Recovering Rare Earth Elements from Aqueous ... - ACS Publications

Research Article www.acsami.org

Recovering Rare Earth Elements from Aqueous Solution with Porous Amine−Epoxy Networks Walter Christopher Wilfong,*,†,‡ Brian W. Kail,§ Tracy L. Bank,§ Bret H. Howard,† and McMahan L. Gray*,† †

U.S. Department of Energy, National Energy Technology Laboratory, 626 Cochrans Mill Road, Pittsburgh, Pennsylvania 15236, United States ‡ Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, Tennessee 37831, United States § AECOM, 626 Cochrans Mill Road, Pittsburgh, Pennsylvania 15236, United States S Supporting Information *

ABSTRACT: Recovering aqueous rare earth elements (REEs) from domestic water sources is one key strategy to diminish the U.S.’s foreign reliance of these precious commodities. Herein, we synthesized an array of porous, amine−epoxy monolith and particle REE recovery sorbents from different polyamine, namely tetraethylenepentamine, and diepoxide (E2), triepoxide (E3), and tetra-epoxide (E4) monomer combinations via a polymer-induced phase separation (PIPS) method. The polyamines provided −NH2 (primary amine) plus −NH (secondary amine) REE adsorption sites, which were partially reacted with C−O−C (epoxide) groups at different amine/epoxide ratios to precipitate porous materials that exhibited a wide range of apparent porosities and REE recoveries/ affinities. Specifically, polymer particles (ground monoliths) were tested for their recovery of La3+, Nd3+, Eu3+, Dy3+, and Yb3+ (Ln3+) species from ppm-level, model REE solutions (pH ≈ 2.4, 5.5, and 6.4) and a ppb-level, simulated acid mine drainage (AMD) solution (pH ≈ 2.6). Screening the sorbents revealed that E3/TEPA-88 (88% theoretical reaction of −NH2 plus −NH) recovered, overall, the highest percentage of Ln3+ species of all particles from model 100 ppm- and 500 ppm-concentrated REE solutions. Water swelling (monoliths) and ex situ, diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) (ground monoliths/particles) data revealed the high REE uptake by the optimized particles was facilitated by effective distribution of amine and hydroxyl groups within a porous, phase-separated polymer network. In situ DRIFTS results clarified that phase separation, in part, resulted from polymerization of the TEPA-E3 (N-N-diglycidyl-4-glycidyloxyaniline) species in the porogen via C−N bond formation, especially at higher temperatures. Most importantly, the E3/TEPA-88 material cyclically recovered >93% of ppb-level Ln3+ species from AMD solution in a recovery− strip−recovery scheme, highlighting the efficacy of these materials for practical applications. KEYWORDS: rare earth element, lanthanide, porous polymer, amine, infrared spectroscopy, water treatment

■

INTRODUCTION Rare earth elements (REEs) largely comprise the lanthanide series of metals and exhibit unique properties that make them ideal for applications such as catalysis, ceramics, battery alloys, magnets, and more. The U.S. Geological Society reported that the 2015 global REE reserves were 130 000 000 tons, with only 1.3% of those controlled by the U.S and a disproportionate 59% controlled by both Brazil (17%) and China (42%). In 2015, the U.S. mined only 4100 tons of REE’s (3.3% of those mined globally). This predicated the need to import the remaining 76% ($150,000,000) of the U.S. REE demand primarily from China, who produced about 85% of the global REEs.1 Despite both (i) the REE demand predicted to grow an average of 5− 9% annually over the next 25 years,2 and (ii) the recognized impact of foreign REE reliance on national security,3,4 the proven environmental hazards of mining activities5 have generated rigorous legislation that make it prohibitively costly to initiate and sustain new REE harvesting operations. © XXXX American Chemical Society

Therefore, it is imperative to develop technologies for domestically recovering and recycling spent REEs. Viable sources for REE recovery are natural water systems, which include rivers, groundwater, seawater, and acid mine drainage streams. The U.S. Department of Energy analyzed the aqueous REE concentration data for hundreds of samples reported in 71 studies of undisturbed water sources and found the following overall REE concentrations: groundwater, 5.7−410 pmol/kg; lake, 1.4−40 pmol/kg; river, 15−270 pmol/kg; seawater, 1.6− 13 pmol/kg (all ppt level concentrations).6 Mining activities have resulted in a drastic increase in some of these aqueous REE concentrations, up to a reported 7−59 ppb for acid mine drainage (AMD) contaminated streams,7,8 due to contact of newly exposed mineral formations to environmental REE Received: March 17, 2017 Accepted: May 12, 2017 Published: May 12, 2017 A

DOI: 10.1021/acsami.7b03859 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. General process for the formation of amine−epoxy monoliths and particles (TEPA-E3 shown), along with one hypothesized adsorbed REE (Ln3+) structure. Note, the remaining two chlorides (not shown) would be coordinated with water molecules.

coordinated to the Ln cation are displaced. The displacement of dimethyl sulfoxide (DMSO) solvent molecules from coordination spheres encompassing various polyamine (tetraethylenepentamine and others)−Ln complexes was observed in one study by infrared spectroscopy (IR).41 Regarding ligand−Ln binding in other work, nitrogen−La, Nd, Eu, Dy, and Yb coordination bonds for complexes comprised of ethylenediamine (ED) and various lanthanum chloride (−Cl−), bromide (−Br−), nitrate (−NO3−), and perchlorate (−ClO4−) salts were observed via FTIR by N−Ln infrared vibrational bands.42 Accompanying these IR bands were red shifts in the complex’s N−H stretching bands, and the generation of a high-frequency N−H stretching band, in some cases, that suggested the electrostatic repulsion of a coordinated counteranion by the amine group inside the ED−Ln coordination sphere. It was further found via IR that certain ethylenediamine−Ln complexes favored the coordinated form of the REE anions for smaller Ln sizes (up to Sm), whereas the ionic form was favored for larger Ln sizes. Overall, these studies highlighted some fundamental potential binding mechanisms of REEs to amine-based ligands, and provide a reference for understanding the potential interactions between amine−epoxy networks and REEs. In this work, we utilized the PIPS method to prepare an array of porous amine−epoxy monoliths and particles (ground monoliths) for the recovery of single-element and mixedelement REEs from both model solutions and simulated acid mine drainage (AMD) solution. Monoliths were prepared by reacting different polyamine monomers like tetraethylenepentamine (TEPA) with di- (E2), tri- (E3), and tetra- (E4) polyepoxide monomers at high temperature in the presence of different porogens/solvents. The amine/epoxide ratios where varied to achieve between Y = 50% and 150% (excess epoxy)

leaching conditions. Recovery of contaminant REEs from these different water sources could not only help alleviate the growing U.S. REE demand, but also mitigate the harmful effects of REE mining. Currently, three primary strategies are employed for preconcentrating, recovering, or separating aqueous REEs and heavy metals, and these involve (i) liquid−liquid extraction;9−14 (ii) membrane separation;15−19 and (iii) adsorption via ion exchange resins,20−23 functionalized organic/inorganic sorbents,20−22,24−30 and polymeric sorbents.31−35 Liquid−liquid extraction typically utilizes expensive ionic liquids for REE capture and a batch centrifugation process for separating off the REE-absorbed extractant, while a mixed system of cationic/ anionic ion exchange resins also generally suffers from complicated regeneration schemes. One promising group of porous polymer sorbents are those derived from amine− epoxide reactions conducted in a porogen, where a strong and chemically resistant polymer network is precipitated out via C− N bond formation. This polymerization-induced phase separation (PIPS) process reportedly results from spinodal and binodal decomposition of the amine/epoxy/porogen system as the monomers polymerize into higher molecular weight species.36 PIPS was previously employed to create millimeter-scale, cylindrical amine−epoxy polymer monoliths used in preconcentrating metals from aqueous sources for analytical purposes.37−39 Typically for metal capture, the interaction of hydrated REEs/lanthanide cations, [Ln(H2O)n]+3, with commonly used ligand amine functional groups occurs through coordination. Here, the lone electron pairs on the nitrogen atoms are donated to vacant valence electron orbitals of Ln3+ species, presumably excluding the f-orbitals,40 to generate ligand−Ln bonds. While forming the amine−Ln complex, solvent molecules previously B


