Introduction

The production of electrical and electronic equipment (EEE) has been increasing continuously, especially in the past three decades, depending on the increasing market share, developments in technology, and luxurious living conditions. In parallel with the production of EEEs, the amount of electrical and electronic equipment completed its economic life, called e-waste, is increasing dramatically. According to the Global E-waste Monitoring 2020 report, the amount of e-waste produced increased from 9.2 Mt in 2014 to 53.6 Mt in 2019. Accordingly, it is predicted that the amount of e-waste is increasing globally, from approximately 61 million tons in 2023 to 74.7 million tons in 2030 [1, 2]. The amount of e-waste, the fastest-growing type of solid waste, constitutes approximately 5% of the annual solid waste produced worldwide [3]. The most serious problem with e-waste is that many countries do not have official regulations on recycling and disposal of e-waste. About 20% of e-waste has been collected and recycled using unofficial or illegal recycling methods. Processing of e-waste under inappropriate conditions leads to air, water and soil pollution and therefore negatively affects human and environmental health [4, 5].

Printed circuit boards (PCBs), provide mechanical support and conductivity between electronic components and are vital parts of electronic devices. Waste printed circuit boards (WPCB) are important secondary sources, as they contain metals (Cu, Zn, Al, etc.) and precious metals (Au, Ag, etc.) with economic value [6]. The metal concentration in WPCBs is almost tens or even hundreds of times higher than in raw ore. The silver, gold, palladium and copper concentrations of WPCBs are approximately 100, 25, 11 and 5 times higher than those in raw ores, respectively [7]. Therefore, the advantages of metal recovery from secondary sources over recovery from primary sources are that it provides significant energy savings and reduces carbon dioxide emissions [8].

Recovery of metals and precious metals from e-waste is generally followed by methods based on pyrometallurgy [9] and hydrometallurgy [10]. The pyrometallurgical method requires special equipment to work at high temperatures (over 1000 °C). Therefore, both installation and energy costs are very high. In addition, it is harmful to the ecological system as it produces dusty and toxic polluting gases during the process. Hydrometallurgical processes are preferred especially for the recovery of metals from secondary sources due to their advantages such as low cost, high recovery rates, low process temperature, and different reagent options such as acids, bases, halides, or microorganisms [11]. Hydrometallurgically, leaching agents such as strong mineral acids (HCI, HNO3 and H2SO4) [12], ammonia-based solutions [13], ionic liquids [14], supercritical fluids [15], and organic acids [16] are generally used in the recovery of Cu from WPCBs.

Aqua regia studies are of great interest because they provide faster than other leaching agent dissolution in the recovery of gold which is one of the precious metals. Although aqua regia is a very effective reagent in dissolving metals, it is not preferred because it is corrosive. In addition, it is very difficult to deal with the highly concentrated acidic solution formed as a result of the gold recovery process. Cyanide leaching is also widely used for gold recovery [17, 18]. However, since this method produces a large number of cyanide compounds that can cause serious harm to humans and the environment, this method has been replaced by alternative methods over time [19]. These methods are generally leaching systems containing thiosulfate and thiourea. Thiosulfate is a non-toxic and low-priced reagent. Although thiosulfate has a high recovery efficiency of gold, it increases the cost due to the high reagent consumption requirement [20]. On the other hand, thiourea leaching has been frequently preferred in recent years due to its rapid reaction with gold, low cost and less environmental pollution compared to cyanide [21, 22].

Most of the literature studies published on the recycling of WPCBs have focused mainly on the effects of policy, legislation, resources or the process conditions of metal recovery [23,24,25,26,27,28]. Some studies have examined the recovery processes of only one metal among the metal contents in WPCBs. Awasthi and Li [29] and Işıldar et al. [8] studied physical, pyrometallurgical and bio-hydrometallurgical methods for metal recycling technologies. Also, in order to accelerate metal leaching, they examined some combined methods, such as mechanical treatment with biodegradable reagents or organic acid leaching [8, 29]. Many different hydrometallurgical methods such as leaching, solvent extraction, membrane technology etc., have an important place in the recovery of metals from secondary sources. In this way, hydrometallurgy causes relatively less environmental pollution compared to pyrometallurgy [30].

