Phosphorus recovery from organic waste for its agronomic valorization: technical and economic evaluation

Research Article Received: 12 May 2021 Revised: 5 August 2021 Accepted article published: 18 September 2021 Published online in Wiley Online Library: 30 September 2021 (wileyonlinelibrary.com) DOI 10.1002/jctb.6926 Phosphorus recovery from organic waste for its agronomic valorization: technical and economic evaluation Marina Fernández-Delgado,a,b Esther del Amo-Mateos,a María T García-Cubero,a,b Mónica Cocaa,b and Susana Lucasa,b * Abstract BACKGROUND: The present work investigates the use of municipal mixed waste compost (MMWC) residue for phosphorus (P) recycling from a technical, economic, and environmental perspective. The study aims to obtain liquid extracts rich in P from MMWC with a low heavy metals content, suitable for their subsequent precipitation in the struvite form. The effect of inorganic (sulfuric/nitric) and organic (oxalic/citric) acids at different molarities and temperatures was studied using technical analysis. The preliminary economic analysis allows the costs of the extraction section for the optimal extraction strategies to be estimated and compared. RESULTS: Based on the multilevel factorial design, the P extraction optimization shows that sulfuric acid 0.5 M and 30 °C provided the highest P extraction yield, 94.2%. However, oxalic acid 0.1 M and 30 °C is the best choice to maximize the P extraction yield, while also minimizing the heavy metals concentration, reaching a P recovery of 2.5 g P/kg. The extraction kinetics for all tested acids were satisfactorily modelled using a second-order model (r 2 >0.99). The preliminary economic analysis, estimated for a P extraction pilot plant of 100 kg h −1 of MMWC, noted that sulfuric acid 0.5 M provided the lowest total investment cost (130 000 €) and a minimum sale price of theoretical struvite at 4.96 €/kg. CONCLUSION: MMWC is a promising raw material for P recycling as fertilizer due to an appreciable P concentration (7 g P/kg), which can be used together with other residues rich in P to obtain an economical and sustainable extraction process. © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). Keywords: municipal mixed waste compost; phosphorus recovery; kinetic study; organic waste recycling; circular economy INTRODUCTION Phosphorus (P) is one of the critical nutrients in the growth of organisms on Earth. It is mainly obtained from phosphate minerals to produce inorganic fertilizers, feed, and detergents. 1-3 Nowadays, demandisgrowing duetothe continuousincreaseinbothpopulation and industry. 4 So, resources could be exhausted in 50 to 100 years 5 and could be critical by 2050. 6,7 Under such circumstances, new phosphorous resources must be found from organic materials. The use of acid solvents and the subsequent precipitation in the form of struvite (MgNH 4 PO 4 ·6H 2 O) is one of the most common ways to recover P from organic waste. 3 The struvite can be used as a solid fertilizer in agriculture as it is easily spread and has a high P content and low concentrations of heavy metals. 8 In addition, struvite releases P slowly as crop growth occurs, so P is released more efficiently, depending on the needs of the plants, which demand P during growth. So, using chemical phosphoric fertilizers can be decreased, thus diminishing the environmental impact, as the slower dissolution of struvite avoids high P concentrations in soil particles or being released through land runoff. 9-11 The European Union (EU) produces around 250 Mt of municipal waste per year. 12 Mechanical-biological treatment (MBT) plants commonly manage this waste. In 2017 alone, there were about 570 MBT plants in the EU, whose treatment capacity was 55 Mt. 13 The organic fraction of municipal waste obtained by MBT is habitually composted and stabilized. The result of this treatment is municipal mixed waste compost (MMWC), which is low-quality compost. MMWC is restricted in agriculture by European regulations, 14 so MMWC is disposed of in landfills. However, landfill restrictions are increasingly stringent to comply with the principle of the circular economy. In this sense, according to European regulations, only 10% of the waste generated may be disposed of in landfills in 2035. 15 Therefore, it is necessary to develop alternatives that take advantage of the MMWC as a raw material. *Correspondence to: S Lucas, Department of Chemical Engineering and Environmental Technology, Institute of Sustainable Processes, University of Valladolid, Dr. Mergelina s/n, 47011 Valladolid, Spain. E-mail: [email protected] aDepartment of Chemical Engineering and Environmental Technology, University of Valladolid, Valladolid, Spain bInstitute of Sustainable Processes, Valladolid, Spain © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 167 The P composition of MMWC ranges from 6 to 19 g kg −1 . 16 Similarly, other studies of P extraction from various wastes reported similar values of initial P concentration in the raw material. For example, the municipal solid waste digestates used by Oliveira et al. 17 had a P content of 8.11 g kg −1 and more than 90% was extracted using nitric acid (HNO 3 ) extraction. Kalmykova & Karlfeldt Fedje 18 studied P recovery from municipal solid waste incineration (MSWI) flyash.They achieved a recovery rate of 70% using acidic leaching-precipitation from MSWI residue with an initial P concentration of 5.9 g kg −1 . To the best of our knowledge, no previous papers reported using MMWC as a raw material for P recovery. Moreover, there are no references related to the economic evaluation of P recovery from organic wastes. P extraction is mainly based on extraction with acid solvents because of its high efficiency and low cost, although metal/metalloids can also be extracted. 