Two poly(amidoamine) crosslinked resins whose repeating units are mimics of L-lysine (LYMA) and of EDTA (LMT85) prove capable of quantitatively and reversibly sorbing manganese ions (Mn2+). Halving times are on the order of 1 min or less, and Mn2+ can be abated to the ppb level.
Polyamidoamines (PAA)-based resins have recently emerged as an advantageous tool for metal ion pollutant sorption in water treatments, complying with nearly all desirable requirements for this application, such as bio- and eco-compatibility, effectiveness in aqueous systems, low cost and easy scaling up of preparation, biodegradability, fast and quantitative sorption even at low pollutant concentration, reversibility, and chemical versatility allowing a virtually unlimited range of target-oriented architectures.1,2 Recently, two novel PAA resins carrying amine and carboxyl groups, LYMA (Scheme 1) and LMT85 (Scheme 2) were prepared and tested for their sorption ability3 adopting a standard set of six heavy metal ions (Cu2+, Pb2+, Ni2+, Co2+, Cd2+, and Zn2+) corresponding to the DIN 38406-16 protocol for trace determination in water.4 LYMA, whose repeating unit contained one carboxyl and two amine groups and was a mimic of L-lysine, proved selective for Cu2+ and Ni2+, the other ions tested being negligibly sorbed. LMT85, whose repeating unit contained two amine and five carboxyl groups and was a mimic of EDTA, proved capable of rapidly and quantitatively sorbing all the ions tested either singly or in mixed solution. The sorption process was reversible and both resins were easily regenerated by acidification.
Mn2+ was not included in the set of ions so far considered. The problem of its removal from waters, typically concerning local environmental contexts and often concurrent with a high iron amount, arises from esthetic (dark brown/blackish layers on pipings that can collapse resulting in dark stains in laundry clothes and/or dark particles suspension in water), organoleptic (unpleasant taste of beverages such as tea or coffee), and health issues (particularly regarding neurological pathologies).5,6 For instance, a Greek study testing the effects of Mn2+ in drinking waters on aged people found a progressively higher prevalence of neurological signs of chronic poisoning with increasing metal concentration (from 0.04 to 2.3 ppm).7 The maximum Mn2+ concentration in drinking water in the US is presently 0.05 ppm.8,9 Only very few studies of Mn-sorbing resins are presently available in the literature,10–13 and the proposed materials do not fulfill all the aforementioned requirements. Among these, it is worth mentioning that an Mn2+ chelating resin containing iminodiacetate groups recently proposed to obtain high-purity Mn for advanced electronics devices.14
Lysine is reported to give complexes with many ions including Co2+,15 in this respect behaving differently from LYMA that does not sorb this ion, and Mn2+. Further investigation has proved that Mn2+ is well sorbed by LYMA, which toward this ion behaves as lysine mimic. The aim of this communication is to report on the performance of both LYMA and LMT85 as Mn2+ sorbing resins by in situ monitoring the sorption kinetics with voltammetric techniques and using ad hoc-developed protocols.
LYMA and LMT85 were prepared by polyaddition of L-lysine and N,N′-ethylenediaminodisuccinic acid, respectively, to 2,2-bisacrylamidoacetic acid, as recently reported.3
Batch Sorption Tests
Batch sorption experiments were carried out with the resin confined in highly porous Japanese tea paper envelopes, in conditions of both large metal excess (to estimate the resin maximum sorption capacity) and of large resin excess (to estimate the lowest metal concentration obtainable), determining after at least 1-day equilibration the residual metal quantity in solution by anodic stripping voltammetry ASV16,17 performed by an Autolab 12 potentiostat controlled by a PC with GPES software. The working cell included a hanging drop mercury electrode HDME (Metrohm VA Stand) as the working electrode, an aqueous saturated calomel electrode SCE as the reference electrode, and a platinum counter electrode. ASV analyses were carried out on solutions made up by 10 mL sample + 50 μL 25% NH3 + 2.5 mL supporting electrolyte (0.1 M Na2B4O7 + 0.3 M NaOH) + 10 μL Zn2+ solution (100 mg dm−3), thoroughly deaerated by N2 bubbling. The standard addition method was applied, the stripping peak current intensities being linearly proportional to the bulk Mn2+ concentration at constant protocol. ASV parameters: stirring 2000 rpm, purge time 300 s, pulse amplitude −0.075 V, deposition potential −1.7 V, deposition time 90 s, equilibration time 5 s, potential scan from −1.62 to −1.25 V (SCE) [the Mn2+ stripping peak appearing at about −1.45 V (SCE)], voltage step 0.004 V, voltage step time 0.5 s, and sweep rate 0.008 V s−1.
