Enrichment and recycling of Zn from electroplating wastewater as zinc phosphate via coupled coagulation and hydrothermal route

2.1 Zn-bearing wastewater pretreatment

Zn-bearing wastewater contained 81 mg/L of Zn, 1.36 mg/L of Fe, 0.62 mg/L of Al and 3.6 mg/L of nitrate at pH 4.8 whilst nitrate, phosphate and chemical oxygen demand were 3.60, 10.26 and 441.17 mg/L, respectively. It was sampled at the effluent of electroplating workplace of Jilin Sanhe Company (Changchun, China). Fig. 1 shows that 0.4 g of commercial polymeric (PAFC, Chuangshi, Zhengzhou, China) was dispersed in 2 L of wastewater to form a mixture. The mixture was stirred at 90 rpm for 10 min. During agitation, 20% NaOH solution was dropwise added into wastewater to keep the pH at 8. A brownish sludge was collected at the wastewater bottom after 2 h and then centrifugated at 5500 rpm for 5 min. Next, about 25 g of centrifugated sludge with water content of 82% was generated. A small portion of sludge was freeze-dried at − 80 °C for characterisation. The relative weight percentages of Fe, Al, Zn, Na and N were 12.6%, 5.3%, 8.7%, 2.1% and 0.03%, respectively. The crystal structure of Fe/Al/Zn-bearing deposits was analysed by X-ray powder diffraction. The surface morphology of the deposits was observed by field-emission scanning electron microscopy (Fig. 2).

Close

Fig. 1

Flowchart of Zn recovery from wastewater.

Export to PPT

Close

Fig. 2

(a) XRD patterns of Fe/Al/Zn-bearing deposits (b) SEM images of Fe/Al/Zn-bearing deposits.

Export to PPT

The sludge was mixed with 3 M nitric acid and 2 M hydrochloric acid at the solid–liquid ratio of 0.5 under constant stirring at 90 rpm and then heated at 80 °C for 3 h. The leachate (named as Solution A) was collected and then characterised by an inductively coupled plasma optical emission spectrometer (ICP-OES, PerkinElmer, USA). Solution A contained 19.3 g/L of Zn, 31.2 g/L of Fe, 10.8 g/L of Al and 184.6 g/L of nitrate. For the prevention of potential exploding accident from nitrate decomposition in the following hydrothermal system, Solution A was then diluted by deionised water to produce Solution B.

2.2 Separation of impure Fe/Al

A hydrothermal experiment was performed to separate Fe/Al from Solution B. Firstly, 20 mL of Solution B, 0.5 g of glucose and 0.16 g of disodium phosphate were mixed in a 50 mL Teflon vessel and then agitated at 125 rpm for 20 min. Then, the Teflon vessel was sealed and heated to 190 °C for 10 h. An Fe/Al-bearing precipitate was generated at the vessel bottom and then collected, and the supernatant (Solution C) was kept for Zn recovery. Fe/Al separation was optimised by varying the dosage of disodium phosphate from 0.16 g to 0.98 g, the heating temperature in the range of 130–190 °C and the reaction time of 2–20 h. A control experiment was also conducted in the absence of disodium phosphate and/or glucose according to the preceding steps.

2.3 Further precipitation of Zn in solution C

In Solution C, the Zn, Fe and Al concentrations were 9.5, 0.03 and 0.18 g/L, respectively. When the solution was adjusted to pH 5 with the addition of 4 g/L of phosphate, a new whitish precipitate was generated and collected. The collected precipitation was washed thrice with deionised water and dried at 80 °C for 5 h.

2.4 Characterisation

Total organic carbon (TOC) and pH of Solutions A/B/C were determined by a pH metre (S210-S, Mettler Toledo, US) and a TOC analyser (TOC 500; Shimadzu, Japan), respectively. Phosphate and nitrate concentrations were detected by an ion chromatograph (IC, Metrohm AG, Switzerland). Solid products were characterised by an X-ray diffractometer (XRD; Rigaku, Japan) and a scanning electron microscope (SEM; FEI Co., USA). The recovered Zn product was also analysed using X-ray fluorescence (Rint 2200, Rigaku, Japan).