19 It is also necessary to find a suitable acid to extract P. Inorganic acids can extract all types of P and different alkali metals from the raw material. In contrast, organic acids have chelating effects, increasing the concentration of metals and metalloids in the liquid extracts. 20 There are previous studies that efficiently extract P using such organic wastes as sewage sludge and pig manure. For instance, Barca et al. 21 studied P extraction from sewage sludge by conventional extraction, comparing the efficiency of 1 M citric acid and 1 M hydrochloric acid; the maximum efficiencies of P extraction were around 65% using citric acid. Liang et al. 4 analyzed P extraction from sewage sludge ash, comparing the efficiency of sulfuric acid (H 2 SO 4 ) and oxalic acid at different molarities. The results were that 95% of the P contained in the raw material could be extracted. Ekpo et al. 22 recovered phosphorous from pig manure by hydrothermal treatment using water, 0.1 M sodium hydroxide, 0.1 M sulfuric acid, 0.1 M formic acid, and 0.1 M acetic acid as solvents. In this study, the use of sulfuric acid reached the highest extraction efficiencies, recovering 94% of the P in the raw material. The present work investigates the potential of P recycling from MMWC residue from a technical, economic, and environmental point of view. The aim is to obtain liquid extracts rich in P from MMWC with a low content of heavy metals, suitable for their subsequent precipitation in the form of struvite. The technical study aimed to optimize the acid extraction of P by analyzing the effect of temperature, the type of acid (inorganic and organic) and acid molarity, as well as the heavy metal concentration in the extracts. Sulfuric and nitric acid were used as inorganic acids and citric and oxalic acid as organic acids. Moreover, the extraction kinetics were determined and modelled to obtain a suitable mass transfer correlation for the subsequent scale-up of the extraction process. The environmental study analyzed the environmental impact of mineral and organic acids, considering eutrophication of aquatic areas and aquatic acidification IChemE indicators. From the economic point of view, a preliminary economic study was carried out to compare the optimal strategies of P extraction with inorganic and organic acids. For this purpose, the total costs of the extraction section of a pilot plant for P recovery from MMWC, along with the market value of struvite that could be theoretically produced, were estimated. MATERIALS AND METHODS Raw material The MMWC was kindly donated by Resíduos do Nordeste (Mirandela, Portugal), stored at −18 °C until use and dried in an oven at 70 °C before performing the experimental runs. Acid extraction Experimental procedure The general procedure for P extraction was as follows: solid:liquid (S:L) extraction was conducted using 250 mL sealed flasks in which 10 g MMWC and acid solvent were blended to achieve the set S:L ratio (10% w/v). The extraction was carried out in an orbital shaker (Incubator Shaker ES-60, Miulab, China) under the experimental conditions established. The extraction was carried out with different organic acids (oxalic and citric) and inorganic acids (sulfuric and nitric) at different molarities (0.1–0.5 M) in a range of temperatures between 30 and 60 °C at 200 rpm for 8 h. After that, the mixture was carefully recovered through filtration (Filter-lab 1300/80 0.45 μm, Filters AOIA S.A., Spain) and stored at 4 °C until analysis. Experiments were performed in duplicate. Determination of the kinetics of the P extraction P extraction kinetics were performed to determine the maximum concentration reached under equilibrium conditions (C e ) and the mass transfer coefficient of P recovery from MMWC (k). Experimental runs were carried out with the same conditions as those set out in the previous section under isothermal conditions (30 °C) to determine the kinetic parameters. The integrated pseudosecond-order equation proposed by Ho and McKay 23 (Eq. 1) was used tomodel the Pkinetics.In thisequation, C t is the P concentration inthe liquidat time t(g P/kgMMWC),C e istheP concentrationatequilibrium conditions (g P/kg MMWC), and k2is the mass transfer coefficient of pseudo-second-order (kg MMWC/(g P ·min)). t Ct =1 k2·C2 e +t Ce ð1Þ Determination of the mass transfer coefficient for S:L extraction Mass transfer has been widely studied in S:L systems, although principally for first-order coefficients. 24-26 As the mass transfer coefficient obtained in this study is of the second-order; the correlation has been adapted as follows (Eq. 2): k2=A2·Re0:5·Sc0:33 ð2Þ Re=ρL·dp·ω ⊘L ð3Þ Sc=⊘L ρL·DA ð4Þ where Re is the Reynolds number (Eq. 3), Sc is the Schmidt number (Eq. 4), and A2is an experimental constant (g P/(kg MS ·min)) that must be determined by regression from the experimental data. The parameters needed from each dimensionless number are the following: DAis the molar diffusivity (m 2 s −1 ), dpis the particle diameter (m), ρLis the liquid density (kg m −3 ), ⊘Lis the liquid viscosity (Pa s), and ωthe agitation velocity (1/s). To calculate Re and Sc,ρLand ⊘Lwere analyzed experimentally (Table 5), dpwas approximated from the particle size characterization (3.16 mm), ωwas an operation variable (200 rpm), and DAwas determined by the Stockes-Einstein equation (Eq. 5), where kBis the Boltzmann constant. DA=kB·T 6·π·⊘L·dp 2  ð5Þ www.soci.org M Fernández-Delgado et al. wileyonlinelibrary.com/jctb © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). J Chem Technol Biotechnol 2022; 97: 167–178 168 Analytical methods The analytical methods used to characterize both solid and liquid samples were previously described. 