All reagents (ultrapure or in a few cases reagent-grade) were purchased by Aldrich and used as received. Before each analysis, the glassware was thoroughly cleaned by repeated washings with concentrated HNO3.
In Situ Sorption Kinetics Experiments
Mn2+ sorption kinetics were studied at different (resin monomeric unit:metal) ratios and at different metal starting concentrations both by in situ ASV monitoring and by final ASV verification of the limiting residual metal concentration. While the latter was achieved by the same ASV reference protocol above reported, an ad hoc protocol had to be developed for the in continuo monitoring, implying pH 7 (as in natural waters), and the absence of ions potentially sorbed by the resin. Accordingly, (a) the borax + ammonia buffer at pH 9.5–10 of the reference protocol was replaced with a 0.0065 M TRIS + 0.1 M TRIS hydrochloride buffer, corresponding to pH 6.8; (b) the Zn2+ addition was omitted; (c) the preaccumulation step was reduced to optimize the quantity of metal deposited on the mercury drop (10 s at 0.0001 M, to be increased to 3, 60, and 90 s at lower concentration ranges). As a consequence of (a) and (b), the baseline is no more horizontal and sensitivity is quite worse (detection limit 10−7 M instead of 10−9 M), but more than sufficient to follow the sorption kinetics, implying a fairly high metal concentration range.
RESULTS AND DISCUSSION
LYMA and LMT85 were prepared as previously described.3 Briefly, LYMA was obtained in a single step by Michael polyaddition of L-lysine sodium salt with methylenebisacrylamide in 1/1.5 stoichiometric ratio (Scheme 1) in aqueous solution and at room temperature after a short initial heating to dissolve the bisacrylamide. In Michael polyadditions with bisacrylamides, the L-lysine sodium salt acts as tetrafunctional monomer and yields crosslinked resins. LMT85 was obtained in two steps (Scheme 2). In the first step, a low-molecular-weight end-capped oligomer was prepared by polyaddition in aqueous solution of the tetrasodium salt of (S,S)-ethylenediamine-N,N′-disuccinic acid to excess sodium salt of 2,2-bisacrylamidoacetic acid. In the second step, the resultant acrylamide end-capped oligomer was radically polymerized at pH 5.5 by means of the persulfate-bisulfite redox system. Both resins were highly swollen in water, LMT85 to a much higher extent than LYMA (weight increase by swelling 3620 vs. 118%), probably owing to the higher number of ionizable groups, hence of charge density at neutral pH, of the former.
(a)“Working with a high excess of resin monomeric units,” not only the residual amount of dissolved Mn2+ appeared to be below the ppb detection limit of the more sensitive reference ASV protocol, but upon subsequent Mn2+ standard solution additions the newly added metal ions appeared to undergo very fast sorption possibly by resin oligomers released through the highly porous paper envelop;
(b)“Working with a high excess of metal ions” the following sorption capacities were estimated:
0.10 g Mn2+ g−1 resin for LYMA (corresponding to 1:2 metal to resin monomeric unit ratio).
0.18 g Mn2+ g−1 resin for LMT85 (corresponding to 1.2:1 metal to resin monomeric unit ratio).
Both resins visibly changed their morphology upon water and metal sorption (in particular, the original grains aggregated into rigid ice-like flakes), but, unlike former cases,2 they both maintained their initial whitish color (Supporting Information, Part A, Fig. S1).