3.1 Enrichment of Zn from wastewater

The Zn-bearing wastewater was effectively treated by commercial PAFC. The Zn removal efficiency steadily increased from 34% to 98.4% with the increase in PAFC dosage from 0.05 g/L to 0.2 g/L (Fig. 3a). After treatment, a Zn-bearing sludge was generated, and the remaining wastewater contained 1.3 mg/L Zn, which met the discharge standard of electroplating industry of China (China, 2008, Daocheng and Lihui 2009). The sludge had water content of 60% and was directly leached with a mixed nitric acid and hydrochloric acid to produce Solution B with Fe, Al, Zn and nitrate concentrations of 15.6, 5.4, 9.7 and 82.3 g/L, respectively, after dilution.

Close

Fig. 3

Removal efficiencies of (a) Zn from wastewater and (b) Fe/Al/Zn from acid leachate without glucose and phosphate, (c) Variation of solution pH and nitration concentration and (d) XRD pattern of Fe-bearing deposits.

Export to PPT

For Zn recovery, Fe/Al impurities were preferentially removed via a one-step hydrothermal route. Without glucose and phosphate, only Fe precipitation was observed (Fig. 3b). The removal efficiencies of Fe/Al were approximately 0 at 130 °C but were elevated to 46.3% and 0.9%, respectively, at 160 °C and further to 54.2% and 1%, respectively, at 190 °C. This increase was accompanied by the loss of 2.4% Zn in the temperature range of 130–190 °C. This finding showed that the separation of Fe/Al impurities from Solution C was unsuccessful. When the temperature was increased from 130 °C to 190 °C, the solution pH dropped from 0.25 to 0.11, and the nitrate concentration slightly decreased from 82.3 g/L to 69.7 g/L (Fig. 3c). With Fe removal, irregular hematite block was generated (Figs. 3d and 4). Fe removal was predominated by Fe hydrolysis at high temperature and accompanied by the release of H⁺ into the solution. The thermal decomposition of nitrate occurred with temperature increase, thereby reducing its concentration (Amano et al., 2015). The removal mechanism of Fe is detailed in Section 3.4.

Close

Fig. 4

SEM images of deposits generated at (a) 160 °C and (b) 190 °C without glucose and phosphate.

Export to PPT

3.1.1

3.1.1 Effect of glucose on Fe/Al separation

Fe precipitation continued when glucose was added to the reaction system (Fig. 5a). When 0.5 g of glucose was added, the removal efficiency of Fe was 88.3% at 130 °C, then increased to 97.9% at 160 °C and further to 98.6% at 190 °C. These values were higher than those without glucose (Fig. 3 b). The Al/Zn removal rates were only 1.3% and 1.7%, indicating that most of Al/Zn were kept in Solution C. Nitrate concentration decreased from 82.3 g/L to 29.6 g/L with the variation of temperature from 130 °C to 190 °C, and this phenomenon was in accordance with the decrease in TOC from 10.8 g/L to 1.8 g/L at 130 °C, 0.7 g/L at 160 °C and 0.5 g/L at 190 °C (Fig. 5b). This finding was affiliated to the redox reaction of nitrate to glucose in the hydrothermal system, where H⁺ was consumed and solution pH increased from 0.25 to 0.82 (Fig. 5b) The removed Fe was involved in the hematite precipitates (Fig. 5c), which appeared as spherical particles with the diameter of 1 μm at 130 °C (Fig. 6a) and shrunk to a diameter of 0.5 μm at 160 °C (Fig. 6b) but were polymerised as block with coarse surface at 190 °C (Fig. 6c). Even when the solution pH was elevated to 0.82, Al/Zn remained in the solution and were not hydrolysed in the temperature range of 130–190 °C. Such discussion is shown in Section 3.4.

Close

Fig. 5

(a) Removal efficiencies of Fe/Al/Zn with glucose; (b) Variation of solution pH, TOC and nitrate concentration and (c) XRD patterns of Fe-bearing deposits.

Export to PPT

Close

Fig. 6

SEM images of the deposits generated at (a) 130 °C, (b) 160 °C and (c) 190 °C with glucose.

