The flexible and reversible preparation of columns for use in high-performance solid phase extraction chromatography by physisorption of organophosphorus acid extractants has been investigated in detail.
Journal of Chromatography A 1676 (2022) 463278 Contents lists available at ScienceDirect Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma Impregnation of preparative high-performance solid phase extraction chromatography columns by organophosphorus acid compounds Meher G Sanku, Kerstin Forsberg, Michael Svärd∗ Department of Chemical Engineering, KTH Royal Institute of Technology, Teknikringen 42, SE-11428 Stockholm, Sweden a r t i c l e i n f o Article history: Received 17 March 2022 Revised 18 May 2022 Accepted 24 June 2022 Available online 25 June 2022 Keywords: Physisorption Impregnation Metal extraction Column Separation a b s t r a c t The flexible and reversible preparation of columns for use in high-performance solid phase extraction chromatography by physisorption of organophosphorus acid extractants has been investigated in detail Two extractants have been evaluated, bis (2-ethyl-1-hexyl) phosphoric acid (HDEHP) and 2-ethyl-1-hexyl (2-ethyl-1-hexyl) phosphonic acid (HEHEHP), but the developed procedure should be broadly applicable to other extractants The liquid-liquid solubility of the extractants in feed solvents consisting of aqueous ethanol solutions of varying composition has been determined The total amount of adsorbed extractant has been quantified by complete desorption and elution with ethanol followed by acid-base titrimetry Column impregnation with feed solutions of varying concentration in the undersaturated region has been systematically evaluated, and the influence of a subsequent water wash step has been explored It is shown that to achieve a robust and reproducible physisorption, the adsorbed amount of extractant should be determined after the wash step, and care must be taken when using indirect methods of measurement Equilibrium Langmuir-type adsorption isotherms as a function of the extractant concentration in the feed solution have been determined Adsorption of HEHEHP is higher than HDEHP for equal feed compositions, but the solubility of HEHEHP is lower, resulting in approximately identical maximum coverage levels The ability of the resulting columns to separate rare earth elements have been verified for a mixture of eight metals using a combined isocratic and gradient elution of nitric acid © 2022 The Author(s) Published by Elsevier B.V This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/) Introduction Chromatography as a separation method has been gaining attention as a promising alternative to solvent extraction – a current cornerstone technology in hydrometallurgy – for many important metals including but not limited to rare earths and precious metals [1–5] Depending on the metals and the purity requirements, solvent extraction can require a large number of mixersettler units in series, which consume vast quantities of often volatile and hazardous organic solvents [6] By attaching a suitable extractant onto the solid phase in a chromatographic column, the extraction process can be made more sustainable and environmentally friendly This is partly due to reduced consumption of solvents and extractants, and improved chemical recycling possibilities, but also because a single column corresponds to multiple equilibrium stages, which eliminates the need for multiple units In high-performance solid phase extraction chromatography [5,7], reverse-phase HPLC columns packed with particles containing ad- ∗ Corresponding author E-mail address: micsva@kth.se (M Svärd) sorbed extractant molecules are used The metals are typically separated by elution with a gradient of a mineral acid, such as nitric acid, in a dynamic process The main challenge with a chromatographic process is to increase the productivity while retaining the purity of individual components [8] Although partly a multivariate optimization problem involving decisions regarding operation variables and fractionation [5], attention should also be directed towards how to reliably and effectively supply the solid phase with a high and stable coverage of extractant The molecular structure of many suitable extractants is composed of a hydrophilic part that interacts with the metal ions and a lipophilic part that interacts with the nonpolar phase, which could be an extraction solvent or the stationary phase of a chromatographic column Bis (2-ethyl-1-hexyl) phosphoric acid (HDEHP) is amongst the most extensively studied extractants [5,7,9–12] with some studies also available on 2-ethyl-1-hexyl (2-ethyl-1-hexyl) phosphonic acid (HEHEHP) [1,10,13] and other acidic as well as neutral extractants [1,10,13–16] Such extractants can be physically adsorbed (physisorption) onto the solid particles of a reverse phase column The stationary material in such columns typically consists of porous silica particles functionalized with e.g octadecyl (C18 ) carbon chains By impregnating the particles with a feed solution https://doi.org/10.1016/j.chroma.2022.463278 0021-9673/© 2022 The Author(s) Published by Elsevier B.V This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/) M.G Sanku, K Forsberg and M Svärd Journal of Chromatography A 1676 (2022) 463278 containing dissolved extractant, the surface of the particles can be coated with a layer of extractant molecules The coverage reached is limited by the appropriate adsorption isotherm under the conditions of the impregnation process Once established, the interactions between the lipophilic part of the extractant and the nonpolar chains on the support material are stable in aqueous solution, even at quite low pH The extractant can be loaded onto the solid support material either before (batch loading) [7,10,12,14,15,17,18] or after (flowthrough loading) [7,9,13,18–20] the column is packed The flowthrough impregnation process has several advantages; it is easy to perform and undo without the need for specialized equipment, the resulting column performance has been repeatedly claimed to be stable over several repeated elutions even under harsh acidic conditions, and the choice of extractant and the coverage level – and thereby the column performance – can be tuned to specific needs Moreover, any gradual decrease in the column performance could easily be restored to initial levels by a re-impregnation step Using 55 wt% methanol in water as feed solvent, extractant coverages in the range 89 to 345 mol HDEHP/m3 column have been attained in previous studies [7,9,13,21] For HEHEHP, a coverage of 103 mol/m3 column has been reported [13] However, the methodology for quantification and validation of the coverage is often not described, or differs, ranging from analysis of breakthrough solution during impregnation [3,7] to complete flushing of extractant before analysis [13] Moreover, it is rarely shown that the impregnation processes are reproducible or under which circumstances Kifle et al evaluated impregnation of columns at a coverage of approx 120 mol/ m3 column (approx 0.3 mmol HDEHP on the column) and reported that the process was repeatable [7] No studies were performed at higher extractant concentrations Conversely, Max-Hansen reported that the ligand concentration after column impregnation did not always reach expected levels (although no column details were reported in the study) [3] In order to serve as the basis for a feasible separation process, it is of crucial importance that the impregnation process can reproducibly and reliably deliver a column with the desired extractant coverage level, and a sufficient stability over repeated elutions under the conditions required to separate the metals for which it is designed Currently, there is ambiguity or a lack of clarity in the available literature with respect to these matters In the present work, the first step for metal extraction using column chromatography, the column preparation step, has been thoroughly studied, for two extractants (HDEHP and HEHEHP; shown in Fig 1) Data on the liquid-liquid solubility of the extractants in the feed solvent mixtures, crucial to avoid liquid-liquid phase separation during impregnation which could lead to obstructed flow, pressure build-up and a damaged column, has been collected Adsorption isotherms, key to knowing how to alter the feed solution composition in order to obtain the required extractant coverage on the stationary phase, have been measured Particular attention is devoted to the repro- ducibility of the impregnation process Two methods of estimating the extractant coverage are contrasted, shedding light on the adsorption behaviour of the extractant Reverse phase C18 -coated mesoporous silica columns have been impregnated with each extractant using an ethanol-water mixture as feed solvent The resulting columns have been evaluated with respect to their ability to separate eight REEs predominant in apatite ore (La, Ce, Pr, Nd, Sm, Gd, Dy