Research Article


sample charging. The particle and pore sizes were measured from SEM images, using MS Visio software. Nitrogen physisorption measurements (Quantachrome, NOVA 2200) at 77 K were further performed on the particles to assess their BET surface areas (SABET), BJH pore volumes (VBJH), and pore structures. Chemical Structure Analysis. An in situ, diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS, ThermoScientific) study of the cross-linking reactions within a model polyepoxide/ polyamine/PEG400 mixture was conducted to provide a fundamental understanding of the formation of the optimum PIPS sorbent. About 6 uL of a solution containing 20 wt % of the E3/TEPA-88/PEG400 mixture in MeOH was first injected onto an aluminum cup set on the DRIFTS sample stage, followed by attaching the DRIFTS dome and flowing N2 over the sample at 40 °C for 60 min to evaporate the MeOH. Once MeOH was evaporated, the E3/TEPA-88/PEG400 film was deposited onto the cup. Subsequently, the cup and the film were heated/reacted using a temperature profile near identical to that experienced by the oven-heated mixtures within the sealed vials, as seen in Figure S1. The DRIFTS film was heated in four segments, which were (i) 40−76 °C at 2.7 °C/min; (ii) 76−90 °C at 1.5 °C/min, (iii) 90−105 °C at 0.5 °C/min; and (iv) held at 105 °C for about 2 h, followed by cooling to 40 °C. Single-beam spectra of the film were averaged from 25 coadded scans collected over 15 s, with a resolution of 4 cm−1. Ex situ DRIFTS analysis of the amine−epoxy particles was also performed to understand the particles’ chemical structure−property relationships. Prior to scanning, the particles were pretreated at 105 °C for 15 min under N2 flow to remove any H2O and CO2 preadsorbed from the ambient air, and were then cooled to 50 °C. Rare Earth Element (REE) Adsorption Screening. Roomtemperature batch adsorptions of single or near equimolar-mixed La3+ (LaCl3 anhydrous, >99.99% trace metals basis), Nd3+ (NdCl3 hexahydrate, 99.9% trace metals basis), Eu3+ (EuCl3 hexahydrate, 99.9% trace metals basis), Dy3+ (DyCl3 hexahydratye, >99.99% trace metals basis), and Yb3+ (YbCl3 hexahydrate, 99.9% trace metals basis) rare earth elements were each accomplished by soaking about 0.3 g of monolith chunks or particles in 12.0 g of aqueous 100 ppm (single) or ∼0.72 mmol/L total (∼0.14 mmol/L each of five mixed REEs) solutions (natural pH ≈ 5.5 ± 0.3) for 40−45 min in a polystyrene vial, which was occasionally shaken during adsorption. After adsorption, the treated REE solutions were separated from the monoliths by decantation and from the particles by syringe filtration, and then their metal concentrations were analyzed by inductively coupled plasma mass spectrometry (ICP-MS, see the Supporting Information) for remaining REE content. The percentages of REE uptake by the materials were calculated from the REE concentrations in the treated solutions and in the controls of either the fresh stock solutions (monoliths) or the filtered stock solutions (particles). Relative maximum, single-element REE uptake capacities (wt %) of key particle sorbents were determined by soaking the particles in 500 ppm Ln3+ solutions for 4 h, filtering the particles, and then analyzing the treated solutions. REE recovery was also performed at pH ≈ 6.4 (±0.3) and ∼2.5 (± 0.1) to further assess the amine−epoxy particle sorbents’ performances under different conditions. Each of 0.2 M NaOH and 1.0 M HCl solutions were used to adjust the solution pH values, which were first checked with pH paper (0.2 or 0.5 pH resolutions) and later assessed by pH probe after 1−2 days. The actual Ln3+ concentrations and probe pH values for all solutions are shown in Table S1. The ICP-MS precision of Ln3+ detection in ∼100 and ∼500 ppm solutions was calculated to be typically between 0.07% and 1.9%, with limits of detection for these species generally between 0.1 and 18.2 ppb. The detection precision for ppb-level Ln3+ species in simulated acid mine drainage (AMD) solution was 1.4−3.1%, with limits of detection between 3 and 10 ppt. REE Recover−Strip−Recover Cycle Testing. An REE recover (R1)−strip (S)−recover (R2) scheme was performed on optimized amine−epoxy particle sorbents to assess their application in practical systems. First, about 0.15 g of particles was loaded on top of a ceramic frit/filter set inside of a closed flow-through column (batch mode). Approximately 6 g of the 0.72 mmol/L 5-REE mixture (pH ≈ 6.4) or a

theoretical reaction of the initial primary (−NH2) and secondary (−NH) amines with the epoxides. This yielded different extents of phase separation and porosities (water uptakes) for the monoliths. General polymerization mechanisms responsible for monolith formation were elucidated by in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS). Ex situ DRIFTS of the E3/TEPA-Y series of particles assisted in explaining their differences in apparent porosity and REE uptake. Screening of the monolith particles by adsorbing REEs at different pH values (∼2.4 to ∼6.4) from 100 ppm Ln3+ solutions revealed that E3/TEPA-88 was the optimum sorbent. The efficacy of this sorbent for REE recovery under practical conditions was confirmed by its cyclic capture of >93% of all ppb-level REEs from simulated, acidic AMD solution.

■

EXPERIMENTAL SECTION

Amine−Epoxy Monolith and Particle Preparation. All reagents were purchased from Sigma-Aldrich. An array of amine− epoxy monoliths were prepared by first dissolving 3.5−4.25 g of D.E.R 332 (diepoxide, E2; 5.7−5.9 mmol epoxide/g), 2.75−3.95 g of N-Ndiglycidyl-4-glycidyloxyaniline (triepoxide, E3; 10.8 mmol epoxide/g), or 2.95−4.05 g of 4,4′-methylenebis(N,N-diglycidylaniline) (tetraepoxide, E4; 9.5 mmol epoxide/g) polyepoxide monomers into 40.0 g of poly(ethylene glycol) (PEG, MWavg = 200 or 400) porogens in 100 mL beakers. PEG200 failed to produce phase-separated monoliths, and so PEG400 was used for all materials described in this work. Once the polyepoxides were dissolved, 0.64−2.22 g of tetraethylenepentamine (TEPA, technical grade; 26.4 mmol of −NH plus −NH2/g) or 0.96− 2.86 g of branched poly(ethylenimine) (PEI, MW ≈ 800/Mn ≈ 600; ∼16.2 mmol of −NH plus −NH2/g) polyamine monomers was added to the beakers containing the polyepoxide plus PEG mixtures (50 g total). These mixtures were portioned into five 22 mL glass vials (10 g each), which were sealed and then reacted in an oven at 105 °C for 3 h. The amounts of amine and polyepoxide were adjusted to give between 50% and 150% (excess epoxy) theoretical reaction of the amines, and the total weight fraction of monomers in PEG was varied between 0.10 and 0.30. Reaction of the polyamines with polyepoxides in the presence of PEG formed either viscous liquids, gels, or “waxes”, depending upon the extents of reaction and phase separation of the porous network during the PIPS process. After reacting, each vial was broken, and its contents were removed and cut (if solid) into eight equal-sized chunks. The chunks were soaked in 200 mL of reverseosmosis-treated (RO) water for 24 h followed by two changes of 200 mL of MeOH for 24 h. Soaking the samples removed the porogen and formed templated gels (wet monoliths), which were then dried at 25 °C under vacuum for between 6 and 17 h to produce dry solid monoliths. Portions of the cut monoliths were reserved and further ground into particles with a mortar and pestel or spatula, washed in 200 mL of MeOH for 4−24 h, filtered, and then vacuum-dried for 6−8 h. Samples were labeled with the scheme, “EX/polyamine-Y”, where “X” represented the number of epoxide groups in the epoxy monomer (2, 3, or 4), “polyamine” is the name of the amine monomer used, and “Y” is the theoretical percent reaction of the amines (−NH2 plus −NH) with the epoxides. The overall scheme for the formation of the porous materials and the hypothetical general mechanism for their REE adsorption is further clarified in Figure 1. Pore Structure and Morphology Analysis. The apparent porosity and success of the phase separation of the amine−epoxy monoliths were ascertained by swelling the monoliths in RO water for 45 min and then measuring the wt % of H2O uptake. Typical particle sizes, visible pore sizes, and overall morphologies of the monoliths and particles were investigated by scanning electron microscopy (SEM, FEI Co. Quanta 600), which was equipped with secondary and backscatter electron detectors. The samples were mounted on conductive tape adhered to aluminum planchets, coated with a Pd/ Au mixture, then analyzed at 10 kV in low vacuum mode to reduce C


Research Article


Figure 2. Swelling profiles of (a) E3/TEPA-Y and (b) E4/PEI-Y monoliths after soaking for 45 min in RO water. The monomer weight fraction for the samples was 0.10. (c) RO water swelling profile for the E3/TEPA-88 monolith prepared with 0.10−0.30 weight fractions of monomers in solution.

