Porous 3D Printed Scavenger Filters for Selective Recovery of Precious Metals from Electronic Waste

Selective laser sintering (SLS) 3D printing is used to fabricate highly macroporous ion scavenger filters for recovery of Pd and Pt from electronic waste. The scavengers are printed by using a mixture of polypropylene with 10 wt% of type‐1 anion exchange resin. Porosities and the flow‐through properties of the filters are controlled by adjusting the SLS printing parameters. The cylinder‐shaped filters are used in selective recovery of Pd and Pt from acidic leachate of electronic waste simply by passing the solution through the object. Under such conditions, the scavenger filters are able to capture Pd and Pt as anionic complexes with high efficiency from a solution containing mixture of different metal ions. By using the Pd/Pt scavenger together with previously reported, highly selective nylon‐based Au scavenger, precious metals, i.e., Au, Pd, and Pt could all be recovered from the electronic waste leachate in a single flow‐through process. One of the main advantages of the printed scavengers is that all recovered metals can be easily extracted from the filters as separate fractions by using aqueous solutions of thiourea or diluted nitric acid. After removal of the captured metals, the scavengers are reusable without significant loss of their ion‐capturing performance.

Development of 3D printing techniques has enabled the preparation of chemically and/or physically functional materials and devices. [1] Fascinating examples have been reported where the potential, industrially relevant applications range from antimicrobial composites to catalytically active materials and even light-emitting diodes. [2] Advances in printing material development have opened the field for intriguing discoveries such as new energy storage devices or drug delivery systems. [3] Typically, these products have been obtained by utilizing printing techniques such as stereolithography, extrusion-based methods, or inkjet printing. [1a,4] The chemical activity is often obtained www.advsustainsys.com concentrations of metals can be cost-effectively and selectively captured by using simple, reusable scavenger devices.
To prepare 3D printed functional filters with precious metal scavenging properties, swollen beads of additive 1 (type-1 anion exchange resin, Dowex 21K) were dried at 90 °C overnight and grinded to a fine powder in a ball mill. The chemically active resin (10 wt%) was mixed with readily available and easily printable polypropylene (PP), which was chosen as the supporting matrix for the chemically active component. The mixture was then used to print filters with desired shape, size (5 mm tall disc with 16.5 mm diameter) and porosity (see the Experimental Section) by using Sharebot SnowWhite SLS 3D printer. The final filters were analyzed using microscopic techniques and X-ray tomography.
Scanning electron microscopy (SEM) is commonly utilized to acquire information about the porosity and composition of novel materials. Unfortunately, nonconducting samples, like organic polymers, are known to have notable charging problem during SEM experiments. This charging effect can be reduced, for example, by image post-processing or by coating the sample with metal, but important information about surface structure could be lost during the treatment. Thus, imaging of insulating samples is challenging using standard electron microscopic methods. In helium ion microscopy (HIM), imaging is based on usage of positive He ions when the aforementioned charge problem can be circumvented by utilization of chargecompensating low-voltage electron beam (a flood gun), enabling the imaging of insulating samples. [9] Hence, the surfaces of the printed objects were thoroughly investigated using HIM (Figure 1). By inspecting the images acquired from the break surfaces of PP/resin filter (1-PP) (Figure 1, left), it is evident that the filters are highly porous throughout the sample. In general, the material appears to be composed of polypropylene beads partially sintered together forming a macroporous structure. Extensive melting of the polymer beads is not observed either in printed pure PP (Figure S1, Supporting Information) or in 1-PP. When zooming in, the particles of anion exchange resin additive 1 can be seen firmly attached to the surface of sintered polypropylene (Figure 1, middle). However, the additive is not encapsulated by the polypropylene matrix. The active resin is thus available for chemical interactions. In addition to overall porosity, using SLS technique, materials with different densities can be printed in a single object enabling fabrication of fully functioning products with solvent impermeable edges and chemically active interior ( Figure S2, Supporting Information).
Detailed and comprehensive information about the the interior structure of the objects and the distribution of the active component was acquired using X-ray tomography (Figure 1, right). The overall porosity of the objects was established from the high estimated porosity percent (≈32%). The pore size distribution is relatively even, as the diameter of most of the pores falls between 20 and 60 µm ( Figure S3, Supporting Information). The data obtained from this analysis demonstrates that the inside structure of the filters corresponds to the HIM images and, even more importantly, that the active material (Dowex 21K) is evenly distributed over the whole object.
The chemical performance of the printed scavengers was evaluated by using a synthetic test solution containing 100 mg L −1 of Ni, Zn, Fe, and Cu and 50 mg L −1 of Al, Cr, Pb, and Sn along with 5 mg L −1 of Pd and Au and 1 mg L −1 of Pt in 5% aqua regia. The chosen composition mimics the metal ratio in an average sample of an aqua regia dissolved Printed Circuit Board (PCB) waste. [10] Synthetic test solution was used due to the fact that the metal concentrations in true WEEE leachates vary considerably depending on the actual waste material. Hence, analysis of a well-defined synthetic mimic solution is needed for estimating the true capacity and the overall performance of the printed scavengers. It should also be noted that the main part of the metal-scavenging experiments with 1-PP Pd/Pt scavenging hybrid material was conducted with a test solution without two key components, Ag and Au, found in electronic waste. From authentic PCB waste, silver can precipitate in presence of chloride, and hence it was not included in any of the test solutions. Similarly, gold can be removed completely and selectively by using the 3D printed nylon scavenger prior to 1-PP filter (see below and Table S4, Supporting Information). [10] Hence, the synthetic test solution mimics the leachate after removal of Au and Ag. The scavengers were also tested with authentic WEEE solutions, which were prepared by using ultrasound-assisted aqua regia leaching to dissolve metals from crushed PCB waste. The efficiency of the scavenger was

