Introduction

To realize a low-carbon society, it is necessary to accelerate the reduction of carbon dioxide emissions through the effective use of natural energy sources, such as photovoltaic (PV) power generation [1]. Since the 2000s, the use of PV panels has expanded rapidly from small-scale individual houses to large-scale mega-solar power plants [2]. As the lifetime of PV panels is typically approximately 25 years, the significant increase in spent PV panels will become a serious issue by 2030 [3, 4]. Most of the installed PV panels are crystalline silicon-based PVs. Silicon PV panels are composed of an aluminum frame, a junction box, a glass plate, a back sheet made of multilayer plastics such as polyvinyl fluoride [PVF: (CH2CHF)n] and polyethylene terephthalate [PET: (C10H8O4)n], an ethylene–vinyl acetate [EVA: (C2H4)n(C4H6O2)m] copolymer as an encapsulant, and silicon and nonferrous metals such as Cu and Ag wires for current collection [5]. Copper is used for busbars in solar panels due to its high electrical conductivity. Ag has also excellent electrical conductivity (better than Cu), but Ag is used only for finger wires due to its high price. The PV recycling processes can be generally classified as dismantling, sorting (by physical separation), and chemical processing [6,7,8]. During the dismantling process, the aluminum frame and junction box are separated from the PV panel. A glass plate is then sometimes treated with a hot-knife method, collected as is, and subsequently used for horizontal recycling [9,10,11]. In the sorting process, PV panels are crushed using shredders and separated into glass, metals (i.e., Cu, Ag, and Si), and resin by combining the gravity, magnetic, and eddy current separation methods. Finally, high-purity glass and metals are recovered by leaching after the resin is removed by burning.

Ag is the most valuable material in PV panels. The average amount of Ag in PV panels is reported to exceed 630 g/t, which significantly affects the economics of PV panel recycling [12]. Ag, which is mainly contained in cell sheets, can be recovered using a simple chemical process such as nitric acid leaching [13]. In nitric acid leaching, a long leaching time and a large amount of acid are required to dissolve the Ag adhered to the resin. In fact, it has been shown that heating is required to achieve high Ag recovery [14, 15], and it is inevitable to lower the recovery cost of Ag. Mechanical and thermal methods have also been investigated for the separation of Ag [16,17,18,19]; however, the thermal treatment for removing resin and recovering the valuable metals emits large amounts of CO2.

Recently, the pulsed discharge method for material separation was thoroughly investigated [20,21,22,23]. In the previous study, the direct pulsed discharge method of electric pulse technology was applied for separating Ag finger wires in a PV cell sheet [24, 25]. A combination of the pulsed discharge method and mechanical pulverization was confirmed to concentrate Ag into fine particles; in addition, an optimal flow was proposed using the electric pulse technique as a pretreatment for milling [24]. In this study, nitric acid leaching experiments were conducted to recover Ag from samples collected using the discharge method and mechanical pulverization, as proposed in a previous study [24]. In addition, the experiment was conducted with samples collected using only mechanical pulverization (without the discharge method). Finally, the results were compared and examined for the possibility of the electric discharge method as a pretreatment to improve Ag recovery by acid extraction.

Materials and methods

Sample preparation

PV cell sheet sample

A waste crystalline silicon solar cell (Shanghai JA Solar Technology, JAM6(K)-60-290/PR, China) was used after removing its aluminum frames and cover glass plates. To detach the cover glass from the cell sheet, the hot knife method (cutting the EVA layer under the glass layer using a heated knife blade) was employed. Each cell sheet was cut into pieces of 156 mm × 156 mm (equivalent to one individual cell) and used in the experiments [24]. Each sample contained eight Cu busbars (four at the top and bottom, respectively) and 102 Ag finger wires in parallel contact with the Cu busbars. The cross section of this cell sheet sample is illustrated in Fig. 1. The chemical composition of the cell sheet reported in [24] is presented in Table 1.