13 Solid samples The dried samples of MMWC were sieved to determine the particle size distribution. The organic matter content was determined by gravimetric analysis. The elemental composition (carbon and nitrogen contents) was determined by a LECO CHN-2000 analyzer. Macro and micronutrients and heavy metals were analyzed by ICP Optical Emission Spectrometry and ICP Mass Spectrometry. Characterization of the P fractions Total phosphorus (TP) was extracted from the MMWC by wet digestion with sulfuric acid and nitric acid (HNO 3 ). The P fractions were quantified by a sequential extraction protocol described by Barca et al. 21 and Lee & Kim. 27 As a result, four P fractions were quantified: (1) Weakly bound P extracted under mild alkaline conditions (0.5 M NaHCO 3 ) (2) Al and Fe bound P compounds obtained under strongly alkaline conditions (0.1 M NaOH) (3) Ca bound P compounds extracted under diluted acid conditions (1 M HCl) (4) Stable P residual compounds extracted under hot concentrated acid conditions (10 M HCl) Then, the P concentration in each extract was determined by spectrophotometry (HITACHI UV 2000 spectrophotometer, Hitachi Healthcare Americas, United States) at a wavelength of 720 nm according to the molybdenum blue method. 28 Liquid samples TOC and TN were determined by a TOC-V 5000 analyzer. P was characterized by spectrophotometry using the molybdenum blue method. 28 Macro and micronutrients, as well as heavy metals, were analyzed, as previously described in section 2.3.1. Physical properties of extracts The density of the liquid extracts was measured by gravimetry (30 °C). The viscosity was characterized by a CFRC-100 Cannon- Fenske Routine Viscometer, size 100 (Cannon Instrument Company, Barcelona, Spain). All determinations were made in duplicate. Economic evaluation A pilot plant with a production capacity of 100 kg MMWC/h and a humidity of 25% was considered for the economic analysis. To estimate the struvite's minimum sale price, only the extraction section was included in the economic analysis to compare the feasibility of the different acidic solvents tested. The heavy metals removal and struvite precipitation section was not considered because it would contribute similarly to the total investment cost for all scenarios considered. On the one hand, the heavy metal removal section would consist of membrane equipment. On the other hand, the struvite precipitation section would consist of a temperature-controlled stirred tank where Mg compounds and salts are added to adjust the pH of the extract for the optimal precipitation of struvite. On this basis, the theoretical struvite production was calculated. Its market value was estimated to verify whether the sale price is competitive and if the process could become economically viable. The Lang Factor method was applied to calculate the economic costs of the pilot plant as this method is commonly used to approximate the costs of process plants. The complete description of this method can be found in the literature. 29 First, the cost of the leading equipment (agitated extraction tank, storage tank, and centrifuge) was considered to compare the setup of each scenario. Next, the agitation power requirements were calculated using the scale-up criterion of the constant extraction mass transfer coefficient. Only these stages were considered because they are within the battery limits of this study. However, it must be considered that, if the heavy metal removal and struvite precipitation stages had been taken into account, the costs obtained, and therefore the sale price of the struvite, would be higher thanthatobtainedinthispreliminary economic study. Then, the Equipment Cost (PCE) was estimated with the CAPCOST software. Finally, the Total Investment Cost (TIC) was calculated using the solid–liquid factors proposed by the Lang Factor method from the PCE. Eqs. 6–8, to estimate these costs, are as follows: Physical Plant Cost PPCðÞ=PCE *3:15 ð6Þ Fixed Capital Cost FCCðÞ=PPC *1:40 ð7Þ TIC =FCC *1:05 ð8Þ The cost of the raw materials used was estimated from the references: MMWC: 12 €/t, 30 process water: 3.16 €/m 3 , 31 sulfuric acid 98% w/v: 77 €/t, nitric acid 50% w/v: 300 €/t, oxalic acid: 1.14 €/kg, and citric acid: 0.75 €/kg. 29 The following assumptions were necessary to estimate the plant's profits and the minimum sale price of struvite: All scenarios had a plant lifetime of 10 years. The annual production cost and the production cost per kg of struvite were estimated operating at 8000 h/y. Equipment amortization was estimated in 10 years. Finally, the minimum sale price could be calculated, considering an internal rate of return (IRR) of 10%, 13 and a net present value (NPV) of the plant of 0 €. Environmental analysis IChemE metrics 32 can be applied to compare any industrial process from an environmental perspective. In this case, the parameters selected were the eutrophication of aquatic areas (equivalent tonnes of phosphate (PO 43− ) per year) and aquatic acidification (equivalent