Pressure leaching provides the opportunity to work in higher temperature and pressure environments than traditional leaching processes. This allows metals to dissolve faster and with higher efficiency. Pressure acid leaching is used industrially to recover copper in the metallurgical industry. WPCBs are compact structures that contain metals, plastics and epoxy resins. Plastic and epoxy resins resist acid reaching the metals in this compact structure, making their dissolution difficult. With pressure leaching, lower acid concentrations and higher metal recovery can be achieved in a shorter time, as high temperature and pressure effects weaken the resistance layer. Only a few articles on pressure leaching of metals from WPCBs has been found in the literature [31, 32]. It is known that ultrasound assisted leaching processes increase the dissolution rate compared to conventional leaching processes. Cavitation bubbles formed by ultrasound collapse around the solid surface, and the boundary layer around the surface breaks and so mass transfer between solid and liquid surface improves. Another effect of ultrasound is that it causes erosion on the particle surface, reducing the particle size and creating micro-cracks on the particle surface [33].

The literature survey has shown that there has not been a study using the multi-objective optimization technique to study the recovery of base metals and precious metals from WPCBs. Also, ultrasound has not been applied in the recovery of precious metals. The purposes of the study were (i) to determine the maximum dissolution conditions of base metals (Cu, Zn) under pressure leaching from printed circuit boards of end-of-life desktop computer motherboards, (ii) to determine the maximum dissolution conditions of precious metals (Ag, Au) by ultrasound in an acidic thiourea medium, using the solid residue from the optimized pressure leaching step the experiment carried out under the optimum conditions of the first stage, (iii) to determine optimum operating conditions for both base and precious metals using Taguchi-based gray relationship analysis, and (iv) to examine the effects of parameters on response variables by ANOVA analysis.

Materials and Methods

Chemicals

The chemical materials used in the experiments are Sulfuric acid (H2SO4, 95%, Merck, Germany), Hydrogen peroxide (H2O2, 35%, Merck, Germany), Thiourea (CS(NH2)2, Apollo Scientific, England), and iron(III) sulfate hydrate (Fe2(SO4)3, Merck, Germany).

The Preparation of WPCB

WPCBs used in this study were obtained from end-of-life desktop computer motherboards in Turkiye. Easily removable rough circuit elements such as capacitors, resistors, metal cooling plates, cables, etc. on WPCBs were separated with auxiliary hand tools such as pliers, screwdrivers, and chisels. Flat sheets were cut into approximately 3 × 3 cm pieces using stainless steel sheet shears and then subjected to a two-stage mechanical grinding process with a Retsch SK 100 cross-impact crusher and Brook Crompton Series 2000 rotary cutter, respectively. The sieve analysis showed that more than 95% of the WPCB (the flat sheets) was below 2 mm and this grain size (− 2 mm) was used in the experiments. The visuals of this stage are given in Fig. S1a, b, c.

Dissolution of Base and Precious Metals in WPCB by Two-Stage Leaching

Stage 1-Recover Base Metals by High-Pressure Leaching

Dissolution experiments of base metals from WPCB were carried out in a 1000 mL stainless-steel autoclave (Parr, Model 4828,) that provides the opportunity to measure temperature, pressure and stirring speed values (see Fig. S2). Acid concentration (H2SO4), amount of oxidizing agent (H2O2), solid/liquid (S/L) ratio, temperature, pressure and reaction time were chosen as experimental parameters. The stirring speed was kept constant at 300 rpm in the experiments. Parameters and their level values used in leaching operations are shown in Table S1. Since 6 parameters were examined at 5 levels in base metal leaching experiments, the L25(56) orthogonal array was chosen as the experimental plan (Table S2). After the solid material was put in the reactor in accordance with the determined solid/liquid ratio, H2SO4 (300 mL) and H2O2 were added at the specified concentrations according to the desired experimental conditions. Oxygen gas was used when necessary to stabilize the overall pressure inside the reactor. H2O2 is one of the most preferred oxidants in acidic solutions because it is strong, safe, effective and relatively inexpensive oxidant [34, 35]. Additionally, an important advantage of using this oxidant is that oxygen and water are released as reaction by-products during leaching [36]. It is a reasonable oxidant that also provides environmental benefits in the leaching process, mainly due to the formation of oxygen as a byproduct of the reaction. The release of oxygen as a result of the decomposition of H2O2 enables oxygen to be used as an oxidative agent in the leaching of Cu and other metals [37]. After the autoclave reached the desired temperature, the experiments were started. At the end of the determined experiment period (according to Table S2), the system was stopped, the reactor was cooled and the pressure was relieved. Then, in order to separate the solid and liquid phases from each other, filtration was carried out with a blue band filter paper in a vacuum environment. The metal contents of the obtained leach solutions were determined by ICP-MS analysis.