Fig. 1
figure 1

Schematic image of the photovoltaic sheet cross section

Full size image
Table 1 Components of the photovoltaic panel sheet (referred from Imaizumi et al. [24]
Full size table

Electric pulse treatment

A cell sample (156 × 156 mm) was used for the underwater electric pulse discharge experiment. The electrodes were arranged diagonally on a Cu busbar, and the electrical explosion experiments were performed five times on each sample (Fig. 2). The experimental method is described in detail in [24]. After the electric pulse discharge experiment, the sample was recovered as three different components, the remaining cell sheet, the particles (under φ1.18 mm) collected by wet sieving, and the separated Cu busbars (Fig. 2). The Ag finger wires between the Cu busbars were separated and collected as particles; however, the Ag finger wires at both ends, where the electrical explosion did not occur, remained in the cell sheet. To recover the remaining Ag in the cell sheet, the sheet was ground using a cutter mill (Vertical Mill, VM-16; Orient Grinding Mill Co.) (Fig. 2). The rotational speed of the cutter mill, the grinding time, and the inner diameter of the screen was set to 1500 rpm, 10 min, and 12 mm, respectively. The cell sample without electric pulse treatment was ground using the cutter mill and subsequently used for the acid-leaching experiments (Fig. 2). The recovered particle samples, the ground sample of the remaining cell sheet, and the ground sample of the cell sheet without electric pulse treatment were labeled as A(I), A(II), and B, respectively. Ag concentration and amount from single cell sample (156 × 156 mm) in sample A(I), A(II), and B are 1.96 wt% (0.27 g), 0.05 wt% (0.12 g), and 0.15 wt% (0.39 g), respectively. Schematic flow diagram of the sample preparation is described in Fig. 2.

Fig. 2
figure 2

Schematic flow diagram of sample preparation for leaching experiments

Full size image

Acid-leaching experiment

From each sample [A(I), A(II), and B], 10.00 g was added to a PTFE vessel containing 100 mL of leachate (60% nitric acid, analytical grade). The leachate was stirred at 180 rpm using a magnetic stirrer (HSH-6A AS ONE). The reaction temperature was set to 25 ℃. After starting the leaching experiment, 250 µL of sample was aliquoted at 1, 2.5, 5, 10, 20, 40, 80, 160, 1500, 4500 and 10,080 min. These samples were filtered through a 0.1 µm membrane filter (Supor(R) Acrodisc, PALL). The filtered samples were diluted and analyzed by ICP–OES (iCAP 6500 Duo, Thermo Fisher Scientific Inc.).

Results and discussion

Results of the acid-leaching experiments

The experimental results are displayed in Fig. 3. After 10,080 min, the Ag elution rate was approximately 95% for samples A(I) and A(II); however, it was only 85% for sample B. Surprisingly, this result clearly indicates that the pulse treatment of PV panels can increase the maximum Ag recovery rate for the ground cell sample after the pulse [A(II)], as well as for particles recovered from the pulse experiment [A(I)].

Fig. 3
figure 3

Ag extraction rate from sample A(I), A(II), and B

Full size image

In the next section, leaching experiments are conducted using samples containing Ag wires in different exposed states to fully explain the differences in the elution characteristics of each sample.

Modeling leaching experiments with different exposed states of Ag wires

Leaching experiments were conducted using samples containing Ag wires in different exposure states. As it was difficult to extract individual components from the actual samples, samples were prepared with different coverage states by peeling the surface of the EVA layer from the PV cell sheet by hand (Fig. 4). Three samples were prepared: an untreated cell sheet, a wire-exposed cell sheet (cell sheet after peeling the surface EVA layer), and an EVA-coated wire (peeled surface EVA layer). In the wire-exposed cell sheet sample, the Ag wire attached to the silicon wafer was exposed, and the EVA layer was peeled off (Fig. 4b). In contrast, as displayed in Fig. 4c, the Ag wires were encapsulated in the peeled EVA layer (the EVA-coated wire sample).

Fig. 4
figure 4

Model samples with different Ag exposure states. (a) Untreated cell sheet sample, (b) wire-exposed cell sheet sample, and (c) EVA-coated wire sample

Full size image

These three samples clearly displayed different results (Fig. 5). The untreated sample exhibited the slowest dissolution of Ag, and the dissolution rate of Ag was only 82%, even after 10,080 min. The peeled EVA layer also exhibited a lower dissolution speed than the wire-exposed sample, and the dissolution rate of Ag at 10,080 min was 91%. In the case of the wire-exposed sample, 55% of the Ag was eluted in the first minute, and the elution rate reached 99% by the end of the experiment.

Fig. 5
figure 5

Ag extraction rate from model samples

Full size image

Component analysis based on the modeling leaching experiments

Based on the results of the leaching experiments, Ag was assumed to exist in four different states (components 1–4) in the pulse-treated cell samples. A schematic representation of each component is displayed in Fig. 6.