tonnes of released H+ ions per year). Data analysis An ANOVA test was carried out with the Statgraphics Centurion XVIII version. The ANOVA test was used to conclude the statistical differences at a confidence level of 95% (P<0.05). RESULTS AND DISCUSSION Raw material composition The MMWC particle size was determined by sieving with the following size distribution: 0.02% greater than 20 mm; 0.37% between 10 and 20 mm; 2.02% between 5 and 10 mm; 47.61% between 1 and 5 mm; and the rest being under 1 mm. The MMWC was not milled before the experimental runs. Phosphorus recovery from organic waste: technical and economic evaluation www.soci.org J Chem Technol Biotechnol 2022; 97: 167–178 © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). wileyonlinelibrary.com/jctb 169 Table 1 shows the complete characterization of the MMWC used in this study. On average, the MMWC had the following composition (g kg −1 dry weight basis): 488.5 ±5.0 organic matter (OM), 309.0 ±4.9 total organic carbon (TOC), 22.6 ±0.5 total nitrogen (TN), 7.0 ±0.3 TP. Influence of the extraction conditions on P recovery P fractionation indicates that 65% of TP in the raw material is formed mainly by compounds with Ca bound P, while 17% of the P belongs to compounds with weak bonds, 13% is P relative to stable compounds, and 5% corresponded to P bound Fe and Al. The Ca and Fe concentrations verify that the majority fraction in the MMWC is Ca bound P. Thus, P extraction has to be performed under acidic conditions. 18,27 Barca et al. 21 recovered P from sewage sludge under these conditions, which showed 86.3% of Ca bound P and 3.5% of Fe and Al bound P. Therefore, due to its composition, the MMWC used in this study is a potential P feedstock, which can be recovered to obtain liquid extracts rich in P suitable to be precipitated in the form of struvite. Three critical factors were identified to extract P efficiently. First, the effects of temperature (30, 45, and 60 °C), type of extraction solvent (sulfuric, nitric as inorganic acids and oxalic, and citric as organic acids) and solvent concentration (0.1, 0.25, and 0.5 M) on the P recovery were analyzed. According to previous results (data not shown), these operating ranges and the literature data on P extraction from solid organic waste were selected. 17,21,33-36 In this study, the constant operating parameters were: S:L ratio 10% w/v, agitation speed of 200 rpm, and operation time of 8 h. These values were selected from previous studies and have not been optimized. 21,37,38 Temperature effect on the P extraction First, the temperature effect on the S:L extraction was analyzed. Table 2 summarizes the TP, TOC, and TN concentrations in the extracts using organic and inorganic acids as solvents. It should be noted that blanks with acids were also performed to consider the nitrogen and organic carbon concentration in the extraction solvents. An increase in temperature involves higher carbon and nitrogen concentrations in the extract due to the solubilization of the organic matter. 22,39 Regarding the TOC concentration, the temperature effect was more significant when oxalic acid was used. A seven-fold increase in the TOC concentration, from 0.6 g TOC/kg MS to 44.0 g TOC/kg MS, was observed when the temperature increased from 30 to 60 °C. However, the effect of temperature was less pronounced for inorganic acids and citric acid. In general, the TOC concentration increased significantly with temperature (by 1.05 and 1.60 times). This produced significant differences (P<0.05) in the range of temperatures studied. Similar results were obtained by Ekpo et al., 22 who reported P extraction from swine manure using a hydrothermal treatment. Increasing the temperature from 120 to 170 °C enhanced the TOC concentration by 1.16 times using sulfuric acid and 1.60 times using formic acid. However, a considerable amount of TOC in the liquid extracts can negatively affect P recovery in the subsequent P precipitation as struvite due to an inhibitory effect of the number of carboxyl groups. 40 Therefore, operating at low temperatures is highly recommended so the TOC concentration in the extracts can be kept low. On the other hand, the total nitrogen concentration in the extracts increased with temperature for all tested acids. In general, the TN concentration ranged from 4 g kg −1 MS to 6 g kg −1 MS, except when nitric acid was used, which reached 11 g kg −1 MS when the extraction was performed at the maximum temperature and acid concentration (60 °C and 0.5 M). The temperature increase in the analyzed range did not significantly improve (P>0.05) the TN concentration in the extracts. Ekpo et al. 41 also concluded that temperature did not significantly affect the nitrogen solubilization from chicken manure and digestate. The nitrogen solubilization could be due to the breakdown of proteins, which is mainly influenced by the extraction temperature and operation time, as pointed out by other authors, such as Sun et al. 39 Regarding P extraction, an opposite trend was observed for temperature: lower P concentrations were achieved at higher temperatures (Table 2). In our study, in the temperature range studied (30–60 °C) for all acids and molarities, P concentration could decrease by up to 50%. However, the results did not change significantly (P>0.05). These results agree with Ekpo et al., 22 who studied the P extraction from swine manure with such solvents as water, sulfuric acid, and organic acids. The possible reason for the decrease