Optimization of Precious Metals Recovery from Electronic Waste by Chromobacterium violaceum Using Response Surface Methodology (RSM)

An effective recovery technology will be valuable in the future because the concentration of the precious metal contained in the source can be a key driver in recycling technology. This study aims to use response surface methodology (RSM) through Minitab software to discover the optimum oxygen level (mgL−1), e-waste pulp density (% w/v), and glycine concentration (mgL−1) for the maximum recovery of gold (Au) and silver (Ag). The method of precious metals recovery used for this study was taken from the bioleaching using 2 L of batch stirred tank reactor (BSTR). A Box-Behnken of RSM experimental statistical designs was used to optimize the experimental procedure. The result of the RSM optimization showed that the highest recovery was achieved at an oxygen concentration of 0.56 mgL−1, a pulp density of 1.95%, and a glycine concentration of 2.49 mgL−1, which resulted in the recovery of 62.40% of Au. The pulp density and glycine concentration greatly impact how much Au is bioleached by C. violaceum. As a result, not all of the variables analyzed seem crucial for getting the best precious metals recovery, and some adjustments may be useful in the future.


Introduction
In the present world, having electronic devices has become a need. Tese devices are essential for human communication and daily convenience. Up to 60 diferent metals can be found in modern electronics, and to make up for any potential shortage, there is a growing interest in secondary sources of these metals, particularly e-waste [1]. According to the European Union (EU) directive, e-waste is any waste of electrical and electronic equipment, including all components, subassembly, and consumables, or any part of the discarded product deemed unnecessary [2]. E-waste contains more than 1000 components, hazardous materials-cadmium (Cd), chromium (Cr), lead (Pb), and mercury (Hg)-and nonferrous metals such as aluminium (Al) and copper (Cu). In addition to base metals, e-waste, e.g., PCBs, also contains precious metals, such as gold (Au), silver (Ag), and platinum (Pt) [3]. Due to their high chemical stability and excellent conducting qualities, these metals are frequently used in the production of electrical products [4]. Although PCBs concentrations are not as high as base metals, this valuable resource is more difcult to fnd in original compounds on the Earth. Surprisingly, the concentrations of precious metals in PCBs can be more than ten times greater than those of commercially mined minerals [5]. An electronic device includes a component primarily made of toxic metals and chemicals. Such materials include beryllium (Be), phthalates, polyvinyl chloride (PVC), brominated fame retardants (BFRs), and antimony [6,7]. A signifcant reason for the rapid generation of e-waste is the high rate of obsolescence in the electronic market, particularly in developing countries [8]. Most electronic products are changed at least every two years, have a relatively short shelf life, and are either abandoned or exported as used goods to underdeveloped nations.
Environmental pollution has resulted from the breakdown and disposal of e-waste. According to data provided by Jaibee et al. [7]; landflls have received more than 90% of unnecessary electrical and electronic waste. As a result, landflls are producing an increasing amount of e-waste, worsening the degradation of the environment. Cancer risk increases along with developmental and neurological issues [9]. Additionally, incinerating has harmed the environment, particularly in the quality of solids, land, and air. When computer chip sludge and acid are dumped on the ground, the soil becomes more acidic, contaminating water sources [10]. Even worse, past research has revealed some health issues brought on by the hazardous metal and chemical efects of e-waste. A review conducted by Pant [11] confrmed the toxic efects of metals such as Pb, Cd, Hg, Cr, arsenic (As), nickel (Ni), selenium (Se), lithium (Li), and americium (Am) on the human body.
Microorganisms, particularly cyanogenic bacteria, have come under consideration as a potential alternative method for the gold cyanidation process in recent years. Microbes such as A. ferrooxidans, C. violaceum, and Aspergillus sp. increased gold and Cu recovery from e-waste, especially mobile-PCBs [12][13][14]. It has long been known that aerobic gram-negative bacteria can produce free cyanide [15], which Sneath and Punjab [16] initially discovered in mesophilic C. violaceum strains. C. violaceum is a mesophilic, motile, gram-negative, and facultative anaerobe [17]. Tis bacterial strain exhibits excellent potential for the ecological gold recovery technique [18,19]. Several physiological and nutritional factors (pH, temperature, glycine concentration, pulp density, and medium composition) may afect the process's end product [20]. Terefore, process optimization is needed to achieve maximal recovery with the fewest number of experiments. In this situation, statistical techniques can identify interactions between factors. Te response surface methodology (RSM) has been used to model, evaluate, and research the relationship and interaction of the individual components [21]. Only a few studies have been done using RSM for bioleaching metals from e-waste [22].
Te recovery of Au and Ag metals from e-waste is widely studied in this research. E-waste mobile-PCBs were converted into ingots and leached using the batch stirred tank reactor (BSTR) technique. Te experimental fndings are verifed using RSM software for various factors, including oxygen level, e-waste pulp density % (w/v), and glycine concentration.

