Selective Recovery of Critical Minerals from Simulated Electronic Wastes Via Reaction-Diffusion Coupling

Abstract

Atom- and energy-efficient chemical separations are urgently needed to meet the surging demand for critical materials that has strained supply chains and threatened environmental damage. In this study, we used reaction-diffusion coupling to separate iron, neodymium, and dysprosium ions from model feedstocks of permanent magnets, which are typically found in electronic wastes. Feedstock solutions were placed in contact with a hydrogel loaded with potassium hydroxide and/or dibutyl phosphate, resulting in complex precipitation patterns as the various metal ions diffused into the reaction medium. Specifically, we observed the precipitation of up to 40 mM of iron from the feedstock, followed by the enrichment of 73 % dysprosium, and the extraction of >95 % neodymium product at a further distance from the solution-gel interface. We designed a series of experiments and simulations to determine the relevant ion diffusivities, D_Nd=5.4×10^-10 and D_Dy=5.1×10^-10 m²/s, and precipitation rates, k_Nd =1.0×10^-5 and k_Dy=5.0×10^-3 m⁹ mol^-3 s^-1, which enabled a numerical model to be established for predicting the distribution of products in the reaction medium. Our proof-of-concept study validates reaction-diffusion coupling as an effective and versatile approach for critical materials separations, without relying on ligands, membranes, resins, or other specialty chemicals.

Keywords: Critical elements; crystallization; interfaces; reaction-diffusion; separations.

Figure 1

Separation of Nd and Dy via reaction‐diffusion coupling. (A) Schematics of the experimental setup. (B) Photographs of the resulting precipitate patterns for individual Dy³⁺, Nd³⁺, and mixed salts, with border color coded with magenta, blue, and black, respectively. (C) Representative optical micrographs of precipitate crystals in the Dy‐only (magenta), Nd‐only (blue), and mixed‐salt (black) experiments. (D) Molar composition along the precipitate from a mixed‐salt experiment measured by ICP‐MS. (E−G) Time‐space plots of Dy‐only (E), Nd‐only (F), and mixed‐salt (G) experiments.

Figure 2

Building a numerical model for predicting the distribution of Nd and Dy in the reaction medium. (A) X‐ray images of NdCl₃ and DyCl₃ diffusion into agarose gel, with no precipitation reactions. (B, C) Evolution of concentration profiles from experimental (black) and best‐fit simulation (red) data for Nd (B) and Dy (C). (D) Images of reaction‐diffusion experiments from NdCl₃ and DyCl₃ single‐salt solutions diffusing into gel loaded with Kdbp. (E, F) Precipitation fronts acquired from experimental (gray) and best‐fit simulation (red) data for Nd (E) and Dy (F). (G) Time‐space plots showing precipitation fronts from mixed salt experiments. (H) Outer precipitation front and (I) inner precipitation front acquired from experimental (gray) and directly calculated simulation (red) data. The numbers in the legends of E,H, and I represent the concentration of Nd³⁺.

Figure 3

Reaction‐diffusion coupling with two precipitating agents in the gel. (A) Photograph of the precipitate pattern with mixed Nd³⁺ and Dy³⁺ in the aqueous solution and a mixture of dbp⁻ and OH⁻ in the gel. (B) Optical micrographs of precipitates at distances from the solution‐gel interface as indicated on top of the panels. All scale bars: 0.2 mm. (C−F) Intensity of averaged image cropped at different regions and spacing analysis for the large‐scale (C,E) and small‐scale (D,F) precipitate bands. The hollow circles mark the ratio of two consecutive bands and refer to the right‐side Y‐axes in E and F. (G) EDS maps of a representative precipitate particle acquired from the segment of 0–0.7 cm.

Figure 4

Reaction‐diffusion separations from feedstocks containing iron. (A) Photograph of the precipitate patterns for increasing Fe³⁺ concentrations. In these experiments, constant concentrations of Nd³⁺ (aq), Dy³⁺ (aq), dbp⁻ (gel), and OH⁻ (gel) were used, and the photograph was captured at t=5 days. (B) Measurements of the position, length, and gap of the leading band for each solution condition. (C) Optical micrographs of precipitates at the beginning, middle, and end of the gel labeled as segments (i), (ii), and (iii), respectively. (D) Powder XRD patterns of precipitates in different segments (solid lines) and pure precipitates for reference (dashed lines). (E) Correlation analysis of EDS maps for segment (i). (F) Molar composition along the precipitate from a mixed‐salt experiment measured by ICP‐MS. The Fe³⁺ concentration was 10 mM for C−F. (G) Molar composition at the end of the gel for different Fe³⁺ concentrations.