Export to PPT

3.2 Synergistic removal of Fe/Al

3.2.1

3.2.1 Effect of phosphate dosage on Fe/Al separation

The temperature of 190 °C was used for the synergistic removal of Fe/Al because it was optimal for Fe removal, as shown in Fig. 7. The removal efficiency of Al was only 1.4% with 5 g/L of phosphate but was increased to 49.3% with 10 g/L of phosphate, further to 96% with 15 g/L of phosphate and steadily to 97.5% with 30 g/L of phosphate. Fe removal also reached 98.7–100% in the phosphate dosage range of 5–30 g/L. However, Zn loss was lower than 2.3% in the phosphate dosage of 5–15 g/L but considerably varied to 23.8% with 30 g/L of phosphate, implying that 30 g/L of phosphate was an overdose (Fig. 7a). The solution pH was 0.82 without phosphate (Fig. 5b) but decreased to 0.5 after 5 g/L phosphate addition and slightly up to 1.28 with the phosphate dosage increased to 30 g/L. After reaction, the concentrations of the remaining nitrate and TOC were 29.6 and 0.46 g/L, respectively, without phosphate but increased to 44.3 and 2.73 g/L, respectively, after 5 g/L phosphate addition. This finding suggested that phosphate inhibited the redox reaction of nitrate to glucose. However, the concentrations of nitrate and TOC slightly decreased to 34 and 0.87 g/L, respectively, with the increase in phosphate dosage to 30 g/L (Fig. 7b and c), indicating a reverse effect to the redox reaction with low dosage of phosphate. Such performance was related to the Fe-bearing minerals, e.g. hematite and phosphate ferric. Generally, free Fe and hematite can be reduced by glucose in the hydrothermal system and then oxidised rapidly by nitrate, thus exhibiting a high catalytic effect to accelerate the redox reaction of nitrate and glucose (Qu et al., 2019). With phosphate addition, free Fe was precipitated as solid phosphate ferric by phosphate and did not participate in the redox reaction, thus leading to the retention of nitrate and glucose in the solution. However, phosphate addition also introduced H⁺ into the solution, which was beneficial for the redox reaction to consume nitrate and TOC gradually. This promoting effort of H⁺ was lower than that of Fe catalysis.

Close

Fig. 7

(a) Removal efficiencies of Fe/Al/Zn; (b) Variation of solution pH, TOC, remaining nitrate and (c) Remaining phosphate concentrations and (d) XRD patterns of deposits.

Export to PPT

After the reaction, phosphate also remained in the solution, and its concentration increased from 0.4 g/L to 4.4 g/L with the addition of 5–30 g/L phosphate, suggesting that the majority of phosphate was involved in the removal of Fe/Al/Zn. The removed Fe/Al/Zn was converted into spheric particles, where its diameter grew steadily from 2 μm with 5 g/L of phosphate (Fig. 8) to 10 μm with 15 g/L of phosphate and remained almost unchanged with 30 g/L of phosphate. This trend was consistent with the Fe/Al removal efficiencies. Corresponding EDS images showed signals of Fe/P in the four precipitates; however, the Al signal was weak with 5 g/L of phosphate, became visible with 10 g/L of phosphate and further distinct with 15 and 30 g/L of phosphate. Additionally, the Zn signal was blurry with the phosphate dosage of 5–15 g/L but became visible at 30 g/L. This trend agreed with the heavy loss of Zn (Fig. 7a). Several weak signals that belonged to the background value in the detection room of the SEM instrument were also observed (Fig. 8a and b) (Newbury et al., 2009). Al and Zn were heavily removed, but their compounds were not found in the corresponding XRD patterns. Only giniite peaks were observed. This finding revealed that the removed Al/Zn was involved in the giniite formation and accordingly, promoted the growth of giniite sphere.

Close

Fig. 8

SEM images and EDS patterns of deposits generated by adding (a) 5, (b) 10, (c) 15 and (d) 30 g/L of phosphate.