Source of Mobile-PCBs Concentrates.
Te e-waste focusing on mobile-PCBs is sourced from a local electronic mobile shop in Selangor, Malaysia. No physical and mechanical methods are applied to the sample before it is transported to the laboratory to avoid the loss of its properties. Te scrap mobile-PCBs (SW110) is frst manually crushed to a uniformity of 500 μm using stainless steel scissor. Te analysis is then performed using the representative samples. Ten, 100 mL of aqua regia solution, 68% concentrated HNO 3 , and 37% concentrated HCl (HNO 3 : HCl � 1 : 3), is used to dissolve 1 g of mobile-PCBs [23]. Te solution is refuxed in a 250 mL beaker for 1 hour at 100°C. After the solution cools, the volume of solution is made up to 100 mL with deionised water. Ten, the solution is passed through glass fbre flters (PALL-GF-A/E-1) to guarantee particle-free suspensions and stored at 4°C for further analysis. Next, the mobile-PCB undergoes an inductively coupled plasma-mass spectrometry (ICP-MS) (Brand Perkin Elmer, Model NEXION 2000) technique to identify the metal content in the metal concentrates. Te APHA Method 6020B ICP-MS analysis was used to conduct the procedure, which involved inserting a sample of fltered efuent into the ICP-MS apparatus. 15 mL of pure mobile-PCBs after the aqua regia leaching was used to determine the Au and Ag concentration using ICP-MS testing. Te fne mobile-PCBs (of particle size less than 1.0 mm) are then sterilised via autoclaving at 121°C for 15 minutes for the bioleaching study [24]. E-waste may contain harmful contaminants if mismanaged. Te frst step in the isolation and purifcation process employed throughout the research was the manual physical separation of various forms of e-waste into distinct categories such as metals, plastics, and glass. Next, precious metals are extracted from e-waste via a bioleaching method utilising C. violaceum. Finally, impurities that cannot be purifed must be properly disposed of to protect the environment and public health.