Stage 2-Leaching of Precious Metals by Ultrasound and Thiourea

In this step, which is aimed at recovering gold and silver, the remaining solid of the experiment carried out under the optimum conditions of the first step was used. The solid residue was washed several times with distilled water and then dried for use in second-step experiments. Experiments were carried out in a 500 mL jacketed three-necked glass reactor. To provide the desired experimental temperature a hot water circulator (Haake L D8), and also a reflux cooler (Brookfield TC-502) to prevent solution losses due to evaporation were connected to the reactor. An ultrasound power supply (Cole Parmer Ultrasonic Homogenizer, 400W, 20 Hz) with a 2 cm diameter probe was used to examine the effect of ultrasound on the dissolution. Since the most efficient operating percentage of the device is between 0 and 240W, the maximum capacity was chosen at 240W. The experimental setup is shown in Fig. S3. In the leaching of precious metals (PML) experiments, ultrasound power, thiourea concentration, acid concentration, and reaction temperature were selected as parameters to be studied. Other parameters for the experimental system, solid/liquid ratio, amount of Fe+3, stirring speed, and reaction time were kept constant as 1/30 g/mL, 6 g/L, 300 rpm, and 120 min, respectively. The parameters and their levels used in PML experiments are shown in Table S3. The L9(34) orthogonal array was chosen as the experimental plan for the PML experiments where 4 parameters were examined at three levels. The experiment plan is given in Table S4. After adding H2SO4 solution and thiourea at a certain concentration according to the experiment plan, Fe2(SO4)3 was added into the reactor so that the total solution content contained 6 g/L Fe+3 ions. The ultrasound probe was immersed in the solution, operated at the desired power, and the reactor was heated to the desired temperature. In all experiments, 5 g of WPCB sample was added to the reactor and then the reactor contents were mixed with a magnetic stirrer at 300 rpm to provide a homogeneous suspension. At the end of the reaction time of 120 min, filtration was carried out under a vacuum using blue band filter paper.

Each experiment in Table S3 and Table S4 was repeated twice in randomized order under the same conditions to observe the effect of noise sources. The dissolution values (%) of the metals were calculated using Eq. 1.

$$\% \,{\text{Dissolution}}\, = \,\frac{{{\text{Amount}}\,{\text{that}}\,{\text{passes}}\,{\text{into}}\,{\text{solution}}\,{\text{as}}\,{\text{a}}\,{\text{result}}\,{\text{of}}\,{\text{dissolution}}}}{{{\text{The}}\,{\text{total}}\,{\text{amount}}\,{\text{of}}\,{\text{metal}}\,{\text{in}}\,{\text{the}}\,{\text{sample}}}}\, \times \,100$$
(1)

Optimization Method

One-response optimization techniques are widely used to find the best solution in different engineering branches because they are easy to implement. The Taguchi method is widely used in leaching processes due to its advantages such as time and cost reduction [38,39,40]. However, the weakness of this method is that it is only effective for single-performance optimization. Therefore, it is widely used in conjunction with grey relation analysis for multiple performance optimizations [41,42,43,44].

Taguchi Analytical Method

The Taguchi approach is a powerful statistical technique that examines all independent combinations of selected factors with a small number of experiments. The method uses orthogonal arrays that allow many parameters to be changed simultaneously and a large number of parameters to be evaluated with very few experiments [45]. The Taguchi method uses signal-to-noise (S/N) ratios to determine the performance characteristic. S/N ratios are evaluated in three categories ‘the bigger the better’, ‘the smaller the better’ and ‘the more normal the better’. They can be calculated using Eqs. (2), (3) and (4) [46, 47].

$${\text{The }}{\mkern 1mu} {\text{ bigger}}{\mkern 1mu} {\text{ the}}{\mkern 1mu} {\text{ better }}{\mkern 1mu} {\text{S}}/{\text{N}} = {\mkern 1mu} - {\mkern 1mu} 10\,\log_{10} \left( {\frac{1}{n}\sum\limits_{i = 1}^{n} {\frac{1}{{Y_{i}^{2} }}} } \right)$$
(2)
$${\text{The }}{\mkern 1mu} {\text{ smaller}}{\mkern 1mu} {\text{ the}}{\mkern 1mu} {\text{ better }}{\mkern 1mu} {\text{S}}/{\text{N}}{\mkern 1mu} = {\mkern 1mu} - {\mkern 1mu} 10{\mkern 1mu} \log_{10} \left( {\frac{1}{n}\sum\limits_{i = 1}^{n} {Y_{i}^{2} } } \right)$$
(3)
$${\text{The }}{\mkern 1mu} {\text{more}}{\mkern 1mu} {\text{ normal}}{\mkern 1mu} {\text{ the}}{\mkern 1mu} {\text{ better}}{\mkern 1mu} {\text{S}}/{\text{N}}{\mkern 1mu} = {\mkern 1mu} - {\mkern 1mu} 10{\mkern 1mu} \log_{10} \left( {\frac{1}{n}\sum\limits_{i = 1}^{n} {(Y_{i} - Y_{0} )^{2} } } \right)$$
(4)

S/N is defined as the signal/noise ratio in the performance statistic. n is the number of repetitions experimentally. Yi is the value of experiment performance, and Y0 is the desired normal value.

