Synthesis and Characterization of Polystyrene-Supported Piperazine-Substituted Triazoles by CuAAC and First Evaluation for Metal Ion Extraction

The goal of this work was to synthesize substituted polystyrene for metal extraction and/or depollution by introduction of substituted piperazines as chelatants starting from Merrifield polymer. After transformation of Merrifield’s resin in azidomethyl polystyrene, click-chemistry using copper (I)-catalyzed Huisgen’s reaction (CuAAC) was performed to prepare different polymers grafted with 1,4-triazoles bearing the piperazines, containing an alkyne as the other counterpart in the CuAAC. The polymers were then first tested for their efficiency to remove various metal ions from neutral aqueous solutions (Fe3+, Ni2+, Cu2+, Zn2+ and Pb2+). The polymers were found to extract Ni2+ and Zn2+ with low efficiencies ≤40%. For Fe3+ and Cu2+, the average extraction was around 80%, and for Pb2+ around 50%. The global selectivity for these polymers was found to be in the order of Fe3+ ≥ Cu2+ > Pb2+ >> Ni2+ > Zn2+.


Introduction
Water pollution by inorganic and organic contaminants is a major concern for water treatment and water stocks management on our planet. These contaminations often occur as a result of human agricultural and industrial activities, but can also be from natural origins, and can be found in all water reservoirs. Metallic ion pollution can have adverse effects on ecosystems and humans, mainly through the drinkable water quality. This kind of pollution is a major concern and solutions to prevent, monitor and treat it are still needed. Metallic cations can be removed from water using clays, natural and synthetic minerals (e.g., zeolites), ceramics and organic polymers bearing the required functions.
For a long time, what we call ion exchange resins, polymers bearing groups chelating a non-toxic metal cation such as sodium or others that can be exchanged with the cation for removal, have been used [1][2][3]. Several polymers can be made, and substitutions introduced, in order to create hydrophilic [4][5][6][7] and hydrophobic polymers [8,9], in order to use them in different applications [10][11][12][13][14]. Ideally, a polymer that can be easily functionalized for a specific application is of great interest. Such polymers can be engineered in a simple way to introduce chelating groups that can complex the metallic ions and find uses in trace detection, depollution or even supported catalysis [15][16][17][18][19][20]. Our previous work was based on amide appendages [23], and we present in this paper the complete synthesis and characterizations of polymer-supported triazole-piperazines using Merrifield polymer (chloromethylated polystyrene) as the starting scaffold, chloride substitution to give azidomethyl polystyrene, followed by CuAAC with piperazine-substituted alkynes as carbamates of propargyl alcohol (Figure 1). The polymers were then first tested for their ability to extract some metallic cations (Fe 3+ , Ni 2+ , Cu 2+ , Zn 2+ and Pb 2+ ) from neutral aqueous solutions by atomic absorption spectrometry. The polymers synthesized here were also used in parallel to elaborate a highthroughput evaluation of chelation by conductometry, which has been published elsewhere [36].

