Factors Affecting Copper(II) Binding to Multiarmed Cyclam-Grafted

pubs.acs.org/Langmuir © 2009 American Chemical Society

Factors Affecting Copper(II) Binding to Multiarmed Cyclam-Grafted Mesoporous Silica in Aqueous Solution St
ephanie Goubert-Renaudin,† Mathieu Etienne,† St
ephane Brandes,‡ Michel Meyer,‡ Franck Denat,‡ B
en
edicte Lebeau,§ and Alain Walcarius*,† Laboratoire de Chimie Physique et Microbiologie pour l’Environnement, UMR 7564, CNRS - NancyUniversit
e, 405 rue de Vandoeuvre, 54600 Villers-les-Nancy, France, ‡Institut de Chimie Mol
eculaire de l’Universit
e de Bourgogne, UMR 5260, CNRS - Universit
e de Bourgogne, 9 avenue Alain Savary, 21078 Dijon Cedex, France, and §Equipe Mat
eriaux a Porosit
e Contr^ ol
ee, Institut de Science des Mat
eriaux de Mulhouse, LRC 7228, CNRS-UHA-ENSCMu, 3 rue Alfred Werner, F-68093 Mulhouse, France †

Received March 13, 2009. Revised Manuscript Received April 24, 2009 Single- as well as multi-anchored cyclam-functionalized silica samples have been prepared by grafting amorphous silica gel (K60) and mesostructured silica (SBA-15) with silylated cyclam precursors bearing one, two, or four triethoxysilyl groups, respectively ascribed to cyclam-mono, cyclam-di, and cyclam-tetra. Their reactivity toward copper(II) has been thoroughly investigated in aqueous solution and discussed with respect to the number of arms tethering the ligand to the silica surface and the structural ordering of the adsorbent in terms of capacity, long-term stability, and speed of access to the binding sites. Less-thancomplete metal ion uptake was always observed, even in excess of cyclam groups with respect to solution-phase Cu(II), suggesting lower stability of immobilized complexes relative to those in solution. Therefore, the number of arms attaching cyclam moieties to the silica walls (one, two, or four) was found to dramatically affect the binding properties of these hybrids toward copper(II), revealing significantly larger capacities when reducing the number of arms (less rigidity constraints in the macrocycle). In parallel, multiarm tethering resulted in better chemical resistance toward degradation as evidenced by UV-visible monitoring of Cu-cyclam complexes in solution (i.e., more ligand leaching from the adsorbent for singly tethered cyclam). On the other hand, electron spin resonance (ESR) experiments did not evidence significant differences between complexes bearing one, two, or four alkyl arms, since all Cu(II)-cyclam surface complexes were found to be hexacoordinated with a strong equatorial ligand field. Comparison of amorphous gels and mesostructured materials indicates that the binding properties of the adsorbents were hardly influenced by their level of ordering, suggesting that accessibility to the binding sites was not the limiting factor. Some advantage belonging to mesostructured adsorbents was however observed with respect to the rate of access to the active centers at pH values close to neutrality (due to faster mass transport), but this was no more the case when operating at lower pH values where the formation of the Cu-cyclam complex became the rate-determining step, as pointed out by electrochemistry.