Grey Relational Analysis

The grey relational analysis (GRA) method is used to evaluate together the parameters acting on multiple output targets. Since Taguchi is not sufficient in this sense, optimization is done by combining it with the grey relational method. In this way, it enables to achieve objectives such as reducing component sizes and costs and improving performance at the same time. GRA enables it to transform the multi-response optimization design into a single-response. This is accomplished in four stages as given below [48]:

Normalization

First, the analysis results are normalized between 0 and 1. There are three different criteria for the normalization procedure: (i) the bigger the better, (II) the smaller the better, and (III) the more normal the better [49]. In this study, it was normalized using the 'the bigger the better' criterion (Eq. 5), as it was aimed to maximize the recovery of base and precious metals.

$$y_{i} (k) = \frac{{x_{i}^{0} (k) - \min x_{i}^{0} (k)}}{{\max x_{i}^{0} (k) - \min x_{i}^{0} (k)}}$$
(5)

While yi(k) represents the normalization value of the grey associative formation, max x0i(k) and min x0i(k) represent the maximum and minimum value of x0i(k) respectively and x0 the optimum value.

Distance matrix

The distance matrix (Δ0i) is determined by Eq. 6.

$$\Delta_{0i} (k) = y_{0} (k) - y_{i} (k)$$
(6)

Δ0i(k) is the deviation value between y0(k) and yi(k). y0(k) is the reference array and is taken as 1. yi(k) is the comparative array.

Grey relationship coefficient

The grey relationship coefficient (γi) gives the relationship between the desired and actual normalized value and it is expressed as follows [49]:

$$\gamma_{i} = \frac{1}{n}\mathop \sum \limits_{k = 1}^{n} \omega_{i} \varsigma_{i} (k)$$
(7)

ωi is the weight of the corresponding response and the weight value is 0.5 for each response.

Analysis of Variance (ANOVA)

Analysis of variance (ANOVA) is a method used to determine the effect level of parameters on the response. ANOVA includes the sum of squares, degrees of freedom, variance, and percent effect [46]. As a result of the calculations, the Fexperimental value and the Ftable values are compared and the contribution of the parameters are determined. If the parameter contribution is Fexperimental > Ftable, the relevant parameter is interpreted as effective, otherwise, the parameter is interpreted as ineffective. At the same time, values less than 0.05 in the 95% confidence interval of the p-value also show the effectiveness of the relevant parameter [48]. The statistical analyzes were performed using the MINITAB 18 program.

Analysis of PCBs and Metals in Solution

To determine the metal content of PCB plate sample, the ground sample was dissolved in aqua regia and analyzed with Inductivity Coupled Plasma-Mass Spectroscopy (ICP-MS, Agilent 7800). Structural identification of the material was made by scanning electron microscope (SEM) equipped with EDS (Zeiss Sigma 300). Analyzes of base and precious metals from the solutions obtained after the leaching process were determined by Inductivity Coupled Plasma-Mass Spectroscopy (ICP-MS, Agilent 7800).

Results and Discussion

Characterization of Raw Material

The chemical composition and SEM–EDS images of the WPCBs are given in Fig. 1. As can be seen from Fig. 1, the WPCB sample contains Cu, Fe, Al, Zn, Sn, and Ni as base metals and, Au and Ag as precious metals. Copper and zinc are major elements. WPCBs differ in terms of metal content (w %) depending on the source. The metal contents of WPCBs obtained from different sources have been reported as 9.70–47.5% Cu, 1–1.22% Zn, 3–10.36% Fe, 2.7–5.8% Al, 2–3.3% Sn, 1.4–2.24% Pb, 0.69–2.0% Ni, 0.014–0.03%Au and 0.06–0.13%Ag [50]. From the SEM image (Fig. 1), it is seen that the sample was heterogeneous both in structure and distribution. This depends on the fact that the printed circuit board has a composite structure.