In Situ Kinetic Experiments
In situ kinetic experiments were performed in a voltammetric cell with a TRIS+TRIS hydrochloride buffer at pH 6.8, affording simulation of a typical natural water pH while avoiding the presence of precipitating ions such as phosphates.
Extremely fast sorption was observed with both LMT85 (Fig. 1) and LYMA; the latter, in spite of showing lower sorption capacity in batch tests, appeared to give faster sorption: the metal peak disappeared within approximately 1 min, so that the voltammetric monitoring protocol, albeit fast (affording ∼1 scan min−1), could not provide enough points for the subsequent kinetic analysis. Also the sorption process appeared quantitative, since the limiting concentration values referred to t, determined by the very sensitive standard ASV protocol, appeared to be below the 1 ppb detection limit.
The very fast sorption rate strictly required a preconditioning step of the resin in a small solvent volume before inserting it in the working cell, to avoid an irregular transient in the very first minutes in which the most significant portion of the kinetics experiment took place (Fig. 2).
Adopting such a precaution, good linearization of all LMT85 kinetics was achieved by a pseudosecond-order kinetic model:18
where qt is the metal ion sorption capacity at time t (in mmol sorbed metal g resin−1), qe the same quantity at equilibrium, and k2 the pseudosecond-order constant in g mmol−1 s−1 or g mmol−1min−1. Integrating eq 1 affords, after rearrangement, the following equation, to be applied for linearization:
A synopsis of the linearized kinetic characteristics is shown in Figure 3 (while the original voltammograms and experimental vs. predicted kinetic curves are reported in detail in the Supporting Information, part B). It may be observed that the slopes regularly decrease with increasing (LMT85 monomeric unit:metal) ratios; furthermore, also the single experiment with 1:5 diluted Mn2+ (represented in figure by asterisks and dash-dotted line) appears fully consistent with the group at higher Mn2+ concentration considering the resin to metal ratio.
From the straight line, slopes and intercepts kinetic constants and halving times, reported in Table 1, were obtained. All halving times were of the order of 1 min or less; they regularly decreased (and the kinetic constants regularly increased) with increasing resin monomeric unit:metal ratio, as visualized in Figure 4.
Table 1. Operating Conditions of the In Situ Sorption Experiments and Corresponding Kinetic Parameters
Both the investigated resins showed an impressive aptitude to Mn2+ sorption. They proved to combine sorption capacity of the order of 0.1–0.2 g g−1 resin with very fast sorption kinetics. The halving times were in the order of 1 min for LMT85 and even less for LYMA, even starting with diluted metal solutions. The residual Mn2+ concentrations after treatment using only slight excess resin were below the ppm detection limit of the reference ASV protocol, well below the law limits (0.05 ppm in the USA). These features, coupled with the easy, inexpensive, and environmentally friendly preparation of both resins, make them an attractive tool for Mn2+ sorption from polluted waters. Considering that the structure of all PAAs is assembled in a modular fashion and their repeating units are known since a long time to behave quasi-independently in ionic reactions,19−21 the metal ion sorption properties of LYMA and LMT85 reported here, as well as those of previous articles on the same3 and other PAA-based resins,2 constitute a reasonable background for planning new copolymeric PAA resins with metal sorption properties finely tuned to specific needs.
As a final observation, the sorption properties of LYMA correspond to the complexing properties of L-lysine, its low molecular weight counterpart as well as chemical precursor, but with qualifications. In particular, L-lysine give complexes with Cu2+, Ni2+, Co2+, and Mn2+ ions,15,22 whereas from the data collected in this and in a previous article3 LYMA sorbs with remarkable efficiency Cu2+, Ni2+, and Mn2+, but shows negligible sorption capacity for Co2+. As regard LMT85, all experiments performed so far indicate a close corresponding between its sorption properties and the complexing properties of its low-molecular-weight analog ETDA.
Financial support by Regione Lombardia, ATP 2009 call, project PAARMENIDE is gratefully acknowledged.