Figure 3. SEM images of E3/TEPA-88 and E4/TEPA-88 monoliths, (a) and (d), respectively, plus particles (b) and (e), produced by grinding the monoliths and then washing with MeOH. (c and f) BET N2 adsorption−desorption isotherms with BJH method pore size distributions. Particles were pretreated at 105 °C under vacuum for 2 h prior to BET analysis.

■

simulated acid mine drainage (AMD) solution (pH = 2.6) was mixed with the particles inside the column, which was sealed with para-film for batch REE adsorption for 40 min (R1). The exact composition of the AMD solution can be found in Table S2. Briefly, the solution contained 4−26 ppb each of La3+, Nd3+, Eu3+, Dy3+, and Yb3+ (REEs, 57 ppb total) plus 14−283 ppm each of Mn, Al, Fe, Mg, Ca, and Na cations (potential fouling metals, 689 ppm total). After adsorption, the column outlet was opened, and slightly pressurized air was pumped into the system to purge out treated REE solution, which was collected for ICP-MS analysis. Once the treated solution was purged, the column outlet valve was again closed, and a pH = 8.5 citrate-based buffer solution was mixed with the REE-adsorbed particles for simultaneous REE stripping plus sorbent regeneration for 40 min (S). Further details of the buffer cannot be revealed due to the filing of a patent application. After the REE strip/sorbent regeneration, the buffer was purged out and collected for analysis. The column was then again set in batch mode, and another 40 min of REE recovery was performed (R2).

RESULTS AND DISCUSSION

Monolith Swelling and La3+ Recovery Screening. Figure 2a and b shows the 45 min, RO water swelling behaviors of monoliths versus the theoretical percent reaction of amine groups with epoxide groups, Y, for materials prepared with (a) E2, E3, and E4/TEPA-Y and (b) E2, E3, and E4/PEI-Y compositions at a monomer (polyamine plus polypoxide) weight fraction of 0.10. The reaction of TEPA and PEI polyamines with E2, E3, and E4 polyepoxide reactants in the PEG porogen yielded cross-linked/polymerized products with consistencies ranging from translucent viscous solutions or gels to opaque white “waxes”. Overall, the swelling profiles of monoliths prepared with both the N-N-diglycidyl-4-glycidyloxyaniline triepoxide (E3) and 4,4′-methylenebis(N,N-diglycidylaniline) tetra-epoxide (E4) species exhibited trends of rising and then falling percent swelling (apparent porosity) with increased Y values. An exception to this trend was for E3/PEI-Y D


Research Article


available amine (and possibly hydroxyl) sites and dissolved REE species, especially for E3/TEPA-88. Figure S2 shows the pH ≈ 5.5, 100 ppm La3+ adsorption screening of (a) E3/TEPA-Y and (b) E4/TEPA-Y (Y = 66− 125%) intact monoliths and ground monolith particles. Results confirmed that the La3+ cation was best adsorbed by the particle morphology, which was further used to test all amine−epoxy formulations. REE Recovery with Amine−Epoxy Particles. Because of the discovered slight acidity of RO water and the presence of ppb-level metal contaminants, ground particles used for REE tests were prepared from monoliths that were washed in only MeOH. Figure 4 shows the recovery of single element La3+,

monoliths, which displayed gradually increasing percent swelling after the sharp increase in swelling from Y = 88% and Y = 100%. Swelling values reached maximums of 892% and 430% for E3/TEPA-88 and E4/TEPA-88, respectively, and 933% and 235% for E3/PEI-88 and E4/PEI-150, respectively. The rough-textured, rigid white structures of these monoliths (inset) corresponded to their high swelling/porosity, and reflect effective phase separation of the polymerized amine− epoxy network into a porous structure that could recover REEs. Interestingly, the trend in swelling for the final dry materials was accompanied by variances in these materials’ consistencies after reacting in the oven. This means that translucent, medium viscosity gels were observed at lower and higher Y values, while opaque white, rigid waxes were observed at optimum Y values. Tsujioka and co-workers also observed varying consistencies and porosities in their amine−epoxy reaction products by adjusting the amine/epoxide stoichiometric ratio, and other parameters.43 The authors attributed transformation of liquid monomer solutions into transparent gels to the generation of 3D cross-linked networks at an amine/epoxide stoichiometric ratio of 1. Additional precipitation of the mixture into an opaque white mass via spinodal or bimodal decomposition signified macroscopic phase separation of the 3D network.36 We ascribe either of these mechanisms of phase separation to the precipitation of our monoliths. In contrast to E3- and E4-based monoliths, E2-based materials exhibited negligible swelling at all Y values. The minimal porosity of these materials coincided with the translucent, “rock-like” textures of their dried products, and showed ineffective phase separation into a porous network. Figure 2c shows that increasing the fraction of monomers from 0.10 to 0.30 in the E3/TEPA-88/PEG mixture exponentially diminished the water swelling of the solid monoliths. These reduced apparent porosities were expectedly accompanied by transitions from opaque white to translucent yellow, and revealed that a 0.10 monomer fraction was required to form porous amine−epoxy networks. Figure 3 shows the SEM images of the most porous (a) E3/ TEPA-88 and (d) E4/TEPA-88 monoliths and their corresponding particles, (b) and (e), respectively. The structures of both morphologies were composed of interconnected, spherelike globules measuring between 35 and 75 nm in diameter. More agglomeration of the globules for the E4- than the E3based monolith indicates less effective phase separation of the former, which directly led to its lower water uptake. The higher water uptake for E3/TEPA-88 was accompanied by visibly smaller mesopores and macropores, with diameters between 17 and 80 nm (10 pores measured by MS Visio software, average = 41 nm), relative to those for E4/TEPA-88 between 14 and 185 nm with an average of 71 nm. These comparative features between the monoliths translated into similar differences for their ground particles (3−230 μm particle sizes) in Figure 3b and e. BET analysis of E3/TEPA-88 particles revealed a bimodal pore structure, with diameters from 3.2 to 8.6 nm and from 8.6 to 85.2 nm, where the latter range was observed by SEM. SABET and VBJH values were 54 m2/g and 0.30 cm3/g, respectively. Predominance of a monomodal pore size distribution, 3.3−11 nm, for E4/TEPA-88 corresponded to its lower SABET of 24 m2/g and VBJH of 0.08 cm3/g as compared to those for E3/ TEPA-88. These SEM and BET data illustrate that the interconnection of smaller mesopores by larger mesopore and macropore channels should facilitate good contact between

Figure 4. REE recovery from pH ≈ 5.5, single-element 100 ppm Ln3+ solutions by (a) E3/TEPA-Y and (b) E4/TEPA-Y ground and washed monolith particles. Y represents the theoretical percent reaction of the primary (−NH2) and secondary (−NH) amine groups with epoxide groups.