and Y) [22] However, the results of the study should be broadly applicable to other RP columns, solvents and extractants Materials and methods 2.1 Materials The different solutions used in this study are described below All solutions were prepared using the individual components as received HNO3 (>69.9%), and ethanol (>99%) were purchased from VWR, acetic acid (>96%) from Merck, HDEHP (D2 EHPA, bis (2-ethyl-1hexyl) phosphoric acid; >97%), Arsenazo III (2,7-bis (2-arsonophenylazo) chromotropic acid) and urea (>99.5%) from SigmaAldrich, HEHEHP (EHEHPA; PC-88A; 2-ethyl-1-hexyl (2-ethyl-1hexyl) phosphonic acid; >95%) from Daihachi Chemical Industry Co., and NaOH (2.5 M) from J.T.Baker Single-element REE standard solutions (10,0 0 mg/L) were purchased from Teknolab Sorbent All chemicals were used as received Milli-Q grade water was used to prepare all the solutions Column conditioner A solution of ethanol and water with a concentration matching the feed solution: 62 wt% ethanol in water Feed solution Acidic organophosphorus solutions of HDEHP or HEHEHP dissolved in 62 wt% ethanol in water The amount of extractant in these solutions was decided based on the solubility, and the resulting solution was verified to be a homogeneous singlephase liquid NaOH solution A 0.25 M solution of NaOH in water was used for titrations REE solution A solution of eight REEs (La, Ce, Pr, Nd, Sm, Gd, Dy and Y), with a concentration of 37.5 mg/L (with respect to each metal) or 300 mg/L (with respect to the total REE content), prepared from standard solutions mixed in equal amounts The HNO3 concentration in the solution was maintained at 0.59 M HNO3 solutions 2.0 M and 5.0 M HNO3 solutions prepared by dilution of concentrated HNO3 (69.9%) Arsenazo III solution A 0.15 mM aqueous Arsenazo III solution, containing 0.10 M acetic acid and 10 mM urea, used for postcolumn reaction 2.2 Experimental setup A modified Thermo Scientific Dionex ICS-50 0+ Ion Chromatography System, shown in Fig 2, has been used in the present work Solutions, kept in a cryostatic water bath (Julabo FP-50, 25±1 °C) for temperature control, were pumped via a degasser through the column using a quaternary gradient pump The column temperature was maintained at 25±2 °C by means of a column thermostat (BioTek Instruments HPLC 582) and the tubing between the solution bottles and the column was thermally insulated A dedicated pump was used for the post-column reaction solution, which was mixed with the eluting solution downstream of the column A 750 μL knitted reaction coil was used to provide the necessary reaction time for Arsenazo III-REE complex formation A Dionex UV–Vis variable wavelength detector (VWD), placed downstream of the column, was used to detect extractant breakthrough signals at 288 nm and chromatograms (Arsenazo III-REE complexes) at 658 nm An automatic fraction collector module was used to collect samples for NaOH titration A Fig Molecular structure of the two extractants M.G Sanku, K Forsberg and M Svärd Journal of Chromatography A 1676 (2022) 463278 Fig Schematic of the HPLC setup used in this work The solutions used at the different channels in the water bath vary with the purpose For column preparation: A – column conditioner, B – acidic organophosphorus solution, C – ethanol, and D – Milli-Q water For REE separation: A – Milli-Q water, B – M HNO3 , C – M HNO3 and E – Arsenazo III solution 150 mm x 4.6 mm (i.d.) column packed with Kromasil (Nouryon) C18 -functionalized mesoporous spherical particles (di˚ pore volume = 0.9 mL/g; ameter = 10 μm; pore size = 100 A; BET-surface = 320 m2 /g; packed density = 0.66 g/mL; carbon content = 20% or 3.5 μmol/m2 ) was used The volume of the column available for the liquid flow (henceforth CV) as well as contributions to the dead volume was measured by tracer analysis using uracil 2.3.3 Titration HDEHP and HEHEHP can be detected at a wavelength of 288 nm However, because of the wide range of extractant concentrations evaluated, a linear relationship between the intensity and concentration according to the Beer-Lambert law is not applicable, and the use of the detector was restricted to qualitative analysis In this work, the amount of adsorbed organophosphorus compound was calculated using NaOH titration Two methods were used: the indirect method (Eq (1)), by titration of the feed collected after passing through the column, and the direct method (Eq (2)), by titration of the ethanol eluate collected after washing the column with water 2.3 