in P concentration could be that, at high temperatures, dissolved P can precipitate with cations of Ca, Mg, and Fe. 22,41 Oliver-Tomas et al. 34 studied P extraction from the organic fraction of municipal solid waste. This study analyzed the temperature effect on P extraction using such inorganic acids as sulfuric, nitric, and hydrochloric acid. The study concluded that high temperatures did not improve the P extraction and that the process can be operated at temperatures of 40 °C. Moreover, working with acids at high temperatures is not recommended in safety terms due to material corrosion and Table 1. Composition of the MMWC Parameters Units MMWC pH 7.5 Moisture a % 20.9 Ash content a % 40.5 Organic compounds b OM g kg −1 488.5 TOC g kg −1 309.0 Principal and secondary nutrients b TN g kg −1 22.5 Pgkg −1 7.0 Kgkg −1 14.5 Calcium (Ca) g kg −1 55.3 Magnesium (Mg) g kg −1 7.2 Sodium (Na) g kg −1 7.2 Micronutrients and heavy metals b Iron (Fe) g kg −1 12.3 Arsenic (As) mg kg −1 10.9 Cupper (Cu) mg kg −1 218 Manganese (Mn) mg kg −1 354 Zinc (Zn) mg kg −1 516 Lead (Pb) mg kg −1 119 Total Chromium (Cr) mg kg −1 161 Nickel (Ni) mg kg −1 52.7 Mercury (Hg) mg kg −1 0.5 Cadmium (Cd) mg kg −1 2.4 Data were shown as the mean value with less than 5% of relative error. a Total weight basis. b Dry weight basis. www.soci.org M Fernández-Delgado et al. wileyonlinelibrary.com/jctb © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). J Chem Technol Biotechnol 2022; 97: 167–178 170 maintenance. From a technical and economic point of view, working at low temperatures, such as 30 °C, could be the most suitable option. Influence of acid type and concentration The effects of the concentration and type of acid were analyzed. As can be seen, the increase in acid concentration enhanced TOC, TN, and P extraction (Table 2). This increment means that when the pH decreased, more chemical bonds were broken, and the concentration of solubilized compounds in the extracts increased. 42 Concerning the TOC concentration, the ANOVA test shows that the results obtained with inorganic and organic acids are statistically different (P<0.05). However, no significant differences were observed between sulfuric and nitric acids or between oxalic and citric acids, respectively. Regarding the molarity, the statistical analysis showed that the results were significantly different (P<0.05) when the acid concentration increased from 0.1 to 0.5 M, but not when the acid concentration increased from 0.1 to 0.25 M. As explained in the previous section, a high concentration of TOC in the liquid extracts can inhibit the subsequent P precipitation as struvite. 40 Therefore, using organic acids at higher molarities could be a disadvantage compared to inorganic acids. Analyzing the TN concentration, the ANOVA test showed that only nitric acid presented significant differences (P<0.05) in comparison to the other acids, whose results were similar. Regarding the acid molarity, the statistical results showed significant differences (P<0.05) when the acid concentration increased from 0.1 to 0.5 M, but not when the acid concentration increased from 0.1 to 0.25 M. The TOC and TN extraction trends were similar. These results can be explained because most TN extracted was organic nitrogen (organic nitrogen in samples ranged from 50% to 85% of the total nitrogen, data not shown). Similar results were achieved by Szögi et al., 36 who extracted P from pig manure solids using acids, obtaining extracts in which most TN was in the organic form. No significant differences were observed regarding P extraction when sulfuric acid and citric acid were used at concentrations of 0.25 and 0.5 M (P>0.05). However, P concentrations were significantly lower (P<0.05) when oxalic acid and nitric acid were used. If the amount of P extracted is compared, sulfuric acid and citric acid could extract 20–50% more P than oxalic and nitric acid when concentrations of 0.25–0.5 M were used. The extraction with sulfuric acid and citric acid is favored because the H + concentration was two times higher than when the extraction with nitric acid and oxalic acid was performed. 20,35 The P content increases with the concentration of acid in all cases. The ANOVA test also showed that an increase in acid concentration led to a significant increase in P concentration (P<0.05). It should be noted that, at low molarities (0.1 M), the use of organic acids favored the extraction of P. Concentrations of P were 2.5 and 3.1 g kg −1 MS when oxalic and citric acid were used at 30°C, higher than when sulfuric or nitric acids were employed (2.1 and 0.4 g kg −1 MS). This fact occurs because organic acids behave like strong acids at low molarities, since they can completely dissolve in an aqueous solution. At low molarities, organic acids can solubilize P more effectively than inorganic acids. However, with increasing molarity, organic acids cannot fully dissociate, and therefore P extraction is favored by inorganic acids. Under these conditions, inorganic acids can extract all the P in the raw material, obtaining better yields. 