www.advsustainsys.com
determined by analyzing the metal content of the solution before and after it was passed through the filter. In all experiments, 8 mL of metal solution were passed through three separate 5-mm-thick cylinder-shaped 1-PP filters with a diameter of 16.5 mm. This was carried out by placing the tightly fitting filters into a 10 mL syringe and pressing the solution gently through them ( Figure S4, Supporting Information).
For the synthetic test solution, over 96% of Pd and 98% of Pt were captured from the solution with high selectivity. The most concentrated metal ions, such as aluminum, iron, and copper, were not retained by the filter while about 50% of Sn and 10-22% of Pb and Zn were initially adsorbed (Figure 2). However, practically, all of zinc and lead as well as most of tin and traces of other metals trapped inside the filters during experiments could be easily removed by subsequent water washes (six times 8 mL), and only a small amount of tin remained as the main impurity. Excellent efficiency and selectivity for capturing Pd and Pt was also obtained when an authentic WEEE leachate was used instead of the synthetic test solution. These results are summarized in Table S6 (Supporting Information).
By using a sequence of nylon filters and the 1-PP filters, the most valuable metals, i.e., Au, Pt, and Pd, could be captured selectively even from an authentic WEEE solution (Figure 3). In this process, the solution was first passed through gold scavenging filters (pure 3D printed nylon) followed by a set of 1-PP filters. Over 98% of Au could be removed from the authentic WEEE solution with three 5-mm-thick nylon filters with a diameter of 16.5 mm, and over 79% of Pd and 89% of Pt could be captured using three subsequent 1-PP hybrid filters.
Both Au and Pd/Pt scavengers are fully reusable as the recovered metals can be removed from the filters without destroying their ability to capture ions. Because Au and Pd/Pt were captured in discrete filters, the stripping could be performed separately to produce cleaner end products. Au was extracted from the nylon filter by using 7 m nitric acid as reported earlier. [10] Pd and Pt were removed from 1-PP hybrid filter one-by-one with thiourea solutions. Pd was extracted first with 0.1 m aqueous thiourea followed by extraction of Pt with 0.3 m aqueous thiourea solution (Table S2, Supporting Information). A total of about 83% and 85% of adsorbed Pd and Pt, respectively, were recovered with thiourea washes (see the Experimental Section). This process allowed separation and isolation of Pd and Pt as separate fractions. In the end, the residual Sn, also trapped by the 1-PP (see above), was removed as a separate fraction by using 4.5 m nitric acid. Unlike nylon scavenger for Au, which was ready to be reused immediately after removal of the captured metal, the 1-PP filters needed to be regenerated with 0.1 m hydrochloric acid before reuse. However, neither the stripping of the metals nor the regeneration of the filters caused significant reduction in adsorption efficiency (Table S3, Supporting Information).
Maximum adsorption capacity of the 1-PP filters was studied by passing 10 mL of solution containing 400 mg L −1 of Pd and Pt in 5% aqua regia through one 5 mm filter five times (Table S5, Supporting Information). Pd and Pt scavenging capacity of the filters was calculated to be about 11 wt% relative to the mass of the additive in the filter. Performance of the 3D printed filter was compared with a similar weight (520 mg) of powdered, non-printed mixture of type-1 anion exchanger resin (Dowex 21K) and polypropylene. The powder mixture was stirred in a 10 mL of concentrated Pd/Pt solution and the maximum capacity was found to be ≈13 wt%. The results show that printing decreases the maximum capacity of the resin only   www.advsustainsys.com slightly, again highlighting the high porosity of the objects and the availability of the chemically active component 1.
The presented results clearly demonstrate that SLS 3D printing can be utilized to fabricate porous but still rigid chemically functional objects. These objects can be printed by using a hybrid material such as 1-PP in which the chemically active component is printed together with easily printable supporting matrix. The results also show that the printing does not destroy the chemical functionality of the active component. The 1-PP filters introduced in this paper can be used as effective and selective scavengers for Pd and Pt from complex mixtures of metal ions such as acidic leachate of WEEE waste. The captured metals can be removed from the scavenger filter in a stepwise manner, enabling the separation of Pd and Pt as individual fractions. After removal of the metal ions, the objects can be regenerated, which makes them fully reusable.
Although we have used WEEE leachate as an example, the scavengers could also be used to capture metal ions from various other source materials. In a broader scope, the concept of using functional printing materials could be extended even beyond scavenging of ions. By developing new, chemically active printing materials, it is possible to open an entirely new chapter in design of chemically functional devices for various chemical processes.