Nd3+, Eu3+, Dy3+, and Yb3+ from pH ≈ 5.5, 100 ppm solutions by (a) E3/TEPA-Y and (b) E4/TEPA-Y (Y = 66−100) particles. Because TEPA has a more well-defined structure, lower cost, and lower viscosity than those for PEI, the TEPA/ epoxy particles are more desirable than those of PEI/epoxy for practical application and were selected for further testing. The overall Ln3+ recovery profiles exhibited by the series of E3/ TEPA-Y particle sorbents paralleled the increasing−decreasing percent swelling profiles of the particles’ intact monolith counterparts, regarding Y values. This suggests that grinding the monoliths into smaller micrometer-size particles shortened the REE diffusion path and better exposed sorbent amine sites to aqueous Ln3+ species. E3/TEPA-88 displayed the highest Ln3+ recoveries of all E3/TEPA-Y particles, with La = 65.5%, Nd = 83.1%, Eu = 58.8%, Dy = 69.7%, and Yb = 97.4%. Although the percent recovery of Yb3+ was consistently higher than that for La3+ at all Y values, no consistent trend in the percent recovery of the other Ln3+ species was observed. E


Research Article


of these species at higher pH values, reaching up to >99% recovery of all metals at pH ≈ 6.4, except for 92% recovery of La by E3/TEPA-88. Interestingly, Yb3+ recovery was minimally affected by pH and was between 97 and >99% in all instances, except for ∼80% at pH ≈ 5.5 by E4/TEPA-88. We attribute the lowest recovery of REEs at pH ≈ 2.5 to the protonation of amine sites into ammonium ions (−NH3+/−NH2+), which would repel cationic Ln3+ species. This proposed ammonium ion formation was inferred elsewhere by zeta potential measurements of an L-cystine/iron oxide nanoparticle REE sorbent dispersed in acidic solution,45 and was directly observed by infrared spectroscopy after contacting a tetraethylenepentamine/silica CO2 gas sorbent with HCl vapor.48 The trend of increasing percent REE recovery with pH up to pH ≈ 6.4 for our amine−epoxy monoliths was similarly observed for La3+ and Nd3+ recovery by the L-cystine/iron oxide nanoparticle sorbent up to a pH of 7. The authors attributed this phenomenon to deprotonation of carboxylic acid groups (−COOH) into anionic carboxylate groups (−COO−) upon exceeding their point of zero charge pH value (pH = 5.1), where these anionic carboxylates exhibited a strong affinity for cationic Ln3+ species. Although zeta potential values for our fresh materials are unknown, pKa values >9 for typical −NH, −NH2, and −OH functional groups that are inherent in our networks indicate that deprotonated species should not be formed at pH ≈ 5.5 and 6.4. Therefore, better REE capture at the higher pH values partly reflects fewer ammonium ion species within the particle networks. Additionally, hydration of Ln3+ at higher pH might drop the pKa value of [Ln(H2O)n)]3+ species to near the pH value of the REE solution (∼6.4 here). Coordination of water to complexed metal atoms was shown to reduce the pKa of H2O from about 15.7 to 8.3 or 7.2, depending upon the ligands involved in the complex.49,50 We speculate that deprotonation of larger H2O molecules in the [Ln(H2O)n)]3+ hydrated complex would form smaller, bound −OH molecules in the form of [Ln(OH)(H2O)n−1]2+ moieties.47 The accompanying reduction in the formal charge of the hydrated species from 3+ to 2+ could weaken the interactions between lanthanide cations and other coordinated H2O molecules in the hydration sphere. These weakened bonds could cause less basic H2O to be easily displaced from the complex by more basic −NH and −NH2 groups of the amine−epoxy network to form TEPA−Ln coordination bonds. Noteworthy and excluding Nd3+, the trend of decreasing percent Ln3+ recovery with decreasing ionic radii from La3+ (1.250 Å) to Eu3+ (1.120 Å) and then increasing percent recovery from Eu3+ to Yb3+ (1.010 Å) at pH ≈ 2.5 and ∼5.5 by both E3/TEPA-88 and E4/TEPA-88 qualitatively resembles the trend for these LnCl3 solubilities in H2O.51 This correlation shows that the tendency for basic amines to coordinate with acidic Ln3+ species follows as the affinity of water toward these cations. Overall, REE recovery by our amine−epoxy particles is primarily driven by complex acid− base reactions. The relative maximum REE uptakes, in wt %, of all five Ln3+ species from highly concentrated 500 ppm solutions at pH ≈ 5.5 and ∼6.4 by E3/TEPA-88 and E4/TEPA-88 particles shown in Figure 6a and b, respectively, were determined after 4 h of batch adsorption. This was to assess the particle formulation with the best REE recovery potential. The data revealed that both sorbents captured similar amounts of Ln3+

This apparent inconsistency could be related to the inherently different −NH/Ln3+ and −NH2/Ln3+ ratios plus the variations in the amounts of electron-donating −OH groups produced by varied extents of the amine−epoxide reactions. E4/TEPA-Y particles in Figure 4b displayed trends in their percent recovery of Ln3+ species somewhat similar to those for E3/TEPA-Y particles, except for higher recovery values of Nd3+, Eu3+, and Dy3+ for E4/TEPA-66 than for E4/TEPA-88. To identify the better sorbent, relative maximum uptakes of each single REE by these particles from pH ≈ 5.5, 500 ppm Ln3+ solutions were assessed. Yet prior to ICP-MS analysis of the Nd3+, Eu3+, and Dy3+ solutions treated by E4/TEPA-66, we discovered that white precipitates accumulated at the bottom of the vials. These precipitates likely arose from the complex formation between aqueous Ln3+ species and unreacted/free amines, which were leached from the insufficiently reacted Y66% particles during testing. These results contrast with the clear solutions observed after REE recovery by the E4/TEPA88 particles, and signify that E4/TEPA-88 particles are more stable than those of E4/TEPA-66 for practical application. Figure 5 shows the effect of pH on the REE recovery of (a) E3/TEPA-88 and (b) E4/TEPA-88 particles, where the starting

Figure 5. REE recovery from single-element 100 ppm Ln3+ solutions at pH ≈ 2.5, 5.5, and 6.4 by (a) E3/TEPA-88 and (b) E4/TEPA-88 particles.

pH ≈ 5.5 solutions were adjusted to pH ≈ 2.5 using 1.0 M HCl and to pH ≈ 6.4 using 0.2 M NaOH. Previous works showed that the pH of REE or heavy metal solutions can affect the metal uptake of sorbents by altering the surface charge of the sorbent’s functional groups,44 such as amines.45,46 In parallel, varying the solution pH can affect the size of the hydrated cations, [Ln(H 2 O) n ] 3+ , in relation to the number of coordinated water molecules, n.47 The adsorption of La3+, Nd3+, Eu3+, and Dy3+ by E3/TEPA-88 and E4/TEPA-88 in our work almost consistently demonstrated better percent recovery F


Research Article


behavior to both easier complexation of smaller ions by the ligands, and greater penetration of smaller Ln3+ species into the cross-linked ligand network. We speculate that the higher Ln3+ capacities of our E3 (triepoxy)-based particles as compared to those of the E4 (tetra-epoxy)-based particles resulted from both faster mass transfer of REE solution (and Ln3+ species) through a more flexible polymerized network facilitated by the E3 monomer, and more accessible amine sites as suggested by the SABET values. These results are further supported by the swelling results of the corresponding intact monoliths and VBJH values of the particles. Comparatively, the maximum Ln3+ uptakes of our materials here are within the range of those reported in the literature, as shown in Table S3, and highlight the competitive performance of the amine−epoxy particles. Additional testing of E3/TEPA-88 and E4/TEPA-88 for their REE recovery from 0.52 to 0.72 mmol/L near-equimolar Ln3+ mixtures at pH ≈ 5.5 and ∼6.4 (see the Supporting Information) was performed to further determine the optimum particle formulation. The results, in parallel with those for REE recovery from the 100 and 500 ppm single-element solutions, confirmed that E3/TEPA-88 was the optimum sorbent for subsequent testing under practical conditions. Practical Testing of E3/TEPA-88 with AMD. Practical application of REE sorbents demands their recyclability, where previously adsorbed Ln3+ species are subsequently stripped and the sorbent is regenerated to facilitate extended use. Figure 7a shows the cyclic performance of the E3/TEPA-88 particles during a three-segment, recover (R1)−strip (S)− recover (R2) scheme. The particles were exposed to the pH ≈ 6.4, mixed Ln3+ solution for REE recovery (R1 and R2), and exposed to a pH = 8.5, citrate-based buffer solution for REE stripping/sorbent regeneration (S). A sodium citrate solution was similarly used in a single-step strip/regenerate scheme to both remove previously recovered REE species from bacterial cells and restore the cells active REE capture sites.52 The authors referenced that citrate should form strong complexes with the REEs. We expect that our citrate-based buffer steals REEs previously adsorbed to the amines of our sorbent by forming more stable ligand-REE bonds. Our results revealed that between 79% and 94% of REE species recovered after 40 min (R1, >98%) were stripped by a 40 min rinse in the buffer solution. Note, both the REE (R1) and the buffer (S) solutions were purged from the column before the subsequent (S) and