Export to PPT

3.2.2

3.2.2 Effect of reaction time on Fe/Al separation

The reaction time on Al removal was optimised, as shown in Fig. 9. The removal efficiency of Fe reached 99.7% in the first 2 h and remained stable in the following time. Compared with that of Fe, the Al removal efficiency was only 57.9% at 2 h but increased to 96% at 6 h and remained nearly unchanged in the time range of 6–20 h (Fig. 9a). This finding showed that Fe was removed prior to Al, and the Al removal was equilibrised at 6 h. During hydrothermal treatment, Zn loss was steadily below 3.4%, suggesting that Zn remained in the solution. Thus, the optimal reaction time was 6 h, which was lower than the 20 h for Al removal in the report of Liu et al.

Close

Fig. 9

(a) Removal efficiencies of Fe/Al/Zn; (b) Variation of solution pH, TOC, nitrate and (c) Variation of phosphate concentrations; and (d) XRD patterns of deposits generated at 2, 6 and 20 h.

Export to PPT

After the reaction, the phosphate concentration was 2.8 g/L at 2 h and reduced to 1.07 g/L at 6 h and slowly to 1.04 g/L in the following reaction time (Fig. 9c); this trend was similar to the variation of nitrate and TOC. Conversely, the solution pH was 0.6 at 2 h but slightly increased to 0.71 at 20 h from the continuous consumption of nitrate and glucose. With Fe/Al removal, giniite spheres with average size of 2–5 μm were generated in the first 2 h but did not further grow the following time course (Figs. 9d and 10).

Close

Fig. 10

SEM images of the deposits generated at 2, 6 and 20 h.

Export to PPT

3.3 Precipitation of Zn from rest solution

After Fe/Al removal, over 98% Zn was kept in the solution, which was in accordance with the remaining 9.5 g/L Zn. However, the remaining Fe and Al were only 0.03 and 0.18 g/L, respectively. Such solution was adjusted to pH 5, where the remaining Zn was removed as white zinc phosphate hydrate aggregates (Fig. 11 a and b) with 98.1 wt% Zn₃(PO₄)₂·2H₂O, and minor impurities of 0.12% Fe and 1.2% Al (Fig. 11c). This result showed that the recycled Zn-bearing product was highly purified.

Close

Fig. 11

(a) XRD pattern, (b) SEM image and (c) Major composition of Zn-bearing deposit.

Export to PPT

This method left Fe/Al concentrations of only 11 and 31 mg/L, which were lower than the concentrations of residual Al left after the use of precipitant, extractant, pH adjustment and conventional hydrothermal methods. For instance, the amounts of Al and Fe removed were 10% and 89% at pH 3, respectively, and then 96.6% and 99% at pH 4.1, respectively, but approximately 29.1% of Zn precipitates at the same time because of the coprecipitation. This result indicated that pH adjustment was not effective in Fe/Al/Zn separation (Park et al., 2014). In addition, Wang et al., reported that 8.2% rare earth elements (REEs), 94.4% Al and 96.9% Fe were precipitated from REE-containing solution with 8-hydroxyquinoline overdose, and 136 mg/L Al remained (Yw et al., 2019). Fe/Al should be further removed before rare earth extraction. Compared with such methods, hydrothermal treatment was commonly used in the synthesis of nanoscale boehmite with the addition of organics. Such organics were used as a dispersing agent to trap hydrated alumina nanoparticles and inhibit the polymerisation of the alumina nanoparticles. However, the Al removal efficiency from the reaction solution was not considered in these studies. Zhan Qu et al., precipitated 99.6% Fe in the leaching solution of electroplating waste sludge as hematite at 160 °C, but Al remained in the supernatant (Qu et al., 2021). Chenggui Liu et al., realised the coprecipitation of Fe and Al from the leaching solution of wastewater coagulation sludge, but the precipitation temperature was as high as 270 °C. Rui Bian et al., hydrothermally precipitated 98% Al from the acid solution, but the reaction temperature reached 240 °C (Bian et al., 2021). These research methods can only precipitate Fe alone under low-temperature conditions. The Fe/Al synergistic removal reaction temperature was much higher than this research and consumed more energy.