Fig. 1
figure 1

The SEM–EDS images and metals content of the raw WPCBs

Full size image

Optimization of High-Pressure Leaching of Base Metals (Cu and Zn)

Single Objective Optimization

The dissolution reactions of copper and zinc in the media of sulfuric acid (H2SO4) and hydrogen peroxide (H2O2) are as follows [51]:

$${\text{Cu }} + {\text{ H}}_{{2}} {\text{SO}}_{{4}} + {\text{ H}}_{{2}} {\text{O}}_{{2}} \, \to \,{\text{CuSO}}_{{4}} + {\text{ 2H}}_{{2}} {\text{O}}$$
(8)
$${\text{Zn }} + {\text{ H}}_{{2}} {\text{SO}}_{{4}} + {\text{ H}}_{{2}} {\text{O}}_{{2}} \, \to \,{\text{ZnSO}}_{{4}} + {\text{ 2H}}_{{2}} {\text{O}}$$
(9)

It is known from literature data that H2SO4 cannot leach copper and that additional oxidants such as hydrogen peroxide are needed for this [51,52,53]. Additionally, the leaching of Cu by sulfuric acid in the presence of dissolved oxygen occurs spontaneously. The basic copper oxidation reaction is given by the following equation.

$${\text{2 Cu }} + {\text{ H}}_{{2}} {\text{SO}}_{{4}} + {\text{ O}}_{{2}} \, \to \,{\text{2 CuSO}}_{{4}} + {\text{ 2H}}_{{2}} {\text{O}}$$
(10)

The dissolution percentages of copper and zinc and their corresponding S/N values are given in Table 1. Since the maximum dissolution of copper and zinc in the waste printed circuit board was aimed, dissolution values were evaluated according to the larger the better performance statistics. Each experiment in Table S3 was repeated twice in randomized order under the same conditions to observe the effect of noise sources. I and II in Table 1 represent the results of experiments performed at different times and in order of random.

Table 1 Performance values for optimization of the recovery of Cu and Zn in pressure leaching
Full size table

The normalization test of copper and zinc was performed with the Anderson–Darling test using the data in Table 1. Normalization test results are shown in Figs. S4 and S5. The Anderson–Darling test is used in the branch of statistics to test whether the data show a certain probability distribution, that is, for goodness-of-fit testing. The test statistic is used to find the p-value for the determined theoretical probability distribution. If the calculated p-value is greater than 0.05, the null hypothesis is accepted and the data is considered to fit the determined probability distribution [54]. The p-value were determined at 0.335 for means of Cu and 0.063 for means of Zn. These values were greater than the significance level (0.05). So, it can be said that the null hypothesis is supported and the data are normally distributed [55].

Data homogeneity was evaluated with Levene’s and multiple comparison tests for a standard deviation of 0.05, and the results are given in Fig. S6. Levene’s homogeneity test is used to determine whether populations have similar variance. If the variances are similar, the data are homogeneous. If p > 0.05 as a result of the Levene test, the assumption of homogeneity is ensured [56]. As shown in Fig. S6, all p-value are greater than 0.05 and the bars do not overlap according to both test results. Therefore, it can be said that the null hypothesis is supported and the data are homogeneously distributed [44, 57].

Performance statistics graphs showing the effect of the parameters on copper and zinc dissolution are given in Fig. 2a. The effects of the parameters on the dissolution of copper and zinc were similar. The effect of pressure on the dissolution of metals was examined between 5 and 25 bars. As the pressure increased from 5 to 20 bar, dissolution increased, and above 20 bar, no significant change in dissolution was observed. This can be explained as the oxygen mass transfer rate increases as the pressure increases up to 20 bar, and the pressure above 20 bar does not have a significant effect on the oxygen mass transfer [58]. The effect of temperature on the dissolution of metals was examined in the range of 30–130 °C. Metal solubility increased with increasing temperature. The reaction rate constant depends on temperature as exponential according to the Arrhenius equation. The increase in temperature increases the reaction rate constant and improves the metal extraction kinetics. In addition, since increasing temperature reduces the viscosity of the solution, it facilitates the mass transfer of oxygen to the surface and accelerates the reaction progress [59]. The effect of acid concentration on metal dissolution was examined at acid concentrations in the range of 1–3 M. Metal dissolution increased up to 2 M acid concentration. This is the result of increasing the number of moles of acid that will react with the metals in the WPCB. Acid concentration above 2 M did not alter metal dissolution. Therefore, it can be said that 2 M acid concentration is a suitable concentration to dissolve Zn and Cu [60]. The increase in the amount of H2O2 had a positive effect on the solubility of metals. Hydrogen peroxide is one of the most powerful oxidizing agents with a reduction potential of 1.77 V. The oxygen released by the decomposition of H2O2 is adsorbed on the material surface and transfers electrons to the solution. The decomposition rate of H2O2 depends on peroxide concentration, temperature and acid concentration. This has a catalytic effect on the dissolution of metals. Since the oxygen released with increasing peroxide concentration will increase, it can be said that the dissolution of metals increases as a result of its catalytic effect on more metal surfaces [61]. The effect of solid/liquid ratio on the dissolution of metals was examined in the range of 0.02–0.1. Metal dissolution increased up to 1/30 g/mL solid liquid ratio and decreased at rates above that. Increasing the solid/liquid ratio will increase the amount of particles per unit volume of solvent. For this reason, the contact between the solid and the fluid will decrease and the amount of dissolution will decrease [59]. The effect of reaction time on the dissolution of metals was examined in the range of 30–90 min. While metal dissolution increased up to 60 min, there was no significant change in the amount of dissolution after the 60th minute. This shows that 60 min is sufficient time for the metals to dissolve.