Single-Culture Cyanogenic Bacteria.
Te bacteria used in this study are single-culture cyanogenic bacteria, namely, C. violaceum (strain DSMZ 30191). Tese bacteria are purchased from the German Collection of Microorganisms and Cell Culture (DSMZ), Braunschweig, Germany, in an actively growing culture (agar slant). Prior to its adaptation in this study, the C. violaceum must undergo at least twiceduplicate subculturing to prove its authenticity, which is done through identifcation techniques. Te C. violaceum is inoculated from single colonies of one loop full of slant in both nutrient agar (BD 213000) and nutrient broth (BD 234000) media for 24 hours under a stationary phase at 30°C. After an overnight confuent culture, isolated organisms are characterised according to basic morphological techniques, including morphology (classifed via size, shape, appearance, and colour), gram staining reaction count, and turbidity measurement through optical density determination. After completion of the identifcation techniques, the C. violaceum is subcultured in LB broth (Difco) under a stationary phase at 30°C, and transferred twice a month. Long-term storage of the organism is held in 2 mL cryotubes with 500 μL of 20% glycerol, and 500 μL of liquid inoculated medium at −20°C.
To obtain the optimum free cyanide formation by C. violaceum, the media used are Luria Bertani broth (LB) comprising (in gL −1 ) of 10 tryptone, 5 yeast extract, 0.75 glycine, and 10 NaCl. All media are autoclaved at 121°C for 15 minutes and cooled to room temperature before being transferred into a 2.5 L batch stirred tank reactor (BSTR) for the bioleaching study. Te dissolution of Au in cyanide solution consists of anodic and cathodic reactions as stated in the Elsner's equations by J.A. Cook [25]: 2.3. Bioleaching Process. A 2.5 L BSTR is used in experiments on the two-step bioleaching of mobile-PCBs. Te acrylic fbre tank used by the BSTR tank has a 2.5 L tank volume (at a height/diameter ratio of 2.0). A mechanical stirrer that penetrates the tank at a 45-degree angle is present. Te BSTR system was designed with a height of 27 cm and an internal diameter of 13 cm in downward aerobic fow (without recycling). Te bioreactor's top provides access to the medium and mobile-PCB feeds. Daily output samples are collected through a sampling valve built 25 cm below the bioreactor's top. Te waste the bioreactor has processed will be used as an efuent product, which will take the form of an ingot solution. To investigate the impact of aeration on the dissolution of precious metals using the BSTR method, pure air was supplied at a fowrate of 0.5 Lmin −1 and was produced by a gas compressor. In contrast, the gas compressor was turned of throughout the BSTR procedure at the nonaerated state (0.0 Lmin −1 ). During the 14-day incubation period, no controls were utilised to regulate the reactor's temperature or pH. Figure 1 displays BSTR employed in the study. In earlier investigations, two-step bioleaching (indirect bacterial leaching) was preferred over direct bacterial leaching. Tis is such that even if microorganisms do not directly interact with the minerals, they do oxidise the ores through the leaching agents they produce [3]. Additionally, it was thought that the two-step procedure would boost metal mobilisation yield and decrease lixiviant usage [1]. First, 2 L of a sterile medium with a pH of 7.0 is added by inoculating it with a single culture of C. violaceum at a rate of 5% v/v. A gas compressor concurrently supplies pure air. 1% (w/v) pulp density of mobile-PCB was applied under aseptic circumstances after two days of incubation. It is known that diferent amounts of metals can be separated depending on the organism and the growth conditions used. Selection of the right microorganisms, i.e., acidophilic or alkalophilic in terms of cyanogenic types of bacteria, has produced efcient metal bioleaching techniques. Te physical stability of isolated metals can be confrmed through the ability of ingots solution to detect the AuCN 2 − solution (chemical formula for gold cyanide anion) at higher concentration, observe its color change over the bioleaching process, as well as the ability of C. violaceum to react well in 14-dat incubation period. Following a 14-day incubation period, a BSTR containing C. violaceum inoculum, medium, and mobile-PCBs was operated to provide the best output (when the observed yield diference was less than 5%). 45 mL of yield samples were drawn from the BSTR sampling output valve every two days and examined. Free cyanide concentration (CN − ) was measured using a cyanide test kit quantofx technique (macherey-Nagel GMbH & CO. KG) and pH was determined using a YSI 556 pH metre. A DR 6000 HACH spectrophotometer was used to measure the optical density (OD) at a wavelength of 600 nm. Te experiment was conducted by measuring the absorbance of each sample using a 3 mL disposable cuvette, and a blank sample of LB medium without C. violaceum was used as a control. Tese parameters' fndings are indicated in units Abs. Tis procedure is carried out in accordance with the standard method OD in microbiology (OD600). 15 mL of pure mobile-PCBs after the bioleaching were then examined for the concentration of the metal ions Au and Ag using an ICP-MS apparatus after being fltered through a membrane nitrocellulose flter to ensure a suspension free of particles.

RSM Analysis.
Design of experiment (DOE) is a useful approach that may be applied in various experimental contexts [26]. By connecting the quantitative data from experiments to mathematical and statistical equations, the design of experiment (DOE) modeling technique known as RSM can determine the relative impact of parameter variation and make parameter optimization easier [27].
Tree factors, including e-waste pulp density, oxygen level and additive (glycine) concentration, were optimized by box-Behnken RSM using the Minitab version 20 software to achieve maximum Au and Ag recovery. Te recovery of precious metals was taken as the RSM system's response. Te quadratic equation model predicting the optimal point was expressed according to the following equation: where y is the response variable, x i and x j are coded independent variables, β 0 , β i , β ii , β ij (i � 1, 2, . . . , k), and β ij (i � 1, 2, . . . , k; j � 1, 2, . . . , k) are regression coefcients for intercept, linear, quadratic, and interaction terms, respectively, and ε is the statistical error. Tis equation represents an empirical relationship between the yield and the independent parameter obtained from response surface methodology modeling by the Box-Behnken (regression equation) method. RSM is a method for predicting and modeling complex relationships between independent variables and one or more responses [27].
Each variable has three coding levels: low (−1), middle (0), and high (+1). To examine the pure error at the design centre, 15 run trials with three components were conducted. Te cubic model was used to select this optimization (Box-Behnken). Table 1 displays the experimental design's list of variables and their respective levels.