3.4 Separation mechanism of Fe/Al in the presence of phosphate

When the electroplating sludge was leached by strong acid, Fe, Al and Zn were dissolved and released into the leachate. Such heavy metals were easily hydrolysed with the increase in leachate pH. Hence, the related hydrolysis performance was analysed by using Medusa software (1.0, Royal Swedish Institute of Technology, Swiss). When the leachate was adjusted to pH 4, the hydrolysis of Fe/Al as Fe(OH)₃ and Al(OH)₃ was completed, but free Zn remained in the leachate (Fig. 12a). For confirmation, a precipitation experiment was conducted. The results showed that when the leachate was adjusted to pH 4, the Zn loss approached 20.3%, along with the removal of 99.8% Fe and 90.6% Al. These findings differed from those provided by the Medusa software because Fe/Al removal produces Fe(OH)₃ and/or Al(OH)₃-bearing colloid with many hydroxyl groups for Zn coordination. Thus, the effective separation of Fe/Al without Zn loss from Zn-bearing solution cannot be achieved by pH adjustment.

Close

Fig. 12

(a) Hydrolysis diagram of Fe/Al/Zn at room temperature, (b) Gibbs free energies of Fe/Al/Zn hydrolysis with glucose and without phosphate (Before reaction: pH = 0.25, concentrations of Fe/Al/Zn were 10372, 3593 and 3093 mg/L, respectively; After reaction: pH = 0.82, concentrations of Fe/Al/Zn were 146.6, 3552.5 and 3045.4 mg/L, respectively) and (c) Gibbs free energies and (d) pH versus temperature of Fe/Al/Zn hydrolysis in the presence of phosphate (Before reaction: pH = 0.25, concentrations of Fe/Al/Zn were 10372, 3593 and 3093 mg/L, respectively, and concentration of phosphate was 15000 mg/L; After reaction: pH = 0.64, concentrations of Fe/Al/Zn were 20.7, 139 and 3038 mg/L, respectively, and concentration of phosphate was 1070 mg/L).

Export to PPT

High temperature is an alternative condition to accelerate Fe/Al removal. When the temperature increased from 25 °C to 50 °C and 190 °C, the Gibbs value of Fe hydrolysis decreased from 4.5 kJ/mol to below zero and further to −67.2 kJ/mol. This finding revealed that free Fe was stable at room temperature but started to hydrolyse at 50 °C and accelerated at 190 °C. With continued Fe hydrolysis, H⁺ was released and accumulated in the solution, thus decreasing the solution pH and inhibiting Fe hydrolysis (Eq. (1)). Thus, a portion of Fe remained in the solution. H⁺ was generated from Fe hydrolysis but can be consumed in the redox reaction of nitrate to glucose (Eq. (2)). Thus, Fe hydrolysis continued, resulting in Fe remaining in the solution at a low level. Additionally, Fe was hydrolysed as Fe oxyhydroxide colloid and then aggregated in the block form without glucose (Su et al., 2020a,b). The added glucose attached to the surface of Fe oxyhydroxide colloid via electrostatic incorporation to retard the colloid aggregation and generate fine particles. In the hydrothermal system, Fe-bearing oxyhydroxide was also converted to hematite via the conjunction reaction between adjacent Fe-OH groups with the release of water molecules (Zhu et al., 2020). Such conversion did not change the morphology of Fe-bearing precipitates. Given that only nitric and hydrochloric acids were used for the sludge dissolution, nitrate and chloridion were only introduced but did not participate in Fe precipitation. Although the Fe-bearing colloid and hematite particles consisted of many hydroxyl groups, the H⁺ on the hydroxyl groups did not dissociate in strong acid solution and were not replaced by cationic Al/Zn. Hence, Al/Zn almost remained in the supernatant. Different from Fe, Al and Zn were stable in the acid solution and did not hydrolyse in the condition. This finding was consistent with the positive Gibbs value of Al/Zn hydrolysis at the temperature of 20–190 °C (Fig. 12b).