Fig. 2
figure 2

The effect of each parameter on the performance statistics of base metals a Taguchi b GRG

Full size image

It can be seen from Fig. 2a. that the optimum conditions for the maximum solubility of copper and zinc, represented by their top points on the graphs, were A4 (Pressure, 20 bar), B5 (Temperature, 130 °C), C2 (H2SO4, 1,5 M), D5 (H2O2, 35%), E3 (S/L ratio, 1/30 g/mL), F3 (Reaction time, 60 min) and A4 (Pressure, 20 bar), B5 (Temperature, 130 °C), C3 (H2SO4 2.0 M), D5 (H2O2, 35%), E2 (S/L ratio, 1/40 g/mL), F4 (Reaction time, 75 min), respectively.

In order to determine the statistical effects of the parameters on the Cu and Zn dissolutions, analysis of variance (ANOVA) was performed and evaluated in terms of parameters with significant effects (p < 0.05 at a level of 95% confidence). Variance analysis results are given in Table 2 Since F values were Fpressure > Ftemperature > Ftable (F(4,4) = 6.39) for copper, and Fpressure > Ftemperature > FH2SO4 > Ftable (F(4,4) = 6.39) for zinc. These parameters were effective parameters for copper and zinc leaching. Also, p-value supported these results. According to the p-values, the most effective parameters in the dissolution of copper were pressure and temperature, while in the dissolution of zinc, pressure, temperature, and H2SO4 concentration. The acid concentration and solid/liquid ratio did not have a significant effect on the dissolution of copper and zinc, respectively. While 29.73% pressure, 28.44% temperature, 19.83% H2O2, time12.85%, and 4.97% S/L ratio contributed to the dissolution of copper, 27.83% pressure, 24.73% temperature, 24.41% H2SO4 concentration, time 9.70%, and 9.92% H2O2 contributed to the dissolution of zinc.

Table 2 The analysis of variance (ANOVA) of Cu, Zn and GRG for recovery of base metals
Full size table

Multi Objective Optimization

The values of normalized S/N ratio, Grey Relation Coefficients (GRC), Grey Relation Grade (GRG) and rank for pressure leaching are summarized in Table 1. As can be seen from Table 1, the highest GRG value of 0.99 was obtained in experiment no 15. The conditions for this experiment were A3 (15 bar), B5(130 °C), C2 (1.5 M H2SO4), D4 (25% H2O2), E2 (1/50 g/mL), F3 (60 min). The main effect plot of the grey relation grade against the process parameters for pressure leaching is shown in Fig. 2b. According to the multiple performance analysis (Fig. 2b), the parameter levels providing the optimum conditions were A4 (20 bar), B5 (130 °C), C3 (2.0 M H2SO4), D5 (35% H2O2), E2 (1/40 g/mL), F3 (60 min).

The results of the grey relationship degree were analyzed by the ANOVA method (p < 0.05 at 95% confidence level). The results are presented in Table 2. The most effective parameter was temperature with a p-value of 0.01 (contribution 33.76%), followed by pressure (p = 0.012, 31.82% contribution), H2O2 (p = 0.052, 13.56% contribution) and H2SO4 concentration (p = 0.074, 10.81% contribution), respectively.

Confirmation Test of Pressure Leaching

The confirmation test was performed under optimum operating conditions (A4 (20 bar), B5 (130 °C), C3 (2.0 M H2SO4), D5 (35% H2O2), E2 (1/40 g/mL), F3 (60 min.) obtained from the Taguchi-based grey analysis to validate the optimization of multiple performance characteristics. At the end of the experiment carried out under optimum conditions of multiple performance characteristics, 99.62% of copper and 98.76% of zinc were dissolved. These results are higher than the values obtained from the initial setting of the parameters used for the Taguchi-based grey multiple performance characteristic optimizations. The results reported by previous researchers confirm the conclusion obtained above [41, 62].