Characterization of Mobile-PCBs.
Te results of the metal content of mobile-PCBs are shown in Table 2 and were previously reported in our study [28]. Te amount of Cu (30.07 ± 0.200% w/w) was in the majority. Te precious metals Au (0.01 ± 0.001%) and Ag (0.02 ± 0.005%) were in signifcantly lower quantities. Te Au content in the mobile-PCBs is higher than that reported by Liang et al. [29] (0.004% Au) and comparable with electronic scrap materials (ESM) used by Tay et al. [30], i.e., 0.02% Au. Although the content of Cu was higher than that reported in Pradhan & Kumar [24], i.e., 12.06% Cu, the gold content was slightly lower (0.08% Au). Te diference in metal contents, especially Au, is due to diferent e-waste sources (computer and scrap), infuenced by the material's heterogeneity and origin.

Statistical Evaluation.
Te regression equation was used to determine the interaction between the three essential factors and precious metals recovery. Equations (5) and (6) shows the regression equation for the RSM data graphs for Au and Ag, respectively: Au recovery(%) � 59.24 Ag recovery(%) � 9.36 + 0.
Te Box-Behnken of RSM was used to optimize diferent factors (oxygen level, e-waste pulp density, and glycine concentration) afecting the recovery of metals during bioleaching. As a function of the independent factors, data on the dependent variables (oxygen level, e-waste pulp density, and glycine concentration) were ftted to a secondorder equation (Equations (5) and (6)). Te 15 experimental and predicted responses are presented in Table 3, and the statistical analysis of these data was performed at a confdence level of 95% (P < 0.05). Te coefcient of determination (R 2 ) measures the degree of ft and is defned as the ratio of the explained variance to the overall variation [31]. All the models had high precision (R 2 near 1), indicating that the model explained the variability in the data.
Based on Table 3 results, the Au and Ag recovered in this study ranged from 1.81 to 59.24% and from 0.02 to 11.65%, respectively. Te value of R 2 in this recovery of Au and Ag is    Table 4. ANOVA led to two modifed quadratic models for both responses [22]. Te P value from the coded coefcient table can be used to identify the component afecting precious metals recovery; the P value must be less than 0.05. Because X 2 has a P value of 0.016, which is less than 0.05, it impacts the Au recovery. Next, the fnding demonstrates that Ag recovery is infuenced by X 1 X 1 . Te lack-of-ft of the model is higher than 0.05 and is not signifcant, denoting that the model can ft experimental data with analytical data from the model.

Response Plots
(1) Au Recovery. Au recovery was accomplished using a twostep bioleaching process and the BSTR technique. Biogenic cyanide interacts with gold during the two-step bioleaching process to generate the water-soluble complex dicyanoaurate [18]. Au leached out using C. violaceum under two-step bioleaching are comparable to the studies of [12,24,29] [35]; with a maximum of Au bioleachability, 60.00, 3.00, 30.00, 13.62, and 11.30% w/w, respectively. Instead of a batch reactor, shaking fasks were used in all prior research. Figures 2(a)-2(c) show the relationship between Au recovery and these three parameters, the relationship of RSM will only focus on the e-waste pulp density (X 2 ). Figure 2(a) illustrates the interaction efect of oxygen level and e-waste pulp density on Au recovery. Te maximum Au recovery was observed at the optimum pulp density of 2.75% (w/v), converging with the optimum oxygen level at 0.5 mgL −1 . It was observed that increasing the pulp density above 2.75% reduced Au recovery. Environmental toxicity increases at higher e-waste pulp density, but bacterial activity decreases. Increasing pulp density would decrease the metal's recovery [33]. For instance, the leach solution must be prepared around the oxygen level's center point (0.5 mgL −1 ) and e-waste pulp density (2.75% w/v) to obtain a higher Au recovery. Besides that, Figure 2(c) represents the interaction efect of glycine concentration and e-waste pulp density on Au recovery. At a constant oxygen level of 0.5 mgL −1 , the recovery of Au is represented by varying the glycine concentration from 0.75 to 4.50 mgL −1 . While it also varies the pulp density from 0.5 to 5.0% (w/v). Increased glycine concentration from 2.2 to 3.0 mgL −1 increased Au recovery. However, due to the toxic efect of glycine on bacterial growth, the Au recovery decreased as the glycine concentration increased over 3.0 mgL −1 .
(2) Ag Recovery. Te result values in Table 3 will develop a response contour plot and surface plot for Ag recovery to facilitate seeing the best recovery results. Tus, in Figure 3, the response contour plot and surface plot for Ag recovery as a function of (a) oxygen level and e-waste pulp density is the best recovery for Ag. Te darkest center can be represented as the highest value of Ag recovery on the contour plot. Furthermore, the parameters associated with this contour and surface plot are oxygen level and e-waste pulp density, where the value for oxygen level is 0.50 mgL −1 and 2.75 (% w/v) for e-waste pulp density. Tis means that the absence of glycine concentrations does not afect the recovery process to get the highest recovery for Ag. Combining two parameters alone can also obtain the highest recovery meaning that the perfect recovery needs to be mixed up oxygen level and e-waste pulp density.