Syntheses of Azidomethyl Polystyrene (A) and Piperazine Propargyl Carbamates (2a-2g)
Organic syntheses (Schemes 1-3, Sections 3.1 and 3.2) were conducted under nitrogen, otherwise stated, using standard methods and procedures. Characterizations were made as described below. Melting points (mp) were measured on a Kofler apparatus after calibration around the observed fusion of the product. Infrared analysis using the attenuated total reflectance technique (ATR/FTIR) was made on a Nicolet FTIR 200 spectrophotometer (Thermo Scientific, Villebon sur Yvette, France) between 4000 and 400 cm´1. Only the main and relevant absorption bands are indicated as vibrations (ν) and angular deformations (δ). Nuclear magnetic resonance spectroscopy (NMR) was recorded on a Bruker Avance 300 WB spectrometer (Bruker, Champs sur Marne, France) at 300 MHz for the proton spectra ( 1 H) and 75.5 MHz for the carbon spectra ( 13 C). Chemical shifts (δ) are indicated after calibration on the residual undeuterated solvent peak in parts per million (ppm) and as follow: s (singlet), d (doublet), t (triplet), dd (doublet of doublet), td (triplet of doublet ), m (multiplet). Coupling constants (J) are given in Hz. High performance liquid chromatography coupled with mass spectrometry (LC-MS) were performed on a Shimadzu LC-10AD VP chromatograph using an Alltima HP C8 3 µ reverse phase column (Alltech, Carquefou, France, length = 53 mm, internal diameter = 7 mm), using a acetonitrile/0.1% formic acid in water gradient at 1 mL¨min´1 using the following sequence: 0-1 min, 30% CH 3 CN; 1-5 min 30% to 100% CH 3 CN; 5-12 min, 100% CH 3 CN; 12-15 min, 100% to 30% CH 3 CN and 15-20 min 30% CH 3 CN. Detection of the species was done using both PDA SPD-M10A VP (diode array, D 2 lamp, 190-400 nm) and ELSD-LT (low temperature evaporative light scattering) detectors. Mass spectrometry was performed on a prep LCMS-2010A (liquid chromatograph mass spectrometer) at the end of the HPLC process between m/z = 100 to 1700 in positive mode by ESI with quadrupole at 5 V and detector at 1.5 kV. Results are given in ELSD purity, retention time (R t ), and m/z of the ion.
Synthesis of Azidomethyl Polystyrene (A) (Scheme 1, Section 3.1) The polymer A was prepared according to literature [37]. Merrifield polymer (2.10 mmol Cl¨g´1, 5.00 g, 10.5 mmol Cl) was suspended in 50 mL DMSO and 5.00 g NaN 3 (76.9 mmol, 7.32 eq.) were added. The reaction mixture was slowly magnetically stirred at 70˝C for 48 h. The cooled mixture was poured in water (50 mL), filtered on sintered glass and washed eight times with water (8ˆ25 mL) and finally dried in an oven at 50˝C overnight. Elemental analysis of the isolated polymer gave 7.63% N and 0.10% Cl contents. Calculated substitution efficiency was 99% and the polymer has a final loading of 1.82 mmol N 3¨g´1 . To a solution of the required N-substituted piperazine 1a to 1g (12.0 mmol) in acetonitrile (45 mL) was added Na 2 CO 3 (1.27 g, 12.0 mmol, 1 eq.). Propargyl chloroformate (1.42 g, 1.17 mL, 12.0 mmol, 1 eq.) was then added dropwise. The reaction mixture was stirred for 48 h at room temperature and then filtered and evaporated under vacuum. The resulting carbamates 2a-g were sufficiently pure to be used without further purification.

Synthesis of Polymer-Supported Triazolic Piperazines 3a-g (Scheme 3, Section 3.2)
Coupling reactions onto the polymer using CuAAC were conducted accordingly to the general procedure indicated below in round bottom flasks equipped with a reflux condenser. To a suspension of 3.00 g of azidomethyl polystyrene A (1.82 mmol N 3¨g´1 , 5.46 mmol N 3 ) in THF (60 mL) was added 6.30 mmol (1.15 eq.) of the alkyne (2a-g), 9.00 ml of triethylamine (6.75 g, 66.7 mmol, 12.2 eq.) and 2.40 mg of copper (I) iodide (12.6 µmol, 4 mol %). The suspension was slowly stirred at room temperature 72 h. After this time, the complete disappearance of the IR band of the azide of the polymer A (2103 cm´1) was observed. The resulting polymer was filtered on sintered glass and washed sequentially with CH 2 Cl 2 , pyridine, and MeOH (60 mL each), the sequential washings were repeated twice. The resulting polymers 3a-g were finally dried overnight in an oven at 50˝C.

Characterizations by ATR/FTIR of Polymer-Supported Triazolic Piperazines
Infrared analyses using the attenuated total reflectance technique (ATR/FTIR) were recorded as previously described for organic compounds (vide supra) Results appear in Figure 2, Section 3.1.