2,20,43,44 On the other hand, Kpombekou-A & Tabatabai 44 suggested that the presence of Ca and Mg ions can contribute to the decrease in the P solubility because these ions increase the pH during extraction. Koostra et al. 43 studied the effect of the pH of sulfuric acid and oxalic acid in P extraction from pig manure. This study found that at pH 4, oxalic acid obtained a P extraction yield of 70% compared to the 50% obtained by sulfuric acid. However, at pH 2, the sulfuric acid could extract practically all the P (~ 99%) from the raw material, exceeding the P yield obtained by oxalic acid. Sulfuric acid led to the best P extraction results (6.6 g P/kg MS at 0.5 M), corresponding to an extraction yield of 94.3%. At low molarities (0.1 M), citric acid showed the best P extraction results (3.1 g P/kg MS, 35.8% extraction yield). These results are in agreement with previous studies. Fang et al. 20 compared the effect of the mineral acids and organic acids on P recovery from incinerated sewage sludge ash. At low acid molarity (0.1 M), the amount Table 2. P, TOC, and TN concentrations in the extracts using inorganic and organic acids as solvents Type of acid Molarity (M) P(gkg −1 MS) TOC (g kg −1 MS) TN (g kg −1 MS) Temperature (°C) Temperature (°C) Temperature (°C) 30 45 60 30 45 60 30 45 60 Sulfuric acid 0.1 2.1 1.9 1.8 27.2 33.0 34.4 4.1 4.8 5.4 0.25 5.5 5.2 5.1 31.7 42.4 43.8 4.7 5.1 5.7 0.5 6.6 6.3 6.1 32.8 44.4 47.4 4.9 5.5 6.4 Nitric acid 0.1 0.4 0.3 0.2 25.1 31.2 30.4 3.9 6.3 6.8 0.25 1.9 1.6 1.2 27.5 29.6 30.4 5.8 7.1 8.5 0.5 5.1 5.0 4.9 40.6 40.8 41.3 9.7 10.5 11.2 Oxalic acid 0.1 2.5 1.7 1.6 0.6 18.7 44.0 4.3 4.3 4.6 0.25 4.3 4.3 3.7 3.5 26.3 45.0 4.5 4.6 5.4 0.5 5.3 5.0 4.8 5.1 35.1 51.2 4.9 5.0 5.8 Citric acid 0.1 3.1 3.2 2.9 7.5 19.7 31.5 4.0 4.3 4.6 0.25 5.0 4.8 4.8 20.6 25.1 32.0 4.1 4.5 4.9 0.5 5.9 5.3 5.2 26.7 27.8 32.1 4.4 4.6 5.0 Data were shown as the mean value with less than 5% of relative error. In the TOC and TN concentrations, only the concentration of these compounds extracted from MMWC have been taken into account. Operation conditions were: 10% w/v, 200 rpm and 8 h. Phosphorus recovery from organic waste: technical and economic evaluation www.soci.org J Chem Technol Biotechnol 2022; 97: 167–178 © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). wileyonlinelibrary.com/jctb 171 of P recovered was two times higher using organic acids (citric and oxalic acid) than mineral acids (sulfuric and nitric acid). However, at the higher acid molarity (0.5 M), the amount of P recovered was similar in all cases, with sulfuric acid obtaining the highest concentrations. Shiba & Ntuli 35 compared the P extraction from sewage sludge using inorganic acids, such as sulfuric and nitric acid. In this study, the percentage of P extracted was 34.3% for nitric acid and 41.6% for sulfuric acid, the latter being selected as a proper solvent. On the other hand, Darch et al. 37 compared citric acid, oxalic acid and maleic acid with a molarity of 0.02 mM for P extraction from tropical forest soils. In this study, citric acid led to the highest concentrations (up to 9 mg kg −1 ), followed by oxalic acid (up to 7 mg kg −1 ). Concentration of heavy metals in the extracts This work aims to obtain a liquid extract rich in P, which can be further precipitated in the form of struvite. Acids usually extract heavy metals from organic wastes. A high concentration of heavy metals in the extract can be a risk, as they can precipitate along with the P, and the final solid product could be contaminated. 45 Therefore, operation parameters have to be selected to maximize phosphorus recovery, while minimizing the concentration of heavy metals in the extracts. Table 3 shows the concentration of secondary nutrients and heavy metals extracted from MMWC under the different experimental conditions tested. Regarding the secondary nutrients, higher concentrations were reached at higher concentrations of acids. This is probably due to solubility changes at low pH, as Quist-Jensen et al. 45 and Wang et al. 46 verified. Mg, Fe, and Mn concentrations increased considerably when acid concentrations increased from 0.1 and 0.25 M (P<0.05). Otherwise, nutrients such as Ca in the liquid extract rich in P can precipitate in phosphate form and inhibit struvite precipitation. 47 Thus, the concentration of secondary nutrients has to be limited. As can be seen in Table 3, nitric and citric acids extracted significantly (P<0.05) more Ca than sulfuric and oxalic acids. The Ca concentration reached after 0.5 M nitric acid extraction was 3517 mg L −1 , while when 0.5 M citric acid was used, the Ca concentration reached 4373 mg L −1 . So, the use of nitric and citric acids for P extraction can show some disadvantages. On the other hand, oxalic acid has an advantage because it extracts fewer Ca + ions that could inhibit struvite precipitation. Ca + ions can react with the carboxyl group of the oxalic acid, forming calcium oxalate, which precipitates. 4 Regarding heavy metals, inorganic acids extracted more Cd than organic acids, extracting more than 90% of the Cd in the raw material. Sulfuric acid extracted significantly more Cr (5 mg L −1 ) compared to the other acids (<2mgL −1 ). Globally, sulfuric acid extracted 66.2 mg L −1 of heavy metals, followed by citric acid (47.9 mg L −1 ), oxalic acid (45.5 mg L −1 ), and nitric acid (24.9 mg L −1 ). Heavy metals present in the extracts could be removed through such technologies as membranes, 48 biosorbents, 49 or ion exchange resins. 