Figure 6. Relative maximum, 4-h batch REE uptakes from pH ≈ 5.5 and pH ≈ 6.4, 500 ppm single-element Ln3+ solutions by (a) E3/ TEPA-88 and (b) E4/TEPA-88 particles.

species at their optimum pH ≈ 6.4, with about 1−1.3 wt % La3+ and 1.7−1.9 wt % each of Nd3+, Eu3+, Dy3+, and Yb3+. In contrast, E3/TEPA-88 exhibited higher REE uptake than E4/TEPA-88 at lower pH ≈ 5.5 for all Ln3+ species, where capacity of the former gradually increased with decreasing ionic radii, from ∼1.1 wt % for La to ∼2.7 wt % for Yb. These profiles for wt % REE uptake were mirrored by those of the Ln3+/−N molar ratios, as shown in Figure S3, which were all less than 0.02. Overall these single-element capacities of the amine−epoxy particles followed a trend regarding Ln3+ size similar to that observed for multielement uptake (recovery from mixed REE solution) by other carboxylic acid-functionalized sorbents.44 The authors attributed this adsorption

Figure 7. Cyclic performance of the E3/TEPA-88 particle sorbent during the recover (R1)−strip (S)−recover (R2) scheme, with REE recovery/ capture from (a) pH ≈ 6.4, mixed Ln3+ solution and (b) pH = 2.6, simulated acid mine drainage solution (Ln3+ = 57 ppb, fouling metals = 689 ppm), and with REE stripping/sorbent regeneration with pH = 8.5, citrate-based buffer. G


Research Article

ACS Applied Materials & Interfaces (R2) segments, respectively. Because of the incomplete stripping of previously recovered Ln3+ species from the amines, some sites remained occupied by the adsorbed metals and thus decreased the percent recovery of the metals from >98% in R1 to between 86% (Yb) and 95% (La) in R2. Because we recovered nearly all REEs from the lower-concentration mixed Ln3+ solution at pH ≈ 6.4 by E3/TEPA-88 after only 40 min, we performed a 4-h recovery from a higher-concentration 20 mmol/L mixed-element solution to assess individual REE selectivities of the sorbent. Figure S5a expectedly shows higher mmol Ln3+/g uptake by the particles for smaller REEs, which corresponded to overall higher selectivities (distribution coefficient-based, Kd Ln3+,1/Kd Ln3+,2) in Figure S5b toward Nd3+, Eu3+, Dy3+, and Yb3+ over La3+ and toward Yb3+ over Nd3+ and Eu3+. Ultimately, the efficacy of E3/TEPA-88 toward REE recovery was confirmed by conducting the (R1)−(S)−(R2) scheme with simulated AMD.53 Critically important, Figure 7b shows that the particles recovered 98.3% of the total REEs, with between 99.0% and 99.9% of La3+, Nd3+, Dy3+, and Yb3+ plus 70.6% of Eu3+, during R1 despite the roughly 12 000/1 ratio of potential fouling metals/REEs. These particles also had strong affinities for Al3+ and Fe3+, which could reflect a tendency for basic amines to complex with other acidic trivalent metals similarly as the Ln3+ of REE salts. Small or 0% recovery of Na+, Ca2+, and Mg2+, all of which have larger ionic radii (0.86−1.02 pm) and smaller positive charges than Al3+ and Fe3+ (ionic radii = 0.50− 0.60 pm),54 suggests that their lower Lewis acidity could diminish their attractiveness toward basic amines. Most importantly, a high selectivity of Kd Ln3+,total/Kd fouling metals = 86.5 demonstrates the sorbent’s preference for valuable REEs over unwanted contaminants, even in the presence of a mixture of anions (inherently sulfates and chlorides). Furthermore, the E3/TEPA-88 recovered 93.5% of the total REEs, with between 91% and 95% of La3+, Nd3+, Eu3+, and Dy3+, and ∼69.7% of Yb3+ in R2, despite achieving only 60−70% stripping of the bound REEs and 48−57% stripping of the potential fouling metals in the (S) segment. The greater than 100% recovery of Eu and Yb, especially Yb, in the (S) segment is unclear and is under investigation. Overall, these promising findings show that amine−epoxy particles are good candidates for cyclic REE recovery from practical aqueous systems, and yet highlight the need to optimize the stripping/regeneration solution to facilitate better sorbent stability and selectively release the fouling metals and REEs in different segments. DRIFTS Analysis. Figure 8 shows the IR absorbance spectra (absorbance = log(I0/I)) of the E3/TEPA-88/PEG400 film at key temperatures, where I0 is the normalized (0−1) IR single beam spectrum of the film at 40 °C after MeOH evaporation (before reacting), and I is the normalized single beam spectra of the film at different reaction temperatures. From 40 to 91 °C, gradual reductions in the N−H stretching band intensities were observed largely for the primary amines (−NH2) at 3354 and 3295 cm−1 and in part for the secondary amines (−NH) at 3295 cm−1, overlapped with −NH2 species. The decreased N− H stretching band intensities of TEPA were accompanied by a corresponding loss in the primary amine N−H bending band intensity at 1598 cm−1 and by diminutions in the E3 epoxide C−O−C band intensities at 1240 and 900 cm−1, all of which signified the polymerization reaction between TEPA and E3. This reaction (largely from −NH2 species) produced a strong C−N−C stretching band at 1178 cm−1 indicative of the monolith’s covalent E3−TEPA−E3 network, and produced a

Figure 8. In situ DRIFTS absorbance spectra collected during the reaction of an E3/TEPA-88/PEG400 film. Absorbance = log(I0/I), where I0 is the normalized (0−1) single beam spectrum of the film deposited after evaporating MeOH at 40 °C for 1 h, and I are the normalized single beam spectra of the film at different temperatures/ times.

strong, narrow O−H stretching band at 3588 cm−1 for new hydroxyls that were attached to the monolith network. Comparison of the IR spectra for various aliphatic and branched amines revealed that primary amines typically display C−N stretching bands between 1023 and 1090 cm−1, whereas C−N stretching bands for secondary amines are often between 1133 and 1190 cm−1.55 Therefore, we propose that the new 1178 cm−1 C−N stretching band represents rigid, secondary amine species that are locked into the polymer network. Note, the new C−N band is at a higher wavenumber than those for pure TEPA’s secondary amines56 (1125 cm−1) and primary amines (1070 cm−1), and further supports a robust polymer network. We speculate that the additional 1207 cm−1 band could reflect C−N stretching of tertiary amines resulting from the reaction of TEPA’s −NH groups with E3’s epoxide groups. It was shown that N,N-bis(3,5-dimethyl-2hydroxybenzyl)methylamine displayed, in part, a strong band at 1207 cm−1 that was attributed to C−N stretching of its tertiary amine group.57 Beginning from 94 °C and up to 105 °C, both the positive C−N and O−H stretching IR bands of the monolith’s network and the negative −NH plus −NH2 (TEPA) and C−O−C (E3) IR bands of the monomers continued to intensify. An increasing 3295 (−NH)/3354 (−NH2) negative IR band intensity ratio was paralleled by the generation of a new O−H stretching band at 3550 cm−1, and indicated that a greater proportion of TEPA’s secondary amine groups relative to primary amine groups reacted with the epoxides at higher temperatures. One study examining amine/epoxide reaction kinetics showed that the ratio of the secondary amine (k2)/ primary amine (k2) reaction rate constants for aromatic trimethylene glycol di-p-diaminobenzoate increased from 0.158 at 100 °C to 0.270 at 130 °C.58 We infer that an increase in TEPA’s −NH/−NH2 reaction rate constants for our system could occur between 41 and 105 °C, given the aliphatic H


