture components for subclass CISOLID.

In this work, an equilibrium model based on the Gibbs free energy minimization method was used for precipitation of Mn²⁺ and Zn²⁺, whose operating conditions were set at 1 bar without utilities. For precipitation processes, the use of sodium hydroxide was preferred instead of calcium hydroxide because it was found to cause a faster precipitation and avoided the precipitation of gypsum (Mocellin et al. 2015, 2017; Xanthopoulos et al. 2017). The same Aspen model (RGibbs) was used for the sludge treatment, where the operating conditions were set at 1 bar and 30 °C.

For the EED system, around 50% of salt rejections per stage was assumed in concordance with the membrane characteristics and water quality, reaching final removal efficiencies close to 95% based in similar solutions (Taylor et al. 2008).

Models for the calculation of concentrations in outflow streams for NF and EED were developed in Excel^®, and the ratios of the molar flows for each species were transferred to Aspen Plus. In both cases, SEP and SEP2 blocks were used as an artifice for the stream simulations, changes in temperature were considered negligible, and the drop pressure was estimated based on the membrane types and basic operational conditions.

Finally, HCO³⁻ ion was artificially added in the model to adjust the charge balance in mentioned operations; that is, to force the water dissociation process in an RStoic block. Also, the pH of each stream was adjusted by removing the mentioned species at the end, and water electrolysis was simulated in the same block for EED alternatives.

The estimation of capital expenditures (CAPEX) and operational expenditures (OPEX) was developed based on the methodology proposed by Turton et al. (2003). However, the costs of the research or development costs are not included due to the nature of the process and the company. For alternative 3, an evaporator was considered in order to treat the unused concentrate flow. By-products storage is not considered.

For alternative 1 and 2, the chemical oxidation and biological treatment costs were set up in concordance with the company's WWTP update. These processes are not required for the other alternatives because of the high resulting water quality. For the compressors in the biological treatment, costs were estimated by means of a design of activated sludge according to the wastewater quality based on the company's reports.

Table 4 shows the process units considered in the cost estimation for all alternatives, some design considerations and the estimated capacity or size. The prevailing construction material corresponds to carbon steel and only the pumps consider pressure factors. For EED, costs were projected from Taylor et al. (2008), and NF costs were estimated based on membrane area installed according to GE Osmonics.

In the case of chemical precipitation, the focus was on the selective recovery of Mn and Zn as a function of the pH (Mocellin et al. 2017). The amounts of NaOH solution (50% by weight) required were 1.4 kg/h (pH ≈ 8.9) and 42 kg/h (pH ≈ 10.1), corresponding to 200 m³/day of wastewater; nevertheless, the amounts of sludge rich in zinc were not significant. On the other hand, as discussed later in this paper, the amounts of manganese sulfate obtained in alternatives 1 and 2 were relatively equal with almost the same impurity levels. Consequently, the selective recovery of zinc was not taken into account for the other alternatives, as it was considered economically insignificant. Around 23 kg/h of H₂SO₄ and 11.08 kW in cooling service were required in order to obtain 154.97 kg/h of manganese sulfate solution of 23% by weight concentration and 6% of impurities.

These two alternatives were proposed as conventional alternatives, which are pretty common unit operations in industry and relatively easy to operate. However, these required additional treatment processes such as chemical and biological oxidation which increase significantly the costs and require a much bigger space. Additionally, impurities were reasonably high but could be improved by refining the filter processes.

For alternative 3, with the NF, it is expected to require fewer unit operations (in comparison to alternative 1 and 2) and also to increase the treated water quality with the premise of selectively separating SO₄²⁻ from Na⁺. Consequently, the SO₄²⁻ solution can be used in order to treat the precipitated manganese. However, for the recovery system values worked upon this paper for NF, the quality of the concentrate was relatively low, because of the Na⁺ concentration with regard to SO₄²⁻ concentration (Figure 7). Increasing the recovery in practice could result in the inoperability of the device, which is why future laboratory-scale studies are planned in order to determine the maximum operable value as well as the possible pretreatment requirements to avoid possible fouling problems.

From the sludge treatment with the concentrate at 75% of the recovery, solutions of manganese sulfate were obtained with a low concentration, and a high impurity content. It is for this reason that sulfuric acid was still required. The optimum amount of recirculated concentrate was 60.2 kg/h, with a requirement of 17.9 kg/h of sulfuric acid. With this, a solution of 22.61% in manganese sulfate content and 10.01% in impurities is obtained.

For the EED in alternative 4, a diluted solution of sulfuric acid (33.53% by weight with <0.01% impurities) and sodium hydroxide (20.38% with 0.03% impurity), oxygen and hydrogen were obtained as product streams. The anionic concentrate required for the sludge treatment was around 75 kg/h; a product solution of 24.04% by weight of manganese sulfate with an impurity content of less than 2% was obtained, so there is no sulfuric acid requirement for treatment sludge.

In this alternative, the sulfates recovery was sought for the sludge treatment. The recirculation of the sodium solution to the system is not contemplated; nevertheless, with the development of the project the viability of this in the treatment plant and in other plant processes is considered in future works. In addition, as with NF, the start-up of experimental tests for EED is scheduled as soon as the company's updated WWTP is completed.

Mn²⁺ is not considered to be precipitated in alternative 5. In this, NF is expected to selectively separate SO₄²⁻ and Mn²⁺ from Na⁺. Moreover, by increasing this ion's concentration, less energy is required in the proposed crystallization unit intended to recover the manganese sulfate.

Additionally, a heat exchanger was used in order to decrease operational cost in the crystallization unit by increasing the inlet temperature by the stream generated. The sludge is diluted with 95 kg/h of water, and the resulting concentration corresponds to 23.38% by weight with 3.68% of impurities. In the same way as the other alternatives, impurity levels could be significantly reduced by increasing the efficiency in the filters or by adding other cleaning units.