Fig. 6
figure 6

Schematic images of components 1 to 4 displaying different Ag states in the PV samples used in acid leaching experiment

Full size image

Component 1: Fully exposed Ag wire or Ag particles

The fully exposed Ag wire (peeled off from the EVA and Si wafer) and Ag particles were designated as Component 1. This component dissolved immediately after the reaction started (1 min) and corresponded to the dissolved part of the wire-exposed cell sheet sample in the modeling leaching experiments within 1 min.

Component 2: Ag wires remaining on the silicon wafer

The Ag wires remaining (attached) on the silicon wafer are designated as Component 2. Component 2 is defined as the remaining part of the wire-exposed cell sheet after dissolving Component 1.

Component 3: Ag wires remaining between EVA layer and Si wafer

The Ag wires remaining between the EVA layer and the Si wafer were assigned to component 3. This component corresponds to the EVA-coated wire sample in the modeling of the leaching experiments and dissolves.

Component 4: Ag wires in untreated (unbroken) cell sheet sample

The Ag wires in the untreated (unbroken) cell sheet samples were assigned to component 4. This component corresponded to the untreated cell sheet sample in the modeling leaching experiments and dissolved.

The proportions of these four components were then estimated for samples A(I), A(II), and B. The proportions were calculated to minimize error based on the least-squares method, so that the sum of the four components matched the actual experiment data. The calculated proportions of each component are listed in Table 2. A comparison between the leaching experiments and the sum of the estimated component balances is shown in Fig. 7.

Table 2 Estimated percentage of components in each sample
Full size table
Fig. 7
figure 7

Comparison between the leaching experiments and the sum of estimated component balance

Full size image

As shown in Table 2, sample A(I) accounted for 40.1% of component 1, 12.3% of component 2, and 47.6% of component 3; however, it did not contain component 4. Sample A(II) was accountable as 35.3% (component 1), 22.2% (component 2), 28.5% (component 3), and 14.0% (component 4). Sample B was accountable as 39.7% (component 1), 20.9% (component 2), 0.0% (component 3), and 39.5% (component 4).

Component 1 was the major component in all samples (35–40%), and the amount of component 2 was less than that of component 1. Focusing on component 3 (Ag wires remaining between the EVA layer and Si wafer), sample A(I) showed the largest value among the three samples. This is somewhat unusual as the cell sheet texture did not remain in sample A(I), and no Ag wire was observed to be covered by the EVA of the silicon wafer. However, in sample A(I), Ag particles covered by Si and aggregation of Ag and Si were observed (Fig. 8). The Ag wire (paste wire) is composed of Ag and Si. The electric pulse experiments were performed in water; therefore, the plasmatized Ag and Si might have rapidly cooled and aggregated in water. As a result, it is suggested that Si acts as the surface layer of Ag particles and that the elution behavior was similar to component 3.

Fig. 8
figure 8

BES images and EDS mapping data of the sample A(I)

Full size image

Finally, sample A(I) did not contain the component 4. This result was expected considering that the particles were collected after pulse treatment. However, both samples A(II) and B contained some amount of component 4, and the amount of component 4 in sample A(II) was significantly smaller than that in sample B. This clearly displays that the exfoliation caused by the pulse discharge occurred even in the region where no direct current flowed.

Although the above are rough estimates based on model experiments, these results display that pulse discharge to PV panels is an effective method for the selective recovery of Ag wires as well as for the subsequent acid extraction. In addition, the effect of the pulse discharge extends to the region where no direct current flows. This method can increase both the silver leaching speed in the acid leaching process and the final silver recovery rate compared with the case without pulsed treatment. This indicated the possibility of a significant improvement in the overall efficiency of the recycling process of Ag from PV panels.

Conclusion

Leaching experiments were conducted to recover Ag from PV panel samples that had been treated with electric pulses. During the electric pulse treatment, the Ag wires in the energized area were plasmatized and recovered as particles smaller than 1.18 mm, whereas the Cu busbars retained their original shape. After the Cu busbars were removed, Ag leaching experiments were conducted on the recovered particle samples and remaining cell sheets. The cell sheets were ground using a cutter mill prior to the leaching experiments. Cell sheets that had not undergone the electric pulse treatment were used in the same manner for comparison. The highest Ag recovery was therefore observed in the particle samples recovered from the electric pulse treatment. In addition, the cell sheets treated with electric pulses displayed a higher Ag leaching rate compared with that of the untreated cell sheets. Model experiments indicated that electric pulse treatment may have efficiently destroyed the structure of the entire cell sheet. A series of experiments displayed that the addition of electric pulse treatment to the PV panel recycling process could lead to more efficient silver recovery.