Introduction Solid-phase extraction is becoming a powerful method to purify samples and/or preconcentrate target analytes prior to their quantitative determination or for remediation purposes. From that point of view, functionalized silica and sol-gel materials are attractive solids because their rigid and highly porous inorganic matrix ensures good accessibility to a high number of active sites and they can bear a very wide variety of organofunctional groups.1-4 Even more attractive are mesostructured silica-based adsorbents (synthesized using surfactants as structure-directing agents)5-8 due to their regular arrangement of uniform mesopores of monodisperse size ensuring enhanced uptake capacity and improved mass transport *To whom correspondence should be addressed. E-mail: alain.walcarius@ lcpme.cnrs-nancy.fr. (1) Tavlarides, L. L.; Lee, J. S. Ion Exch. Solvent Extr. 2001, 14, 169. (2) Jal, P. K.; Patel, S.; Mishra, B. K. Talanta 2004, 62, 1005. (3) Im, H.-J.; Yost, T. L.; Yang, Y.; Bramlett, J. M.; Yu, X.; Fagan, B. C.; Allain, L. R.; Chen, T.; Barnes, C. E.; Dai, S.; Roecker, L. E.; Sepaniak, M. J.; Xue, Z.-L. ACS Symp. Ser. 2006, 943, 223. (4) El-Nahhal, I. M.; El-Ashgar, N. M. J. Organomet. Chem. 2007, 692, 2861. (5) Sanchez, C.; Soler-Illia, G. J. D. A.; Ribot, F.; Lalot, T.; Mayer, C. R.; Cabuil, V. Chem. Mater. 2001, 13, 3061. (6) Vinu, A.; Hossain, K. Z.; Ariga, K. J. Nanosci. Nanotechnol. 2005, 5, 347. (7) Hoffmann, F.; Cornelius, M.; Morell, J.; Fr€oba, M. Angew. Chem., Int. Ed. 2006, 45, 3216. (8) Fryxell, G. E.; Mattigod, S. V.; Lin, Y.; Wu, H.; Fiskum, S.; Parker, K.; Zheng, F.; Yantasee, W; Zemanian, T. S.; Addleman, R. S.; Liu, J.; Kemner, K.; Kelly, S.; Feng, X. J. Mater. Chem. 2007, 17, 2863. (9) Mercier, L.; Pinnavaia, T. J. Adv. Mater. 1997, 9, 500. (10) Walcarius, A.; Etienne, M.; Lebeau, B. Chem. Mater. 2003, 15, 2161.

9804 DOI: 10.1021/la900892q

rates in comparison to their nonordered homologues.9-12 Such organic-inorganic hybrids are usually prepared by postsynthesis grafting2,13-15 or by a direct route involving sol-gel co-condensation.4,16-18 In addition to their adsorption and extractive concentration properties, other attractive features have been exploited in various fields including catalysis,14,19 separation sciences,20-22 electrochemistry,23,24 optics,25,26 or biology.27-29 (11) Walcarius, A.; Etienne, M.; Sayen, S.; Lebeau, B. Electroanalysis 2003, 15, 414. (12) Mouawia, R.; Mehdi, A.; Rey
e, C.; Corriu, R. J. P. New J. Chem. 2006, 30, 1077. (13) Vansant, E. F.; Van der Voort, P.; Vrancken, K. C. Characterisation and Chemical Modification of the Silica Surface; Elsevier: Dordrecht, 1995. (14) Brunel, D.; Cauvel, A.; Fajula, F.; DiRenzo, F. Stud. Surf. Sci. Catal. 1995, 97, 173. (15) Moller, K.; Bein, T. Chem. Mater. 1998, 10, 2950. (16) Sanchez, C.; Ribot, F. New J. Chem. 1994, 18, 1007. (17) Wen, J.; Wilkes, G. L. Chem. Mater. 1996, 8, 1667. (18) Special issue of J. Mater. Chem. 2005, 15, issue 35-36. (19) Clark, J. H.; MacQuarrie, D. J.; Tavener, S. J. Dalton Trans. 2006, 4297. (20) Collinson, M. M. TrAC, Trends Anal. Chem. 2002, 21, 30. (21) Li, W.; Fries, D. P.; Malik, A. J. Chromatogr., A 2004, 1044, 23. (22) Wu, R.; Hu, L. G.; Wang, F. J.; Ye, M. L.; Zou, H. J. Chromatogr., A 2008, 1184, 369. (23) Walcarius, A.; Mandler, D.; Cox, J. A.; Collinson, M. M.; Lev, O. J. Mater. Chem. 2005, 15, 3663. (24) Walcarius, A. C. R. Chim. 2005, 8, 693. (25) Sanchez, C.; Lebeau, B.; Chaput, F.; Boilot, J. P. Adv. Mater. 2003, 15, 1969. (26) Escribano, P.; Julian-Lopez, B.; Planelles-Arago, J.; Cordoncillo, E.; Viana, B.; Sanchez, C. J. Mater. Chem. 2008, 18, 23. (27) Avnir, D.; Coradin, T.; Lev, O.; Livage, J. J. Mater. Chem. 2006, 16, 1013. (28) Gupta, R.; Chaudhury, N. K. Biosens. Bioelectron. 2007, 22, 2387. (29) Vallet-Regi, M.; Colilla, M.; Izquierdo-Barba, I. J. Biomed. Nanotechnol. 2008, 4, 1.