Synthesis and Characterization of Azidomethyl Polystyrene (A) and Propargyl Carbamate Derivatives of Piperazine 2a-g
Following the method of Mioskowski et al. [37], Merrifield's polymer was treated with an excess of sodium azide (7.32 eq.) in DMSO at 70˝C for 48 h (Scheme 1). The azidomethyl polystyrene (A) prepared using this method was submitted to elemental analysis, which gave results of 7.63% for N and 0.10% for Cl. The substitution yield was of 99%, and the polymer has a final loading of 1.82 mmol N 3¨g´1 . The presence of the azido group was confirmed by IR spectroscopy on which its strong characteristic absorption was seen at 2103 cm´1 ν N3 ).
Following the method of Mioskowski et al. [37], Merrifield's polymer was treated with an excess of sodium azide (7.32 eq.) in DMSO at 70 °C for 48 h (Scheme 1). The azidomethyl polystyrene (A) prepared using this method was submitted to elemental analysis, which gave results of 7.63% for N and 0.10% for Cl. The substitution yield was of 99%, and the polymer has a final loading of 1.82 mmol N3·g −1 . The presence of the azido group was confirmed by IR spectroscopy on which its strong characteristic absorption was seen at 2103 cm −1 νN3). The required coupling partners for A were prepared using commercially available N-substituted piperazines (1a-g) and reacting them with propargyl chloroformate in the presence of sodium carbonate in acetonitrile at room temperature for 48 h (Scheme 2). The corresponding propargyl The required coupling partners for A were prepared using commercially available N-substituted piperazines (1a-g) and reacting them with propargyl chloroformate in the presence of sodium carbonate in acetonitrile at room temperature for 48 h (Scheme 2). The corresponding propargyl carbamates 2a-g were obtained in good yields (88%-99%), without the need of a purification step. Structures of the products 1a-d were confirmed by their characteristic signals in NMR (ca. δ = 2.5 (triplet, "C-H) and 4.7 (doublet, OCH 2 C"CH) ppm in 1 H-NMR; 75/78 (C"C) and 154 (C=O) ppm in 13 C-NMR), IR spectroscopies (ca. 3250 (ν "C-H ) and 1700 (ν C=O ) cm´1), as well as by LC-MS.

Synthesis and ATR/FTIR Analysis of Polymer-Supported Triazolic Piperazines 3a-g
Piperazine propargyl carbamates 2a-g (1.15 eq.) and azidomethyl polystyrene (A) were coupled together following a CuAAC procedure (4 mol % CuI in the presence of Et3N) in order to obtain the corresponding polymer-supported 1,2,3-triazoles-piperazines 3a-g (Scheme 3). The reactions were conducted at room temperature for 72 h, after which time the azide vibration (νN3 = 2103 cm −1 ) completely disappeared from the resulting polymers. FTIR spectra of the synthesized polymers 3a-g did not show any residual vibration band from the azido group (νN3, ca. 2100 cm −1 ), as previously mentioned, in either of the alkynes (ν≡C-H, ca. 3250 cm −1 ), as shown in Figure 2. Characteristic vibration for the carbamate bonds (νC=O, ca. 1700 cm −1 ), and other vibrations due to the triazole, the piperazine and the substituents, can be observed when compared to the spectrum of polymer A. The reactions were considered as quantitative based on ATR/FTIR analyses and gave polymers with triazolepiperazine loadings between 1.17 and 1.37 mmol piperazine·g −1 (Table 1). Scheme 2. Synthesis of piperazine propargyl carbamates 2a-g from N-substituted piperazine 1a-g and propargyl chloroformate.

Synthesis and ATR/FTIR Analysis of Polymer-Supported Triazolic Piperazines 3a-g
Piperazine propargyl carbamates 2a-g (1.15 eq.) and azidomethyl polystyrene (A) were coupled together following a CuAAC procedure (4 mol % CuI in the presence of Et 3 N) in order to obtain the corresponding polymer-supported 1,2,3-triazoles-piperazines 3a-g (Scheme 3). The reactions were conducted at room temperature for 72 h, after which time the azide vibration (ν N3 = 2103 cm´1) completely disappeared from the resulting polymers. FTIR spectra of the synthesized polymers 3a-g did not show any residual vibration band from the azido group (ν N3 , ca. 2100 cm´1), as previously mentioned, in either of the alkynes (ν "C-H , ca. 3250 cm´1), as shown in Figure 2. Characteristic vibration for the carbamate bonds (ν C=O , ca. 1700 cm´1), and other vibrations due to the triazole, the piperazine and the substituents, can be observed when compared to the spectrum of polymer A. The reactions were considered as quantitative based on ATR/FTIR analyses and gave polymers with triazole-piperazine loadings between 1.17 and 1.37 mmol piperazine¨g´1 (Table 1). did not show any residual vibration band from the azido group (νN3, ca. 2100 cm −1 ), as previously mentioned, in either of the alkynes (ν≡C-H, ca. 3250 cm −1 ), as shown in Figure 2. Characteristic vibration for the carbamate bonds (νC=O, ca. 1700 cm −1 ), and other vibrations due to the triazole, the piperazine and the substituents, can be observed when compared to the spectrum of polymer A. The reactions were considered as quantitative based on ATR/FTIR analyses and gave polymers with triazolepiperazine loadings between 1.17 and 1.37 mmol piperazine·g −1 (Table 1).