With phosphate addition, the synergetic removal of Fe/Al transpired, and two new precipitation reactions of Fe/Al occurred via Eqs. (3)–(5). Such reactions were also analysed by HSC Chemistry software (9.5, Metso Outotec, Finland), as shown in Fig. 12(c). Fe was rapidly hydrolysed with phosphate, alongside Al removal. This effect differed from that in the absence of phosphate. For instance, Al hydrolysis did not occur in the reaction without phosphate, and positive Gibbs values were observed in the range of 25–190 °C (Fig. 12b). After phosphate addition, the Gibbs value of Al/phosphate reaction was reduced to below zero at 150 °C and further dropped to −16.9 kJ/mol at 190 °C. This finding suggested that the reaction of Al to phosphate occurred at 150 °C and accelerated at 190 °C. Thus, 96% Al removal was achieved with the addition of phosphate. Although more than 99% and 96% Fe/Al were removed, only giniite sphere was formed, and no Al-bearing mineral was recorded. Hence, the removed Al might be involved in Fe precipitation. The giniite product has a special formula of Fe₅(PO₄)₄(OH)₃·2H₂O, in which Fe³⁺ is embedded in the space structure of phosphonate-based hybrid bond and is easily replaced by free Al³⁺ to form a stable Fe/Al-bearing spheric structure (Martins et al., 2021, Rueff et al., 2016). The byproducts of Al(PO₄)₄(OH)₃·2H₂O and Fe₅(PO₄)₄(OH)₃·2H₂O were also reported as antirust coating and cathode material (Han et al., 2016, Mandal et al., 2010), and accordingly, had potential application in an industrial scale. Such byproducts were fine micro spheres and showed ideal functional groups to coordinate noble metal and/or enzyme, and/or to be served as a catalyst for contaminant removal (Kim et al., 2007, Muneyama et al., 1996). The byproducts exhibited good performance in the removal of Cr, Cu and Ni, similar to β-ketoenamine linkage, Fe₃O₄-modified biochar and common adsorbents of activated carbon and molecular sieve (Han et al., 2016, Wang et al., 2020, Yang et al., 2019, Zhong et al., 2020) (Dai et al., 2019, Fl et al., 2021, Liu et al., 2021a,b).

The hydrothermal precipitation of Al and phosphate was also performed without Fe to investigate the Al-bearing product in the presence of phosphate. The results showed only typical AlPO₄ peaks and irregular blocks (Fig. S1), indicating that the major product of Al and phosphate was AlPO₄ when Fe was absent in the solution. Given that pure AlPO₄ crystal was not formed in Fe/Al-bearing solution, the removed Al was involved in the Fe precipitation of giniite sphere. Thus, a new route for the precipitation of Fe/Al with phosphate (Eqs. (3) and (5)) different from the formation of Fe/Al oxyhydroxide was proposed. Nearly 11 mol H⁺ was released to generate 1 mol giniite. Such H⁺ was involved in the redox reaction of nitrate to glucose. When the solution pH increased correspondingly, giniite generation was promoted, resulting in the removal of Fe/Al at a high level.

After the reaction, the solution was at pH 0.6 and contained 0.14 g/L Al, where the corresponding Gibbs value for Al hydrolysis was −1.2 kJ/mol, which was close to zero. This finding showed that the precipitation reaction of Al to phosphate was in equilibrium. When the solution pH was above 1, the precipitation of the remaining Al continued, but that of Zn also started (Figs. 7a and 12d), which was not desirable for Zn recovery.

3.5 Environmental applications

The proposed method showed its advantages in separating Fe/Al efficiently and recovering Zn as high-purity products. Other wastes, e.g. electronic waste (Olafisoye et al., 2013) and smelting residue (Gorai et al., 2003), also contained heavy metals and Fe/Al impurities and can be treated with this technique for recycling heavy metals. Other low-grade precious metal minerals, e.g. brookite, galena and millerite (Chen 2010), were also rich in Fe and Al that could be recovered using this method. Additionally, the recovered Fe/Al products were highly purified and have potential application in the production of electronic devices (Rueff et al., 2016), biological materials (Nedkov et al., 2016) and battery materials (Hong et al., 2003). However, this method also generated rest solution with abundant nitrate and phosphate. Future investigations should focus on the recovery of other valuable metals from Fe/Al-bearing hazardous waste and the recycling possibility of rest solution with high concentrations of nitrate and phosphate.