Research Article


Figure 9. (a) IR absorbance spectra (absorbance = log(1/I)) of E3/TEPA-Y ground and washed monolith particles, TEPA/KBr, and E3/KBr at 50 °C after pretreating at 105 °C in flowing N2 for 15 min and then cooling. (b) IR absorbance intensity ratio profiles, corrected for different E3/TEPA molar ratios (see the Supporting Information).

structure of TEPA. Interestingly, at 94 °C we observed a dramatic reduction in the IR intensity at 3166 cm−1. This band was previously assigned to an overtone vibration of amine groups,59 and can be intensified upon hydrogen bonding of TEPA’s −NH and −NH2 groups with electron-donating O−H groups of PEG200.60 We therefore postulate that as the precipitated monolith is formed, TEPA previously interacting with PEG400 in the E3/TEPA-88/PEG400 film was removed from the liquid phase, thereby restricting contact between amine and hydroxyl groups. The strong IR features observed here at 105 °C for the film-based monolith, especially those of the 3166 cm−1 band, correlated to the precipitation of white monolith structures in both the DRIFTS aluminum cup and the oven-heated vial after the reactions were completed. Because a 40 °C background single-beam spectrum was used to generate absorbance spectra at different temperatures, potential temperature-dependent variations in the background of the IR absorbance spectra could influence their interpretation. However, upon cooling the monolith film from 105 to 40 °C, we observed the same key features: (i) negative −NH/− NH2 and C−O−C bands, along with a stronger negative intensity at 3354 cm−1 than at 3295 cm−1; (ii) positive C−N and O−H bands; and (iii) a negative amine···O−H band at 3166 cm−1. Interestingly, cooling the monolith dramatically altered its hydrogen-bonded structure. This was evidenced by the generation of a broad O−H stretching band centered at 3398 cm−1 in parallel with the red-shifting and merging of the intense 3588 and 3550 cm−1 O−H bands into a broader shoulder band at 3510 cm−1. These IR features reveal the transformation of more isolated hydroxyls at higher temperature into hydrogen-bonded species at lower temperature. Furthermore, the extent of these O−H interactions at higher and lower temperatures correlated to the apparent strength of the templated monolith gel formed in the oven-heated vial; the monolith felt more flexible immediately after being removed from the 105 °C oven and more rigid after cooling to ambient temperature, ∼20 °C. The IR absorbance spectra (absorbance = log(1/I)) of the oven-prepared, ground and washed E3/TEPA-Y monolith particles at 50 °C are shown in Figure 9a. Note, comparing these IR spectra with that of PEG400/KBr in Figure S6 indicates

that PEG400 was effectively removed from the particles. These spectra, along with the in situ results presented in Figure 8, help clarify differences in the particles’ chemical structures that affected their REE recovery performance. The N−H, C−N, and C−O−C band positions for these monolith particles were well aligned to those of the DRIFTS-formed monolith. Clear narrowing of the IR features between 3800 and 3000 cm−1 with increasing theoretical percent reaction of the amines (Y values) highlights restricted hydrogen bonding among the particles’ phase-separated networks. These results correlate with the increased apparent porosities (H2O swelling) of the particles’ corresponding intact monoliths. The IR absorbance ratio profiles in Figure 9b were derived from corrected measurements taken from the spectra in Figure 9a (see the Supporting Information for correction), and help to explain the structure−property relationship of the particles. These profiles generally present falling 1598/1468 (C−H, TEPA) and rising 1178/1468 values from Y = 66% to 100%. These data confirm that a greater proportion of primary amines were reacted away by the epoxides at higher Y values to form the polymerized network. Because information regarding the nature of secondary amines is ambiguous in these spectra, actual Y values cannot be ascertained. Still, data for the N−H (−NH2) and C−N moieties support overall greater polymerization at higher epoxide/amine ratios. The trends in the −NH2 and C−N profiles were accompanied by expectedly declining 3398 (−OH; −OH···OH−)/2810 (C−H, TEPA) and 3166 (amine; amine···OH−)/2810 profiles, which represent diminished hydrogen bonding for particles comprised of more polymerized networks. Overall, the DRIFTS data support the hypothesis that better phase separation of highly polymerized amine−epoxy networks facilitated effective distribution of REEadsorbing amine (and potentially hydroxyl) functional groups throughout the particles. These distributed groups are relatively, minimally involved in hydrogen bonding with the network and are free to adsorb incoming Ln3+ species from aqueous solutions. This contrasts both the less polymerized (lower Y values) and less porous networks that contain densely packed and hydrogen-bonded functional groups, which may be unavailable to recover REEs. I


Research Article


■

CONCLUSIONS Porous amine−epoxy monoliths and particles were synthesized via the polymer-induced phase separation (PIPS) process for rare earth element recovery from aqueous solutions. Optimization of the epoxide/amine ratio, that is, theoretical percent reaction of the amines (Y values) at different monomer concentrations in PEG, gave both the maximum water swellings/apparent porosities and the highest REE recoveries for the triepoxy-based E3/TEPA-88 and tetra-epoxy-based E4/ TEPA-88 materials. In situ DRIFTS results revealed that the formation of the highly porous E3/TEPA-88 monolith network resulted from the polymerization of TEPA and E3 monomers via carbon−nitrogen bonds, which precipitated the monolith from the porogen. Ex situ DRIFTS data further indicated that a low degree of hydrogen bonding among well-dispersed amine and hydroxyl groups of this polymerized network contributed to both high apparent porosity of and high REE recovery by the particles at different pH values. Consecutive recovery (recover−strip−recover) of >93% of ppb-level REE species from simulated acid mine drainage solution at a fouling metal/ REE ratio of 12 000/1 and a pH value of 2.6 are promising results toward the practical application of these amine−epoxy particles.

■

AECOM, nor Oak Ridge National Laboratory, nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by tradename, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.7b03859. Temperature profiles for an oven-heated vial with PEG and that used in the DRIFTS cell; metal concentrations of all tested REE solutions; comparison of La3+ recovery by monoliths and particles; literature review of REE sorbent materials; REE recovery from mixed-element solutions; REE selectivities; and process for correcting the IR absorbance (log(1/I)) spectra (PDF)