Metal Ion Extractions by the Polymer-Supported Triazolic Piperazines 3a-g
After 24 h incubation on a 250 mg scale (~0.3 mmol) of the polymers (see Table 1) in 20 mL of 50 mg¨L´1 solutions of the salts (1 mg of salt, see Section 2.4 for details), the percentages of extraction for each metal were calculated by AAS, each experiment having been carried out in triplicate. The results for each polymer as a function of the metallic cations are presented in Figure 3. Extraction results showed than Fe 3+ is well extracted by all the synthesized polymers. The maximum levels of extraction for this metal were obtained with polymers 3a (85.40% ± 3.01%) and 3b (83.67% ± 3.87%).
When looking at all the results, as an average of all polymers, we can observe that the metals that are the best extracted are iron (III) and copper (II), at around 80%, lead (II) is extracted at around 50%, and nickel (II) and zinc (II) at 15%-30%. These results suggest that, by selecting the right polymer, a discrimination and selectivity for at least iron (III) and copper (II), and maybe lead (II), can be done over nickel (II) and zinc (II).
Since the polymeric structure differs only by the R substituent on the nitrogen (Scheme 3), a Extraction results showed than Fe 3+ is well extracted by all the synthesized polymers. The maximum levels of extraction for this metal were obtained with polymers 3a (85.40%˘3.01%) and 3b (83.67%˘3.87%).
When looking at all the results, as an average of all polymers, we can observe that the metals that are the best extracted are iron (III) and copper (II), at around 80%, lead (II) is extracted at around 50%, and nickel (II) and zinc (II) at 15%-30%. These results suggest that, by selecting the right polymer, a discrimination and selectivity for at least iron (III) and copper (II), and maybe lead (II), can be done over nickel (II) and zinc (II).
Since the polymeric structure differs only by the R substituent on the nitrogen (Scheme 3), a tentative analysis has been done to try to understand the influence of the substituent's nature on the extraction efficiencies. Polymers 3a and 3b have an alkyl group on the terminal nitrogen, and are thus amines. For 3c-e, and aromatic ring is the substituent, and the compounds are aniline derivatives.
Finally, for 3f and 3g, they include carbonyl linkages, and are in the carbamate and amide families, respectively. Considering only the R substituent, the availability of the nitrogen electron doublet for chelation should theoretically be in the following order: 3a,b > 3c-e > 3f,g. It should be noted, however, that polymer 3e also includes a pyridyl group, capable of metal chelation.
For iron (III), the nature of the N-terminal substituent of the piperazine does not seem to have a determinant effect on the chelation of this metal. Extraction efficiencies are within the same range of 70%-85%. Only a small decrease in efficiency can be observed in the previously cited order (3a,b > 3c-e > 3f,g), which may not be that significant.
In the case of Cu 2+ , a more obvious effect seems to take place, especially for carbamate (3f) and amide (3g) containing N-substituents. Polymers 3a and b are a little more efficient (ca. 90%) than 3c-e (ca. 80%), but there is a more obvious decrease of around 30% while using polymers 3f and 3g.
When looking at lead (II) extraction, the effects are more puzzling. N-alkyl-substituted polymers 3a and 3b are not as efficient as was the case for the two previous metallic ions, with~50% and~20% extraction, respectively. Aryl-substituted polymers 3c-e are better (60%), while 3f and 3g are going down to the 3a level.
For Ni 2+ and Zn 2+ , none of the polymers extracted them well. No conclusion can be drawn other than that these polymers are not suitable for their extraction.
The electro-negativities of the substituted nitrogen of the piperazine may partially explain some of the relative extraction efficiencies, at least for iron (III) and copper (II). For lead (II), the alkyl series (3a,b) is not really matching the previsions, while the other series do. For Ni 2+ and Zn 2+ , none of the results can give an idea about this influence since they are not well extracted, and the levels are the same for all polymers studied here. These results may indicate that the awaited "pendant design" is not the major mode of complexation encountered in this series. It may be, in part, done by an integrated mode, or it may mainly be driven by the triazole one ( Figure 1). Iron (III) and copper (II) are known to be "nitrogen-philic" metals, especially for trinitrogen-/azide-type species. In a previous work, such an effect has been observed [21][22][23].
Another influence of these terminal N-substituents can be the hydrophilic/hydrophobic balance of the polymers. This may explain the access and/or stability of complexes formed between the metal the triazole nucleus. This effect has already been observed in the case of copper (II) and lead (II) in a work on PVC-supported triazolic amides [38].
We have also tried to rationalize the interactions between the polymers and the metal ions based on their electropositivities, ionic radii and water solvation. The extraction efficiencies, based on the metal cation size or electro-positivities, cannot be taken into account since only lead (II) has a superior size and lower electro-positivity when compared to the transition metals having similar radii.
When studying the water solvation of the different metals used here, maybe some explanation can be found when trying to link metal and polymer properties. In Table 2, these are presented in decreasing order of removal of metallic cations. The pH of the initial solution of the cation (pH i ) and the average pH after extraction by all polymers (pH f ) are reported. We can observe that the pH of the solutions were between 3.2 and 7.4, and that there is no direct link between them and the extraction order. The pH at which the extraction is performed is important since a speciation can already be performed as a function of the metal ion [39]. In these cases, a "pH 50 ", a pH at which 50% of the cation is extracted, can be calculated. The pH 50 values can be correlated in a linear way with another value, the first hydrolysis constant in water (pK 1 ), related to the reaction M(OH 2 ) m n+ = M(OH 2 ) m´1 (OH) (n´1)+ + H + [39,40]. When looking at pK 1 , it can be observed that when the value increases, the extraction of the cation decreases almost accordingly. pH 50 is very dependent on the substrate used for cation removal, but always is in linear relation with pK 1 as previously mentioned. From published studies, the order of extraction and pH 50 /pK 1 values increases in the order of Cu ě Pb >> Zn > Ni. From our results, it seems that for Fe 3+ , Cu 2+ and Pb 2+ , the solutions as a function of the polymers are between pH 50 and pH 90 . For Ni 2+ and Zn 2+ , the extractions were never excellent, maybe being in a zone between pH 10 and pH 40 . These results thus give the following order of extraction: Fe 3+ ě Cu 2+ > Pb 2+ >> Ni 2+ > Zn 2+ . a pH of the salt solution before incubation. Water pH was 6.66; b pH of the salt solution after incubation; c pK 1 values from references included in [40].
All the results and analyses cannot definitively identify the discrete complexation behavior of the metal by the polymers at the solid/liquid interface. However, we were able to obtain results that begin to give a better understanding of the subtle equilibrium between all factors influencing the extraction efficiencies of such polymers.