Column preparation 2.3.1 Solubility of organophosphorus compounds The (liquid-liquid) solubility of the organophosphorus extractant compounds (HDEHP and HEHEHP) in aqueous ethanol solutions has been determined using an iterative process Initially, 0.4 to g of the respective organophosphorus compound was added to 3.6 to 15 g of ethanol to form a homogeneous solution Water was added to this solution dropwise until the solution turned turbid, indicating liquid-liquid phase separation Ethanol was again added until the clear point was reached The process was then repeated several times The cloud points (onset of liquid-liquid separation) and the clear points (homogeneous solution) thus form two curves, which flank the true solubility curve Experiments were performed in a total of 20 vials to produce 104 and 114 data points (counting both cloud and clear points) for HDEHP and HEHEHP, respectively q = c1 F t − c2 V (1) q = c2 V (2) where q is the estimated amount of adsorbed acid (in mmol), c1 is the inlet concentration (in M) of acid feed solution to the column, F is the feed flow rate (in mL/min), t is the feed duration (in min), V is the volume of NaOH solution consumed (in mL) and c2 is the concentration of NaOH solution (in M) The difference between the acid amount obtained by both methods should correspond to the difference between the amounts of extractant adsorbed strongly and weakly (weakly adsorbed acid is removed in the water wash step) All samples were titrated with 0.25 M NaOH solution using phenolphthalein as indicator To establish the accuracy and validity of the process, titration was also performed using solutions of known amounts of HDEHP and HEHEHP Titration of samples from the water wash step was not done due to the presence of two liquid phases 2.3.2 Resin impregnation The retention of ligands on the column is a result of hydrophobic interactions between the C18 chains and the aliphatic moieties of the HDEHP and HEHEHP molecules Before column impregnation, any retained acid from previous runs was eluted with ethanol (14 CVs) Then 20 CVs of column conditioner was run through the column at mL/min This was followed by equilibration of the column with organophosphorus feed solution at varying flow rates (0.85 to mL/min for HEHEHP and 0.61 to mL/min for HDEHP) chosen to ensure constant inlet pressure (72.5 ± bar) Preliminary runs were performed to measure the amount of feed solution required to achieve an equilibrium coverage level Finally, the column was washed with Milli-Q water (at least 20 CVs) For improved reproducibility, special attention has been paid to transition steps that can lead to formation of two phases For example, the ethanol wash and the water wash steps are performed at low flow rates of 0.1 or 0.2 mL/min for at least 1.2 CVs followed by gradual increase of flow rate and flushing the column with ethanol/water at a higher flow rate The higher flow rate was set to mL/min for ethanol and mL/min for water unless otherwise specified 2.3.4 Data presentation The performance indicator of interest in this study is the amount of organophosphorus acid adsorbed on the stationary phase (the extractant coverage) Presenting coverage values only as amount of acid adsorbed on the column (q; in mmol) restricts the comparison of data to a single column For the results to be comparable across different columns, it is beneficial to also present them in a more generalized form Coverage values given in units of e.g mmol/m3 of internal column volume, or mmol/m2 of available stationary phase surface area, could be scaled with the column dimension, provided that the columns are identically packed with the same stationary phase particles Kifle et al compared the results of an impregnation process on different columns, including C8 and C18 columns as well as two C18 columns with different surface area [7] This study clearly shows that the amount of adsorbed acid (presented as mmol/g silica) is affected by both the hydrophobicity (length of the carbon chain) of the column material and by the M.G Sanku, K Forsberg and M Svärd Journal of Chromatography A 1676 (2022) 463278 physical properties of the material such as surface area and pore size The results of a C18 column are hardly transferable to a C8 column, but more studies are required even to compare two C18 materials (of different pore size and surface area) and to know under what conditions they are comparable In the present study, since only one type of column was used for