Optimization of Leaching Precious Metals (Au and Ag)

At this stage of the study, the solid remaining from the experiments carried out under optimum conditions determined from the Taguchi-based grey method of the pressure leaching study was used. In the experiment carried out under optimum conditions, gold was never dissolved and only 3.39% of silver was dissolved. The SEM–EDS image of the remaining solid as a result of the test performed under the optimum conditions of the pressure leaching is shown in Fig. 3. In Fig. 3 the metal and other structures in the solid were examined morphologically at different magnification ratios and the metal contents were determined quantitatively using EDS graphics.

Fig. 3
figure 3

The SEM–EDS image of the remaining solid from performed experiment under the optimum conditions of the pressure leaching

Full size image

The gold and silver dissolution reactions in the presence of thiourea (2CS(NH2)2) and Fe+3 in acidic medium are as follows [63,64,65]:

$${\text{Au}}_{{({\text{s}})}} + {\text{ Fe}}^{{ + {3}}} + {\text{ 2CS}}\left( {{\text{NH}}_{{2}} } \right)_{{2}} \to {\text{Au}}\left[ {{\text{CS}}\left( {{\text{NH}}_{{2}} } \right)_{{2}} } \right]_{{2}}^{ + } + {\text{ Fe}}^{{ + {2}}}$$
(11)
$${\text{Ag}}_{{({\text{s}})}} + {\text{ Fe}}^{{ + {3}}} + {\text{ 3CS}}\left( {{\text{NH}}_{{2}} } \right)_{{2}} \to {\text{Ag}}\left[ {{\text{CS}}\left( {{\text{NH}}_{{2}} } \right)_{{2}} } \right]_{{3}}^{ + } + {\text{ Fe}}^{{ + {2}}}$$
(12)

Single Objective Optimization

Since the maximum dissolution of gold and silver in the waste printed circuit board was aimed, dissolution values were calculated according to the larger the better performance statistics. Dissolution percentages obtained at the end of the experiments were repeated twice and S/N ratios are given in Table 3.

Table 3 The values of dissolution and S/N ratios of Au and Ag in WPCB
Full size table

The Anderson–Darling test was performed using the data in Table 3 to determine the normalization test for gold and silver. Normalization test results are shown in Figs. S7 and S8. The p-value of the means of Au and means of Ag were 0.435 and 0.206, respectively. Since these values are greater than 0.05, the null hypothesis is supported and the data are normally distributed. Multiple comparisons and Levene’s test were used for data homogeneity (for standard deviation α = 0.05) and their graphs are shown in Fig. S7 and Fig. S8. As seen from the figures, the bars did not overlap and the p-values were greater than 0.05. Therefore, it can be said that the null hypothesis is supported and the data are homogeneously distributed [66].

The performance statistics graph showing the effect of the parameters on the dissolution of Au and Ag is given in Fig. 4a. The effect of parameter levels on the solubility of Au and Ag was similar. The effect of ultrasound on precious metals was examined at 0–240 W ultrasound power. The dissolution of precious metals increased with increasing ultrasound power. In heterogeneous solid–liquid systems, cavitation bubbles formed in the solution environment by the effect of ultrasound collapse near the solid surface. To fill this gap, a non-symmetrical liquid jet is formed towards the solid surface. This breaks the stagnant layer at the solid–liquid interface and facilitates the wetting of the solid surface by the liquid. Since cavitation effects cause many micro cracks to appear on the solid surface, it increases the penetration rate of the solution and improves the dissolving efficiency [33]. The effect of thiourea concentration on the dissolution of Au and Ag was examined in the concentration range of 20–60 g/L. Increasing thiourea concentration increased the solubility of precious metals. This means that the dissolution of Au and Ag is highly dependent on the thiourea concentration. Similar results have been reported by Lee et al. [21]. The effect of H2SO4 concentration on the dissolution of precious metals was examined at 10–30 g/L acid concentrations. While precious metal solubility increased up to 20 g/L acid concentration, it decreased after this concentration. As the acid concentration increases, the increase in the H+ ion concentration, which is effective in dissolving, has a positive effect on the solubility of precious metals. As the concentration increases after a certain concentration value, the H+ ions in the solution react very quickly around the particle and reach a saturation value. This creates a difficult-to-dissolve film layer around the solid and therefore the dissolution rate decreases [67]. The effect of temperature on the dissolution of precious metals was examined in the range of 20–60 °C. Increasing temperature decreased the solubility of precious metals. This is the result of increasing temperature increasing the decomposition of sulfur compounds, causing loss of thiourea [68].

Fig. 4
figure 4

The effect of each parameter on the performance statistics of precious metals a Taguchi b GRG

Full size image

It can be seen from Fig. 4a that the optimum conditions for maximum solubility of Au and Ag were A3 (Ultrasound power, 240W), B3 (Thiourea, 60 g/L), C2 (H2SO4, 20 g/L), D1 (Temperature, 20 °C). When the experiment plan given in Table S4 is examined, it is seen that this study combination was studied in experiment no 9.