Conclusions
In this work, we studied the chemical grafting of piperazine-chelating units onto commercial poly[styrene] (Merrifield's resin) using the CuAAC procedure between the azided polymer and selected piperazine-N-propargyl carbamates. The synthesized polymers were characterized by FTIR. They were then tested for their efficiency to extract metallic ions from aqueous solution, including Fe 3+ , Ni 2+ , Cu 2+ , Zn 2+ and Pb 2+ . All polymers were found to extract Fe 3+ and Cu 2+ well.
Most of the polymers were found to have a good extraction potential for Pb 2+ . None of the polymers were able to extract Ni 2+ and Zn 2+ with good efficiencies, thus showing certain selectivity when compared to the other studied metallic ions. As an added proof, the extraction efficiencies were studied on 13 metallic cations by a "high-throughput" approach using conductometry [36]. Even if the method was less precise and sensitive, similar average affinities were found for Fe 3+ , Ni 2+ and Pb 2+ (13%-17% lower).
In this series, we obtained far better extraction results than with our first series of triazolic polymers, based on poly(styrene) and poly(vinyl chloride) backbones, bearing propargyl amides and propiolic anilides [23,38].
Even if no clear interpretation can be done with the results for the interfacial chelation process, the good extraction properties encourage us to continue polymer modifications using CuAAC in order to find new triazole-linked polymeric complexing agents for depollution, trace detection and catalytic applications [41][42][43][44][45][46][47]. Further studies will be reported in due course.