50 This step would provide an extract liquid rich in P that could produce phosphate fertilizer. Optimal extraction conditions Experimental data previously described were statistically analyzed. A multilevel factorial design was applied to obtain the most suitable conditions for P extraction with the software Statgraphics Centurion. Two scenarios were considered. Scenario A only considered maximizing the P concentration without considering the content of nutrients and heavy metals in the liquid extracts. On the other hand, Scenario B considered maximizing the P concentration, while minimizing the content of nutrients and heavy metals so that P could be further precipitated as struvite. The results obtained from Scenario A were that the most suitable solvent is 0.5 M sulfuric acid (desirability of 99.0%), reaching a P extraction yield of 94.2%. On the other hand, the best conditions obtained from Scenario B were 0.1 M oxalic acid (desirability of 74.5%), reaching a P extraction yield of 35.8%. In any case, the final decision about the type of acid to be used in the extraction will be established based on the economic evaluation, as shown in section 3.4. Kinetics of P extraction In this section, the kinetics of extraction for all tested acids at 30 ° C were determined and modelled with a second-order kinetic model (Eq. 1). This kinetic study, which plays a significant role in the extraction process, is carried out to determine the kinetic extraction parameters for extracting P from MMWC. Moreover, an appropriate correlation for the mass transfer coefficient, required for scaling up the extraction process, is proposed. Figure 1 shows the experimental points and the predicted kinetic model for sulfuric acid (Fig. 1(A)), nitric acid (Fig. 1(B)), oxalic acid (Fig. 1(C)), and citric acid (Fig. 1(D)). The experimental P concentration data (C t ), shown in Fig. 1, have been adjusted to a second-order kinetic model (Eq. 1), as explained in section 2.2.2. After 8 h, extraction concentrations reached the equilibrium, which allows the parameters k 2 and C e to be determined. In Table 4, the parameters k 2 , predicted C e, and adjusted coefficient of determination (r 2 ) are shown. As can be seen, the model provided a perfect fit to the experimental data (r 2 >0.996). Also, the estimated equilibrium concentrations are similar to the experimental data obtained at 480 min (Fig. 1), with a maximum margin error of ±5%, confirming the validity of the proposed fitting model. On the other hand, Table 4 shows that the k 2 coefficient decreased when the acid molarity increased for the four acids studied. The decrease in k 2 varied between 64% and 85%, depending on the acid used. In addition, the parameter C e increased along with the molarity (Table 4). This was expected because, at high molarities, acids have a higher capacity for P extraction, as shown in section 3.1.2. The parameter C e varied between 0.40 ±0.01 g P/kg MMWC (nitric acid 0.1 M) and 6.61 g P/kg MMWC (sulfuric acid 0.5 M) (Table 4). Also, a mass transfer correlation was calculated to validate these results. At high acid molarity (0.5 M), the highest predicted equilibrium concentration was 6.61 ±0.01 g P/kg for sulfuric acid. Compared to a low acid concentration (0.1 M), citric acid provided the highest value (3.10 ±0.03 g P/kg). These results agree with the conclusions shown in section 3.1.2. Thus, the organic acids provided better P extraction results when low concentrations were used. In contrast, comparing the kinetics with 0.5 M for both inorganic acids, non-significant differences (P>0.05) in extracted P concentrations were observed due to the possible complete acid digestion of the raw material. Correlation of the mass transfer coefficient for the S:L extraction process A correlation of the mass transfer coefficient has been proposed, based on dimensionless numbers to justify the effect of molarity on the mass transfer coefficient. Table 5 shows the parameters needed to calculate the A2and the predicted mass transfer coefficient (k2). An increase in molarity from 0.1 to 0.5 M involved increased density and viscosity www.soci.org M Fernández-Delgado et al. wileyonlinelibrary.com/jctb © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). J Chem Technol Biotechnol 2022; 97: 167–178 172 Table 3. Nutrients and heavy metal concentrations in the extracts using inorganic and organic acids as solvents Compound Units Sulfuric acid Nitric acid Oxalic acid Citric acid Molarity (M) Molarity (M) Molarity (M) Molarity (M) 0.1 0.25 0.5 0.1 0.25 0.5 0.1 0.25 0.5 0.1 0.25 0.5 Nutrients Mg mg L −1 273 415 420 166 198 287 208 348 372 259 265 319 Na mg L −1 634 663 678 498 546 639 527 545 556 530 539 548 Ca mg L −1 532 589 725 1147 2757 3517 51 64 81 1938 4227 4373 Fe mg L −1 15.4 619 786 7.5 9.5 11.1 132 830 868 261 372 384 Mn mg L −1 16.1 27.9 28.3 4.2 8.4 17.6 8.7 23.3 25.2 19.9 26.7 27.9 Total nutrients mg L −1 1470.5 2313.9 2637.3 1822.7 3518.9 4471.7 926.7 1810.3 1903.1 3007.9 5429.7 5651.9 Heavy metals Cd μgL −1 16 201 217 6.4 21.6 224 12.0 14 15.1 28.3 83.5 103 Hg μgL −1 0.6 1.6 3.1 1.4 1.54 1.7 1.3 1.8 2.1 2.1 2.6 2.7 Ni mg L −1 0.5 2.5 2.6 0.3 1.9 2.4 0.6 1.4 2.0 0.6 0.9 0.9 Pb mg L −1 0.3 3.2 3.7 0.1 0.2 7.5 0.6 2.3 2.3 1.1 4.0 5.0 Cu mg L −1 0.2 3.4 6.5 1.2 1.3 4.7 2.4 5.0 8.9 1.6 2.1 2.1 Zn mg L −1 8.9 46.4 47.3 2.4 5.8 8.3 14 28.6 30 23.7 36.2 38.6 As μgL −1 366 730 923 155 667 712 440 602 675 446 516 525 Cr mg L −1 0.1 2.9 5.0 0.1 0.1 1.0 0.4 1.1 1.5 0.4 0.6 0.7 Total heavy metals mg L −1 10.4 59.3 66.2 4.2 10.0 24.9 18.5 39.0 45.5 27.8 44.3 47.9 Extraction conditions: 10% w/v. 200 rpm. 30 °C and 8 h. Data were shown as the mean value with less than 5% of relative error. Phosphorus recovery from organic waste: technical and economic evaluation www.soci.org J Chem Technol Biotechnol 2022; 97: 167–178 © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). wileyonlinelibrary.com/jctb 173 