Analysis of variance (ANOVA) was performed at a 95% confidence level (p < 0.05) to determine the statistical effects of parameters on Au and Ag dissolutions. The results are summarized in Table 4. The F-values of ultrasound power and thiourea concentration for gold and silver were found to be greater than the F table value (F(3,2) = 19.16). According to p-values, the most effective parameters were ultrasound power and thiourea amount for the dissolution both of Au and Ag, respectively. Ultrasound power and thiourea amount contributed 48.03% and 37.98% to the dissolution of gold, and 53.11% and 41.54% to the dissolution of silver, respectively. The contribution of temperature was 12.52% for gold and 4.98% for silver. The acid concentration did not have a significant effect on the dissolution of precious metals.

Table 4 The analysis of variance (ANOVA) of Au, Ag and GRG for recovery of precious metals
Full size table

Multi Objective Optimization

The values of normalized S/N ratio, Grey Relation Coefficients (GRC), Grey Relation Grade (GRG) and rank for Au and Ag leaching are shown in Table 3. As can be seen from Table 3, the highest GRG value of 1.00 was obtained in experiment no 9.

The main effect plot of the GRG against the process parameters is shown in Fig. 4a. According to the multiple performance analysis (Fig. 4b), it is seen that the parameter levels that provide the optimum conditions are A3 (Ultrasound power, 240W), B3 (Thiourea, 60 g/L), C2 (H2SO4, 20 g/L), D1 (Temperature, 20 °C).

Fig. 5
figure 5

The SEM–EDS image of solid samples remaining at the end of the two-step hydrometallurgical process experiment

Full size image

Since this combination is the same as the optimum level combination found in the Taguchi analysis, the optimum operating conditions found by the Taguchi analysis were also confirmed by the grey analysis [69]. Therefore, the optimum dissolution values for gold and silver can be accepted as 85.82% and 99.15%, respectively, without the need for any confirmation experiment.

The ANOVA table (p < 0.05 at 95% confidence level) of the grey relationship degree is presented in Table 4. The most effective parameter was ultrasound power with a p-value of 0.045 (contribution 43.41%), followed by thiourea amount (p = 0.048, 40.38% contribution) and temperature (p = 0.125, 14.18% contribution).

The SEM–EDS image of the solid waste remaining at the end of the two-step hydrometallurgical process is given in Figs. 3 and 5. When these figures are compared with Fig. 1, it is seen that WPCB has changed structurally at the end of the procedures. It can be said that metal extraction from EDS samples and ICP-MS results was significantly successful. At the end of the whole process, the metal removal in the raw WPCB was as follows: Cu 99.62%, Zn 98.76%, Au 85.82%, Ag 99.15%, Fe 84.73%, Ni 96.76%, and Sn 0.22%.

Conclusion

The main results of this study, in which base and precious metals were recovered from the printed circuit boards of the motherboards (flat sheets) of end-of-life desktop computers, by a two-stage hydrometallurgical process, were as follows:

  1. 1.

    According to both single and multi-optimization methods, the most effective parameters in the recovery of base metals were determined as pressure and temperature.

  2. 2.

    The optimum conditions for first stage determined by grey relationship analysis were pressure, 20 bar; temperature, 130 °C; H2SO4, 2 M; H2O2, 35%; time, 60 min and S/L ratio, 1/40 g/mL. Under these conditions, Cu and Zn were recovered 99.62% and 98.76%, respectively, and Au was not dissolved. These dissolution values were higher than the dissolution values corresponding to the maximum value according to the GRA rank result (Table 1). A higher dissolution value was achieved with multiple analysis.

  3. 3.

    According to the ANOVA table, the most effective parameter in the dissolution of gold and silver was ultrasound power, followed by thiourea concentration.

  4. 4.

    According to Taguchi analysis of precious metals, the experimental conditions that provided maximum solubility were A3 (Ultrasound power, 240 W), B3 (Thiourea, 60 g/L), C2 (H2SO4, 20 g/L), D1 (Temperature, 20 °C). At the end of the grey analysis, the combination providing the optimum conditions was found to be the same as the single-objective optimization combination. Optimum operating conditions found in Taguchi analysis were also confirmed by grey analysis. 85.82% of Au and 99.15% of Ag were extracted from waste printed circuit boards under an optimum condition, 240 W ultrasound power, 60 g/L thiourea concentration, 20 g/L H2SO4 concentration and 20 °C temperature.