all experiments, extractant coverage is presented in terms of mmol adsorbed acid per entire column Corresponding values of mmol adsorbed acid per mmol of carbon on the stationary phase surface (as estimated by the supplier) are given in Table A.1 and A.2 of the Appendix The data can be converted to other forms based on the information provided in Section 2.2 according to Eq (3) Since most literature data is only presented in terms of mmol per entire column with little details about the properties of the material used, in the introduction section, data reported in literature is presented in terms of mmol/m3 of internal column volume for crude comparison q q = n Vc ρ p aϕ (3) where n denotes the amount of C18 groups in the packed column (in mmol), Vc the hollow column volume (in mL), and where ρ p is the packing density (in g/mL), a the specific surface area (in m2 /g), and ϕ the carbon content (in mmol C18 /m2 ) of the stationary phase Fig Experimental data and regressed exponential curves of the liquid-liquid solubility of HDEHP in ethanol-water solutions Black and white symbols - experimental data; black line - regressed curve; pink bands - 95% confidence bands; solid symbols – cloud points; hollow symbols – clear points; red triangle – maximum feed concentration used for impregnation experiments 2.4 REE separation Two columns were prepared, with coverages of 0.5 mmol of HDEHP and HEHEHP, respectively, according to the method described in Section 2.3.2, and evaluated for separation of eight REEs (La, Ce, Nd, Pr, Sm, Gd, Dy and Y) under different elution conditions Before each run, remaining traces of metals were initially removed from the column by eluting with CVs of M HNO3 solution, after which the column was conditioned with at least CVs of HNO3 solution of the same concentration as the elution solution A 50 μL sample of REE solution was then injected and elution was performed under isocratic conditions for 30 A 10 HNO3 gradient up to M HNO3 was appended after the isocratic step to ensure all REEs in the column were completely eluted The HNO3 concentration was controlled using the quaternary pump by means of mixing water with M or M HNO3 solution The column and solution temperatures were kept constant at 25 °C A constant flow rate of mL/min was used throughout the experiments Fig Experimental data and regressed exponential curves of the liquid-liquid solubility of HEHEHP in ethanol-water solutions Black and white symbols - experimental data; black line - regressed curve; pink bands - 95% confidence bands; solid symbols – cloud points; hollow symbols – clear points; red triangle – maximum feed concentration used for impregnation experiments Results and discussion 3.1 Validation of titration method Table shows the concentrations of HDEHP and HEHEHP obtained by titration (ct ) for solutions of known extractant concentrations (cf ) as well as for pure water and ethanol Between – repeat analyses were carried out for each solution Negligible amounts of acid were detected in the pure solvents as expected The deviation between repeat experiments is consistently below 1.4%, indicating good reproducibility irrespective of differences in concentration and extractant However, the relative error with respect to the known concentration is generally higher (mean 4.8%, ranging up to 13%) Most of the errors obtained are positive, and a part of this can be attributed to the small but systematic error involved in using phenolphthalein as indicator There is a trend of larger relative errors obtained for lower acid concentrations, as should be expected 3.2 Column preparation 3.2.1 Solubility of organophosphorus compounds The liquid-liquid solubility of HDEHP and HEHEHP in ethanolwater solutions is shown in Figs and as sets of experimentally determined cloud- and clear points The experimental data has been regressed to fit an exponential function, Eq (4), shown as solid black lines in the graphs together with associated 95% confidence bands The corresponding fitting parameters are given in Table together with goodness of fit values (R2 ) cs = A · exp x B +C (4) M.G Sanku, K Forsberg and M Svärd Journal of Chromatography A 1676 (2022) 463278 Table Validation of the titration method against solutions of known concentration Calc Experiment cf (mmol) ct (mmol) RE∗ (%) ct (mmol) RE∗ (%) HDEHP (24 mM) HDEHP (60 mM) HDEHP (167 mM) HDEHP (239 mM) HDEHP (300 mM) HDEHP (525 mM) 0.213 0.213 1.494 0.225 0.24 1.623 6% 13% 9% 0.225 0.23 1.63 6% 8% 9% 4.269 4.263