■

REFERENCES

(1) Rare Earths - Statistics and Information. U.S. Geological Society [online] 2015; http://minerals.usgs.gov/minerals/pubs/commodity/ rare_earths/ (accessed Jan. 1, 2016). (2) Alonso, E.; Sherman, A. M.; Wallington, T. J.; Everson, M. P.; Field, F. R.; Roth, R.; Kirchain, R. E. Evaluating Rare Earth Element Availability: A Case with Revolutionary Demand from Clean Technologies. Environ. Sci. Technol. 2012, 46 (6), 3406−3414. (3) Congressional Research Service. Humphries, M. Rare Earth Elements: The global Supply Chain [online] Dec. 16, 2013; https://fas. org/sgp/crs/natsec/R41347.pdf (accessed Jan., 2016). (4) Senate Armed Services Committee. Clapper, J. R. Statement for the Record: World Wide Threat Assessment of the US Intelegence Community [online] Feb. 9, 2016; https://www.dni.gov/files/ documents/SASC_Unclassified_2016_ATA_SFR_FINAL.pdf (accessed March, 2016). (5) U.S. Environmental Protection Agency. Paul, J.; Campbell, G. Investigating Rare Earth Element Mine Development in EPA Region 8 and Potential Environmental Impacts [online] Aug. 15, 2011; https://nepis. epa.gov (accessed Jan. 10, 2017). (6) Noack, C. W.; Dzombak, D. A.; Karamalidis, A. K. Rare Earth Element Distributions and Trends in Natural Waters with a Focus on Groundwater. Environ. Sci. Technol. 2014, 48 (8), 4317−4326. (7) Gammons, C. H.; Wood, S. A.; Nimick, D. A. Diel Behavior of Rare Earth Elements in a Mountain Stream with Acidic to Neutral pH. Geochim. Cosmochim. Acta 2005, 69 (15), 3747−3758. (8) Cravotta Iii, C. A. Dissolved Metals and Associated Constituents in Abandoned Coal-Mine Discharges, Pennsylvania, USA. Part 1: Constituent Quantities and Correlations. Appl. Geochem. 2008, 23 (2), 166−202. (9) Wang, Y.; Huang, C.; Li, F.; Dong, Y.; Zhao, Z.; Sun, X. The Development of Sustainable Yttrium Separation Process From Rare Earth Enrichments Using Bifunctional Ionic Liquid. Sep. Purif. Technol. 2016, 162, 106−113. (10) Riano, S.; Binnemans, K. Extraction and Separation of Neodymium and Dysprosium From Used NdFeB Magnets: An Application of Ionic Liquids in Solvent Extraction Towards the Recycling of Magnets. Green Chem. 2015, 17 (5), 2931−2942. (11) Noack, C. W.; Dzombak, D. A.; Karamalidis, A. K. Determination of Rare Earth Elements in Hypersaline Solutions Using Low-Volume, Liquid−Liquid Extraction. Environ. Sci. Technol. 2015, 49 (16), 9423−9430. (12) Rout, A.; Binnemans, K. Liquid-Liquid Extraction of Europium(iii) and Other Trivalent Rare-Earth Ions Using a Non-Fluorinated Functionalized Ionic Liquid. Dalton Trans. 2014, 43 (4), 1862−1872. (13) Vander Hoogerstraete, T.; Wellens, S.; Verachtert, K.; Binnemans, K. Removal of Transition Metals From Rare Earths by Solvent Extraction With an Undiluted Phosphonium Ionic Liquid: Separations Relevant to Rare-Earth Magnet Recycling. Green Chem. 2013, 15 (4), 919−927. (14) Sun, X.; Luo, H.; Dai, S. Mechanistic Investigation of Solvent Extraction Based on Anion-Functionalized Ionic Liquids for Selective

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Walter Christopher Wilfong: 0000-0002-1733-6083 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Vyacheslav Romanov and Lei Hong for use of the DRIFTS system. We also thank Phillip Tinker for the pH probe measurements and CHNSO elemental analyses, and Anne Marti with help using the BET. This project was funded by the Department of Energy, National Energy Technology Laboratory, an agency of the United States Government, through a support contract with AECOM. This research was supported in part by an appointment to the National Energy Technology Laboratory Research Participation Program, sponsored by the U.S. Department of Energy and administered by the Oak Ridge Institute for Science and Education. A portion of this technical effort was performed in support of the National Energy Technology Laboratory’s ongoing research under the RES contract DE-FE0004000. Neither the United States Government nor any agency thereof, nor any of their employees, nor J


Research Article

ACS Applied Materials & Interfaces Separation of Rare-Earth Ions. Dalton Trans. 2013, 42 (23), 8270− 8275. (15) Wannachod, T.; Leepipatpiboon, N.; Pancharoen, U.; Nootong, K. Separation and Mass Transport of Nd(III) From Mixed Rare Earths Via Hollow Fiber Supported Liquid Membrane: Experiment and Modeling. Chem. Eng. J. 2014, 248, 158−167. (16) Ramakul, P.; Mooncluen, U.; Yanachawakul, Y.; Leepipatpiboon, N. Mass Transport Modeling and Analysis on the Mutual Separation of Lanthanum(III) and Cerium(IV) Through a Hollow Fiber Supported Liquid Membrane. J. Ind. Eng. Chem. 2012, 18 (5), 1606−1611. (17) Wannachod, P.; Chaturabul, S.; Pancharoen, U.; Lothongkum, A. W.; Patthaveekongka, W. The Effective Recovery of Praseodymium From Mixed Rare Earths Via a Hollow Fiber Supported Liquid Membrane and its Mass Transfer Related. J. Alloys Compd. 2011, 509 (2), 354−361. (18) Lam, K. F.; Kassab, H.; Pera-Titus, M.; Yeung, K. L.; Albela, B.; Bonneviot, L. MCM-41 “LUS”: Alumina Tubular Membranes for Metal Separation in Aqueous Solution. J. Phys. Chem. C 2011, 115 (1), 176−187. (19) Wen, B.; Shan, X.-q.; Xu, S.-g. Preconcentration of Ultratrace Rare Earth Elements in Seawater with 8-hydroxyquinoline Immobilized Polyacrylonitrile Hollow Fiber Membrane for Determination by Inductively Coupled Plasma Mass Spectrometry. Analyst 1999, 124 (4), 621−626. (20) Chevis, D. A.; Johannesson, K. H.; Burdige, D. J.; Tang, J.; Bradley Moran, S.; Kelly, R. P. Submarine Groundwater Discharge of Rare Earth Elements to a Tidally-Mixed Estuary in Southern Rhode Island. Chem. Geol. 2015, 397, 128−142. (21) Chevis, D. A.; Johannesson, K. H.; Burdige, D. J.; Cable, J. E.; Martin, J. B.; Roy, M. Rare Earth Element Cycling in a Sandy Subterranean Estuary in Florida, USA. Mar. Chem. 2015, 176, 34−50. (22) Wang, F.; Zhao, J.; Wei, X.; Huo, F.; Li, W.; Hu, Q.; Liu, H. Adsorption of Rare Earths (III) by Calcium Alginate−Poly Glutamic Acid Hybrid Gels. J. Chem. Technol. Biotechnol. 2014, 89 (7), 969−977. (23) Karadaş, C.; Kara, D. Preconcentration of Rare Earth Elements Using Amberlite XAD-4 Modified With 2,6-Pyridinedicarboxaldehyde and Their Determination by Inductively Coupled Plasma Optical Emission Spectrometry. Water, Air, Soil Pollut. 2014, 225 (5), 1−10. (24) Zheng, X.; Wang, C.; Dai, J.; Shi, W.; Yan, Y. Design of Mesoporous Silica Hybrid Materials as Sorbents for the Selective Recovery of Rare Earth Metals. J. Mater. Chem. A 2015, 3 (19), 10327−10335. (25) Florek, J.; Mushtaq, A.; Lariviere, D.; Cantin, G.; Fontaine, F.G.; Kleitz, F. Selective Recovery of Rare Earth Elements Using Chelating Ligands Grafted on Mesoporous Surfaces. RSC Adv. 2015, 5 (126), 103782−103789. (26) Zhao, Z.; Sun, X.; Dong, Y. Synergistic Effect of Doped Functionalized Ionic Liquids in Silica Hybrid Material for Rare Earth Adsorption. Ind. Eng. Chem. Res. 2016, 55 (7), 2221−2229. (27) Florek, J.; Chalifour, F.; Bilodeau, F.; Larivière, D.; Kleitz, F. Nanostructured Hybrid Materials for the Selective Recovery and Enrichment of Rare Earth Elements. Adv. Funct. Mater. 2014, 24 (18), 2668−2676. (28) Aguado, J.; Arsuaga, J. M.; Arencibia, A.; Lindo, M.; Gascón, V. Aqueous Heavy Metals Removal by Adsorption on Amine-Functionalized Mesoporous Silica. J. Hazard. Mater. 2009, 163 (1), 213−221. (29) Ghoul, M.; Bacquet, M.; Morcellet, M. Uptake of Heavy Metals From Synthetic Aqueous Solutions Using Modified PEISilica Gels. Water Res. 2003, 37 (4), 729−734. (30) Kubova, J.; Nevoral, V.; Stresko, V. Determination of Rare Earth Elements in Mineral Waters by Inductively Coupled Plasma Atomic Emission Spectrometry. J. Anal. At. Spectrom. 1994, 9 (3), 241−243. (31) Pan, Y.; Liu, Z.; Wang, W.; Peng, C.; Shi, K.; Ji, X. Highly Efficient Macroporous Adsorbents for Toxic Metal Ions in Water Systems Based on Polyvinyl Alcohol-Formaldehyde Sponges. J. Mater. Chem. A 2016, 4, 2537. (32) Li, L.; Ma, W.; Shen, S.; Huang, H.; Bai, Y.; Liu, H. A Combined Experimental and Theoretical Study on the Extraction of Uranium by