values from 0.4% to 3.2% and from 14% to 26%, respectively. On the other hand, molar diffusivity decreased, as it is inversely proportional to viscosity, according to Eq. 5. The decrease in Re along with molarity occurred because the physical properties of the liquid extract increased. However, regarding the tendency of Sc, the major contributing factor was DA. So the lower the value, the higher the Sc because, according to Eq. 4, these numbers are inversely proportional. In this way, using the dimensionless numbers and the experimental k2, the constant A2could be calculated. For the four acids, A2showed the same trend, decreasing as the molarity increased. Once the regression that allows A2to be calculated had been obtained, k2could be predicted (Table 5). The median absolute deviation of predicted and experimental k2values is lower than 15%. From the results, it can be concluded that the most influential factor in the mass transfer coefficient is the Re number. An increase in the acid molarity (lower pH) causes a decrease in the value of the Re number, obtaining lower transfer coefficient values (Table 4). Bong et al. 43 also reached a similar conclusion. Their study analyzed the effect of solids loading in agitated vessels for the solid–liquid mass transfer to obtain a mass transfer correlation for a range of solid loading (0.08 to 0.4 v/v). The conclusion was that the mass transfer coefficient (k2) and 0 1 2 3 4 5 6 7 8 0 60 120 180 240 300 360 420 480 C t (g/kg MMWC) Time (min) (A) Experimental 0.1 M Experimental 0.25 M Experimental 0.5 M Model 0.1 M Model 0.25 M Model 0.5 M 0 1 2 3 4 5 6 7 0 60 120 180 240 300 360 420 480 C t (g/kg MMWC) Time (min) (B) Experimental 0.1 M Experimental 0.25 M Experimental 0.5 M Model 0.1 M Model 0.25 M Model 0.5 M 0 1 2 3 4 5 6 7 0 60 120 180 240 300 360 420 480 C t (g/kg MMWC) time (min) (C) Experimental 0.1 M Experimental 0.25 M Experimental 0.5 M Model 0.1 M Model 0.25 M Model 0.5 M 0 1 2 3 4 5 6 7 0 60 120 180 240 300 360 420 480 C t (g/kg MMWC) time (min) (D) Experimental 0.1 M Experimental 0.25 M Experimental 0.5 M Model 0.1 M Model 0.25 M Model 0.5 M Figure 1. Experimental and model P kinetic extraction using inorganic and organic acids as solvent: (A) sulfuric acid, (B) nitric acid, (C) oxalic acid, (D) citric acid. Operation conditions were: 10% w/v, 200 rpm, and 30 °C. Table 4. Kinetic model parameters of P extraction from MMWC using pseudo-second-order kinetic equation Type of acid Molarity k2C e r 2 (M) (kg MMWC/(g P ·min)) (g P/kg MMWC) Sulfuric acid 0.1 0.050 ±0.001 2.17 ±0.03 0.9992 0.25 0.043 ±0.002 5.49 ±0.06 0.9996 0.5 0.018 ±0.001 6.61 ±0.01 0.9986 Nitric acid 0.1 0.310 ±0.039 0.40 ±0.01 0.9966 0.25 0.066 ±0.003 1.96 ±0.03 0.9991 0.5 0.048 ±0.005 6.03 ±0.07 0.9995 Oxalic acid 0.1 0.093 ±0.009 2.48 ±0.04 0.9993 0.25 0.030 ±0.001 4.40 ±0.05 0.9995 0.5 0.022 ±0.001 5.34 ±0.07 0.9995 Citric acid 0.1 0.072 ±0.003 3.10 ±0.03 0.9997 0.25 0.051 ±0.003 4.99 ±0.04 0.9998 0.5 0.025 ±0.002 5.84 ±0.01 0.9980 www.soci.org M Fernández-Delgado et al. wileyonlinelibrary.com/jctb © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). J Chem Technol Biotechnol 2022; 97: 167–178 174 Table 6. Lang factor method for estimating the main costs for the pilot plant for struvite production with the following extraction conditions: 30 °C, 10% w/v, 200 rpm, 100 kg h −1 of MMWC, and 25% of humidity Yield Struvite Equipment cost TIC Production cost Minimum sale price g P/kg DM Kg h −1 €€€/year €/kg €/kg struvite H 2 SO 4 0.5 M 6.60 3.92 30 000 130 000 98 000 3.13 4.96 HNO 3 0.5 M 5,07 3.57 20 000 110 000 200 000 7.00 8.68 Oxalic acid 0.1 M 2.51 1.47 30 000 130 000 150 000 12.76 16.80 Citric acid 0.1 M 3.05 1.84 40 000 170 000 100 000 6.79 11.93 -400 -300 -200 -100 0 100 200 300 400 -80% -60% -40% -20% 0% 20% 40% 60% 80% NPV (thousands of €) Price variaton (%) (A) MMWC Struvite Sulfuric acid -800 -600 -400 -200 0 200 400 600 800 -80% -60% -40% -20% 0% 20% 40% 60% 80% NPV (thousands of €) Price variation (%) (B) MMWC Struvite Nitric acid -600 -400 -200 0 200 400 600 -80%-60%-40%-20% 0% 20% 40% 60% 80% NPV (thousands of €) Price variation (%) (C) MMWC Struvite Oxalic acid -800 -600 -400 -200 0 200 400 600 800 -80%-60%-40%-20% 0% 20% 40% 60% 80% NPV (thousand of €) Price variation (%) (D) MMWC Struvite Citric Acid Figure 2. Sensitivity analysis for the proposed scenarios. (A) Extraction with sulfuric acid 0.5 M; (B) Extraction with nitric acid 0.5 M; (C) Extraction with oxalic acid 0.1 M; (D) Extraction with citric acid 0.1 M. Table 5. Physical properties and dimensionless numbers relative to the extractant liquids used Type of acid Molarity ρL⊘LDA×10 13 Re ×10 4 Sc ×10 10 A2×10 8 Predicted k2×10 2 (M) (kg m −3 ) (cP) (cm 2 s −1 ) (kg MMWC/(g P·min)) (kg MMWC/(g P·min)) Sulfuric acid 0.1 1008 2.50 5.61 2.24 4.38 9.51 5.02 0.25 1024 2.56 5.39 2.21 4.56 8.12 4.31 0.5 1041 3.39 4.15 1.69 7.85 3.24 1.81 Nitric acid 0.1 1013 2.24 6.27 2.49 3.53 59.78 30.96 0.25 1017 2.41 5.83 2.33 4.06 12.58 6.60 0.5 1022 2.60 5.41 2.17 4.70 9.11 4.84 Oxalic acid 0.1 1013 2.46 5.71 2.27 4.26 17.73 9.33 0.25 1016 2.59 5.44 2.17 4.68 5.57 2.95 0.5 1017 2.90 4.85 1.94 5.88 4.01 2.17 Citric acid 0.1 1015 2.48 5.67 2.26 4.31 13.69 7.21 0.25 1024 2.59 5.43 2.18 4.66 9.58 5.09 0.5 1044 2.93 4.80 1.96 5.85 4.50 2.45 Phosphorus recovery from organic waste: technical and economic evaluation www.soci.org J Chem Technol Biotechnol 2022; 97: 167–178 © 2021 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). wileyonlinelibrary.com/jctb 175