Similarly to alternative 4, for alternative 6 a diluted solution of sulfuric acid (29.0% by weight with <0.01% impurities) and sodium hydroxide (17.0% with 0.02% impurity), oxygen and hydrogen were obtained as product streams for the EED. The crystallizer and manganese sulfate solution preparation followed the same methodology as alternative 5, but in this cases a 22.35% concentration solution with 2.48% of impurities was obtained.

Tables 5 and 6 show the manganese sulfate product streams and the water quality obtained for the proposed alternatives, respectively.

The volumes and quality of the final solution were very similar for each alternative, except alternative 5 and 6 because of the remaining manganese in the liquid phase after the crystallization. The highest content of impurities was presented in the alternative 3, which is mainly due to the quality of the concentrate for the simulated recovery. In the case of alternative 1, 2 and 3, a high efficiency in filters and more cleaning unit operations are required.

Although for alternative 1 and 2 pure sulfuric acid was used, there was a higher content of impurities than in the alternative 4 and 6 (which operates with the recovered anionic concentrate in its entirety). This is due to the fact that in the alternative 1 and 2 a high efficient separation operation was not considered, so the final solution had a higher content of untreated wastewater. The highest manganese sulfate quality corresponds to alternative 4.

It is expected that the worst water quality corresponds to the alternative 1 and 2, because only precipitation operations are considered, so the chemical oxidation and biological treatment operations to be completed in the WWPT update by the company are still a requirement for wastewater disposal.

Although alternative 3 obtained an improvement in water quality in comparison with alternative 1 and 2, sulfuric acid is still a requirement. Alternative 5 did not assume the use of acid because of the crystallization, but the resulting water quality was lower than alternative 3 since precipitation could be considered as a pretreatment unit.

For the particular case of the alternative 4 and 6, the five effluents (treated water, concentrates, O₂ and H₂) can be considered as by-products with high commercial value depending on the pretreatment units. However, these alternatives have other implications such as high energy requirement.

Table 7 shows the CAPEX of each alternative in USD for the year 2019 (chemical engineering plant cost index = 619.2). Alternative 1 and 2 have the highest investment costs because of the operation unit dimensions. However, for the company, having the infrastructure for the processes of biological and oxidation treatment, these costs are significantly reduced for this specific case, but this was not taken in consideration.

Table 8 shows the OPEX associated with each alternative, the basis of calculation being 200 m³/day of industrial wastewater treated as a continuous operation corresponding to 360 days of operation per year.

For the biological treatment proposed by the company, a volume of 1,784 m³ and an oxygen requirement of 39.5 kg/h were obtained; the compressor requirement was 2.48 kW. On the other hand, the quantities of water for the needed equalization process (200 m³/day) are not taken into account because it corresponds to another process stream in the company, which is why the biological treatment is relatively viable for this specific case.

The reason why alternative 4 and 6 have a considerably high OPEX is mainly due to the energy consumption of the EED (∼360 kW), despite obtaining several streams of by-products and treated water of better quality. OPEX for alternative 1 and 2 were unexpectedly high for a conventional treatment and this was because of the high requirements for reagents, dimensions of construction and other OPEX parameters directly dependent on CAPEX such as maintenance, depreciation, and local taxes. Alternative 3 and 5 have the lowest CAPEX and OPEX because they required less reagents and unit operations were significantly smaller, but more studies for pretreatment of wastewater and post-treatment of manganese sulfate solution are required.

Table 9 shows the developed WQI for the manganese, zinc and sulfates removal for each alternative. If the quality of the effluent is taken into account by the developed WQI, it is observed that the differences between alternative 1 and 2 with respect to alternative 3–6 are even more remarkable. It means these unconventional alternatives tend to be much more feasible as long as environmental authorities demand higher water quality. The fact that unconventional alternatives have the highest degree of removal of the species studied is expected, with less associated costs.

The results can be summarized as follows.

• (1)

  From the results obtained in the simulation of processes, selective recovery of Zn with respect to Mn is possible. However, in this specific presented case, it is economically unfeasible.

• (2)

  Even though alternatives 1 and 2 are merely based on a conventional unit's operation, CAPEX and OPEX were high due to the wastewater quality to be treated and the quantities which require equipment with huge dimensions or capacities.

• (3)

  From the economic point of view, alternatives based on NF were the best ones (3 and 5), because of the simplicity or low requirement for equipment or unit operations. Despite this, more studies for pretreatment of wastewater and post-treatment manganese sulfate solution are required. For the specific case of alternative 3, it is viable but dependent on the recovery obtained in the NF, looking for a concentrate with better quality and in smaller quantity, which implies less waste to be disposed of. In a similar way alternative 4 is feasible, but further studies are required to achieve the expected NF recovery and other strategies intended to reduce the energy requirement in the crystallization unit and improve the recovered material.

• (4)

  The viability of alternative 4 and 6 is a function of the efficiency of the EED, as well as the cost of energy. They turned out to be the alternatives with the greatest potential for exploitation; nevertheless, their high OPEX, mainly due to energy consumption, could limit their implementation.

• (5)

  This type of industrial wastewater can be considered as a source of raw materials for the production of manganese sulfate under an in situ recycling scheme. However, more studies are required in order to know the final viability in different scenarios in industry in general. Which is why future experiments in laboratory scale are scheduled by the time the WWTP proposed by the company is completed.

This work was funded by MINCIENCIAS – Ministry of Science, Technology and Innovation of Colombia, and approved in the call 786-2017 for projects that aspire to obtain tax benefits for investment in CTeI, project: 6765-786-59834.