Amino-Derived Metal−Organic Frameworks Through Post-Synthetic Strategy. ACS Appl. Mater. Interfaces 2016, 8, 31032. (33) Jiang, L.; Zhang, W.; Luo, C.; Cheng, D.; Zhu, J. Adsorption Toward Trivalent Rare Earth Element from Aqueous Solution by Zeolitic Imidazolate Frameworks. Ind. Eng. Chem. Res. 2016, 55 (22), 6365−6372. (34) Song, Y.; Li, J.; Ye, G.; Xu, J.; Jiang, M. Polyamidoxime/ Poly(vinyl alcohol) Composite Chelating Fiber Prepared by Emulsion Spinning and Its Adsorption Properties for Metal Ions. Ind. Eng. Chem. Res. 2015, 54 (49), 12367−12373. (35) Gao, Y.; Zhang, S.; Zhao, K.; Wang, Z.; Xu, S.; Liang, Z.; Wu, K. Adsorption of La3+ and Ce3+ by Poly-γ-Glutamic Acid Crosslinked with Polyvinyl Alcohol. J. Rare Earths 2015, 33 (8), 884−891. (36) Hara, A.; Inoue, R.; Takahashi, N.; Nishida, K.; Kanaya, T. Trajectory of Critical Point in Polymerization-Induced Phase Separation of Epoxy/Oligoethylene Glycol Solutions. Macromolecules 2014, 47 (13), 4453−4459. (37) Rahmi, D.; Takasaki, Y.; Zhu, Y.; Kobayashi, H.; Konagaya, S.; Haraguchi, H.; Umemura, T. Preparation of Monolithic Chelating Adsorbent Inside a Syringe Filter Tip for Solid Phase Microextraction of Trace Elements in Natural Water Prior to Their Determination by ICP-MS. Talanta 2010, 81 (4−5), 1438−1445. (38) Wang, S.; Zhang, R. Preconcentration and Determination of Copper in Aqueous Solution with Epoxy Resin-based Monolithic Column Containing Large Interconnected Pores. Sep. Sci. Technol. 2007, 42 (5), 1079−1092. (39) Wang, S.; Zhang, R. Column Preconcentration of Lead in Aqueous Solution with Macroporous Epoxy Resin-based Polymer Monolithic Matrix. Anal. Chim. Acta 2006, 575 (2), 166−171. (40) Maron, L.; Eisenstein, O. Do f Electrons Play a Role in the Lanthanide−Ligand Bonds? A DFT Study of Ln(NR2)3; R = H, SiH3. J. Phys. Chem. A 2000, 104 (30), 7140−7143. (41) Di Bernardo, P.; Zanonato, P. L.; Melchior, A.; Portanova, R.; Tolazzi, M.; Choppin, G. R.; Wang, Z. Thermodynamic and Spectroscopic Studies of Lanthanides(III) Complexation with Polyamines in Dimethyl Sulfoxide. Inorg. Chem. 2008, 47 (3), 1155−1164. (42) Forsberg, J. H.; Moeller, T. Rare Earths. LXXIX. Syntheses and Properties of Ethylenediamine Chelates of the Tripositive Lanthanide Ions. Inorg. Chem. 1969, 8 (4), 883−888. (43) Tsujioka, N.; Ishizuka, N.; Tanaka, N.; Kubo, T.; Hosoya, K. Well-controlled 3D Skeletal Epoxy-based Monoliths Obtained by Polymerization Induced Phase Separation. J. Polym. Sci., Part A: Polym. Chem. 2008, 46 (10), 3272−3281. (44) Zhang, H.; McDowell, R. G.; Martin, L. R.; Qiang, Y. Selective Extraction of Heavy and Light Lanthanides from Aqueous Solution by Advanced Magnetic Nanosorbents. ACS Appl. Mater. Interfaces 2016, 8 (14), 9523−9531. (45) Ashour, R. M.; Abdel-Magied, A. F.; Abdel-khalek, A. A.; Helaly, O. S.; Ali, M. M. Preparation and Characterization of Magnetic Iron Oxide Nanoparticles Functionalized by l-cysteine: Adsorption and Desorption Behavior for Rare Earth Metal Ions. J. Environ. Chem. Eng. 2016, 4 (3), 3114−3121. (46) Singh, S.; Barick, K. C.; Bahadur, D. Surface Engineered Magnetic Nanoparticles for Removal of Toxic Metal Ions and Bacterial Pathogens. J. Hazard. Mater. 2011, 192 (3), 1539−1547. (47) Wilkinson, F. A. C. a. G. Advanced Inorganic Chemistry: A Comprehensive Text, 4th ed.; John Wiley & Sons, Inc.: New York, 1980. (48) Yu, J.; Chuang, S. S. C. The Structure of Adsorbed Species on Immobilized Amines in CO2 Capture: An in Situ IR Study. Energy Fuels 2016, 30, 7579. (49) Gunnlaugsson, T.; Brougham, D. F.; Fanning, A.-M.; Nieuwenhuyzen, M.; O’Brien, J. E.; Viguier, R. Effect on pKa of Metal-Bound Water Molecules in Lanthanide Ion-Induced Cyclen “Cavities. Org. Lett. 2004, 6 (26), 4805−4808. (50) Tyler, L. A.; Noveron, J. C.; Olmstead, M. M.; Mascharak, P. K. Modulation of the pKa of Metal-Bound Water via Oxidation of Thiolato Sulfur in Model Complexes of Co(III) Containing Nitrile Hydratase: Insight into Possible Effect of Cysteine Oxidation in Co− Nitrile Hydratase. Inorg. Chem. 2003, 42 (18), 5751−5761. K


Research Article

ACS Applied Materials & Interfaces (51) Victor, W. S. Some Physical Properties of Rare Earth Chlorides in Aqueous Solution. Ph.D. Thesis, Iowa State University, 1960. (52) Park, D. M.; Reed, D. W.; Yung, M. C.; Eslamimanesh, A.; Lencka, M. M.; Anderko, A.; Fujita, Y.; Riman, R. E.; Navrotsky, A.; Jiao, Y. Bioadsorption of Rare Earth Elements Through Cell Surface Display of Lanthanide Binding Tags. Environ. Sci. Technol. 2016, 50 (5), 2735−2742. (53) Zhao, F.; Cong, Z.; Sun, H.; Ren, D. The Geochemistry of Rare Earth Elements (REE) in Acid Mine Drainage From the Sitai Coal Mine, Shanxi Province, North China. Int. J. Coal Geol. 2007, 70 (1−3), 184−192. (54) Shannon, R. t. Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides. Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr. 1976, 32 (5), 751−767. (55) Stewart, J. E. Vibrational Spectra of Primary and Secondary Aliphatic Amines. J. Chem. Phys. 1959, 30 (5), 1259−1265. (56) Miller, D. D.; Chuang, S. S. C. The Effect of Electron-Donating Groups and Hydrogen Bonding on H2S Capture Over Polyethylene Glycol/Amine Sites. J. Phys. Chem. C 2016, 120 (2), 1147−1162. (57) Dunkers, J.; Ishida, H. Vibrational Assignments of N,N-bis(3,5dimethyl-2-hydroxybenzyl)methylamine in the Fingerprint Region. Spectrochim. Acta, Part A 1995, 51 (5), 855−867. (58) Wang, X.; Gillham, J. K. Competitive Primary Amine/Epoxy and Secondary Amine/Epoxy Reactions: Effect on the Isothermal Time-to-vitrify. J. Appl. Polym. Sci. 1991, 43 (12), 2267−2277. (59) Silverstein, R. M. In Spectrometric Identification of Organic Compounds; Yee, J., Ed.; John Wiley & Sons, Inc.: Hoboken, 2005; p 502. (60) Srikanth, C. S.; Chuang, S. S. C. Spectroscopic Investigation into Oxidative Degradation of Silica-Supported Amine Sorbents for CO2 Capture. ChemSusChem 2012, 5 (8), 1435−1442.

L


Recovering Rare Earth Elements from Aqueous ... - ACS Publications

Recommend Documents