Experimental Section
Chemicals and Solvents: Synthetic solutions used were prepared from PerkinElmer spectroscopy standards (1000 mg L −1 ). Polymers used for 3D printing were purchased from Advanc3d Materials. Dowex 21K, thiourea, nitric acid (≥65%), and hydrochloric acid (≥37%) were purchased from Merck. High-purity water of 18.2 MΩ cm resistivity was used throughout the experiments. Water was purified using Elga Purelab Ultra.
Sample Preparation: Synthetic solution containing 100 mg L −1 of Ni, Zn, Fe, and Cu and 50 mg L −1 of Al, Cr, Pb, and Sn along with 5 mg L −1 of Pd (and Au for a solution for synthetic test with gold; see Table S4 (Supporting Information)) and 1 mg L −1 of Pt in 5% aqua regia was prepared using PerkinElmer spectroscopy standards. Synthetic solution used for capacity tests containing 400 mg L −1 of Pd and Pt in 5% aqua regia was also prepared using PerkinElmer spectroscopy standards. The authentic WEEE solution was prepared by ashing milled printed circuit boards for 4 h at 950 °C. 1 g of the ashed WEEE was placed in a 50 mL centrifuge tube and 10 mL of aqua regia was added. Ultrasoundassisted leaching was conducted at 25 °C by performing six threeminute cycles. Pressure was released from the tubes between the cycles. Elma Elmasonic P was used as the ultrasound bath. Samples were filtered using Whatman 41 filter papers and diluted in 1:5 ratio before being used for adsorption tests. Detailed descriptions of HIM imaging, X-ray tomography, and inductively coupled plasma optical emission spectrometry (ICP-OES) measurements can be found from the Supporting Information.
Metal Recovery and Stripping: Recovery experiments were conducted by placing three 5-mm-thick and 16.5-mm-diameter tightly fitting filters in a 10 mL syringe. Before the adsorption tests, filters were prepared by passing 8 mL of water through them. Eight milliliter of the synthetic or authentic solution was then pushed through the filters and metal concentrations were measured before and after the adsorption experiment. To consider the amount of solution trapped in the objects, the final metal concentrations were compared to those of solution that had been passed through a syringe containing three 5-mm-thick and 16.5-mm-diameter pure PP filters. Hence, the recovery efficiency was calculated as the percentage difference in metal concentrations between initial solution (synthetic or authentic) passed through the pure PP filters and the concentration detected after 1-PP treatment. After water washes, recovery efficiency (Table S1, Supporting Information) was obtained by adding the concentrations present in water wash fraction to the metal concentrations left in the solution after the treatment with 1-PP and by comparing this value to the initial metal concentration of untreated synthetic solution.
Stripping tests were performed by using filters that had been used for recovery tests with synthetic solution. First, six washing cycles were performed using water (five times 8 mL), followed by four 8 mL cycles of 0.1 m thiourea. Next, four 8 mL washing cycles of 0.3 m thiourea were conducted, followed by four 8 mL cycles of 4.5 m nitric acid. Finally, filters were regenerated by passing through 16 mL of 0.1 m hydrochloric acid. All experiments presented in this paper were performed in triplicate.
Capacity experiment was conducted by placing one 5-mm-thick and 16.5-mm-diameter filter in 10 mL syringe and passing 10 mL of the synthetic solution containing 400 mg L −1 of Pd and Pt in 5% aqua regia through the filter for five times. Results of this experiment were compared to the result received by stirring the same weight (520 mg) of powder used for manufacturing the filters with 10 mL of the same synthetic solution for 30 min.
3D Printing: Filters were designed using FreeCad v. 0.16 and sliced by using Slic3r v. 1.2.9. Printing was done using Sharebot SnowWhite SLS 3D printer. A layer thickness of 0.1 mm was used for all 3D-printed objects. For PP and 1-PP, the used laser power was 40% with a rate of 60 000 (2400 mm s −1 ). Build plate temperature of 119-123 °C was used. For nylon, laser power of 40% with a rate of 64 000 (2560 mm s −1 ) was used. Build plate temperature was set to 160-163 °C for the duration of the print.

Supporting Information
Supporting Information is available at the Wiley Online Library or from the author.