Process Optimization and Model Validation.
Optimization of the variables was then identifed using the response optimizer function in the Minitab software version 20. Figures 4 and 5 show the optimization results for Au and Ag recovery as a function of oxygen level (X 1 ), e-waste pulp density (% w/v) (X 2 ), and glycine concentration (X 3 ), respectively. Ten, to confrm the validity of the models, an independent bioleaching experiment using BSTR was performed under optimal conditions as prescribed by the software. Point prediction and experimental recovery of precious metals (Au and Ag) under optimal statistical conditions for model validation and accuracy was included in Table 5.
A maximum Au recovery desirability of 61.32% was predicted for the response variable targets with an oxygen level of 0.56 mgL −1 , e-waste pulp density of 1.95% (w/v), and glycine concentration of 2.49 mg/L. It is worth noting that the optimized Au recovery is obtained by decreasing the total value of e-waste pulp density, previously used at 2.75% w/v. However, the predicted Ag recovery is still low, even using the variables under optimum conditions (with an oxygen level of 0.52 mgL −1 , e-waste pulp density of 1.77% (w/v), and glycine concentration of 3.40 mg/L). Te confrmatory results showed that the obtained responses for Au (62.40%) and Ag (10.33%) recovery agreed well with the predicted responses. Based on the experiment conducted on the Ag recovery from e-waste, it can be concluded that the experimental recovery value does not guarantee getting the highest predicted value. As a result, more parameters must be involved than just the three necessary ones in order to obtain the best fnding. It is important to stress the study of pH, particle size, and the use of various microbes. pH is a general factor but greatly impacts the Au and Ag recovery process. E-waste has been reported to be alkaline, increasing the medium's pH. Te dissolution of precious metals is very slow, below pH 6.8 or above pH 9.5 [36]. Ten, the particle size of the e-waste source is also a factor in increasing the efectiveness of the 2-step bioleaching process. When generally solid e-waste is put into a liquid medium and reacts together, the particles between the two cannot combine, but only the particles on the surface of the e-waste will come into contact with the medium. Terefore, breaking the e-waste into small pieces, such a thing allows for an increase in the surface area ratio (area per total sample) and guarantees the activity of microorganisms to react in the dissolution of Au metal. Generally, a sieve fraction of 0.5 mm to 1.0 mm is selected for experimental research [37].

Conclusion
Two-step bioleaching process using BSTR was optimized to leach only Au metal from e-waste mobile-PCBs using Box-Behnken RSM by a cyanogenic bacterium C. violaceum. One quadratic model was proposed by RSM, which can be utilized as an efcient tool to predict Au recovery through bioleaching. Te maximum recovery occurred at an oxygen level of 0.56 mgL −1 , pulp density 1.95%, and glycine concentration 2.49 mgL −1 , which led to the extraction of 62.40% of Au. Te pulp density and glycine content greatly infuence the bioleaching of Au by C. violaceum, as the recovery of the metal would be reduced by increasing pulp density. Glycine concentrations beyond 3.0 mgL −1 were associated with a reduction in Au recovery. However, under the optimal condition suggested by the Ag model, only 10.33% of Ag was extracted (at an oxygen level of 0.52 mgL −1 , pulp density 1.77%, and glycine concentration 3.40 mgL −1 ). It can be inferred from the experiment on Ag recovery from e-waste that the experimental recovery value does not ensure receiving the greatest predicted value. Terefore, further investigation and improvement regarding parameters infuencing metal bioleaching should be undertaken.

Conflicts of Interest
Te authors declare that they have no conficts of interest.