Published on Web 07/10/2009

Langmuir 2009, 25(17), 9804–9813

Goubert-Renaudin et al.

Numerous polysiloxane-immobilized ligands have been used for the uptake of toxic metal ions.1-3 Most of them incorporate amines, thiols, or more generally N- and/or S-containing groups. Even if they operate quite well, investigation into their long-term stability have not been often considered. This is however especially questionable for silica-based adsorbents bearing alkaline groups, which are prone to induce chemical degradation of the material in aqueous solutions.30 Among the family of N-containing ligands, macrocyclic polyamines are of special interest, particularly 1,4,8,11-tetraazacyclotetradecane known as cyclam, whose remarkable ability to strongly bind various transition and heavy metal cations31-33 has been largely exploited in catalysis or medicine.34-36 If the chelating properties of cyclam derivatives in solution have been extensively reviewed,31-33 much less is known for such macrocycles once immobilized onto solid phases, especially in aqueous media.37-39 A few groups have used organic polymers containing macrocyclic polyamines for binding metal cations.40-42 Since these works lie beyond the scope of the present paper, only those pertaining to polysiloxane-immobilized cyclam chelates and their metallic complexes are briefly summarized hereafter. Most studies involving silica-based materials covalently functionalized with cyclam derivatives have been performed by the groups of Guilard,38,39,43-48 Corriu,46-53 and some others.37,54-58 Materials were either silica gel bound cyclam (obtained by (30) Etienne, M.; Walcarius, A. Talanta 2003, 59, 1173. (31) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chem. Rev. 1995, 95, 2529. (32) Meyer, M.; Dahaoui-Gindrey, V.; Lecomte, C.; Guilard, R. Coord. Chem. Rev. 1998, 178-180, 1313. (33) Elias, H. Coord. Chem. Rev. 1999, 187, 37. (34) Geduhn, J.; Walenzyk, T. S; Koenig, B. Curr. Org. Synth. 2007, 4, 390. (35) Liang, X.; Sadler, P. J. Chem. Soc. Rev. 2004, 33, 246. (36) Delgado, R.; Felix, V.; Lima, L. M. P.; Price, D. W. Dalton Trans. 2007, 2734. (37) Bradshaw, J. S.; Krakowiak, K. E.; Tarbet, B. J.; Bruening, R. L.; Griffin, L. D.; Cash, D. E.; Rasmussen, T. D.; Izatt, R. M. Solvent Extr. Ion Exch. 1989, 7, 855. (38) Barbette, F.; Rascalou, F.; Chollet, H.; Babouhot, J. L.; Denat, F.; Guilard, R. Anal. Chim. Acta 2004, 502, 179. (39) Cuenot, F.; Meyer, M.; Bucaille, A.; Guilard, R. J. Mol. Liq. 2005, 118, 89. (40) Percelay, L.; Louvet, V.; Handel, H.; Appriou, P. Anal. Chim. Acta 1985, 169, 325. (41) Amigoni-Gerbier, S.; Desert, S.; Gulik-Kryswicki, T.; Larpent, C. Macromolecules 2002, 35, 1644. (42) Kavakli, C.; Tuncel, S. A.; Salih, B. Sep. Purif. Technol. 2005, 45, 32. (43) Etienne, M.; Goubert-Renaudin, S.; Rousselin, Y.; Marichal, C.; Denat, F.; Lebeau, B.; Walcarius, A. Langmuir 2009, 25, 3137. (44) Gros, C.; Rabiet, F.; Denat, F.; Brandes, S.; Chollet, H.; Guilard, R. J. Chem. Soc., Dalton Trans. 1996, 1209. (45) Dubois, G.; Tripier, R.; Brandes, S.; Denat, F.; Guilard, R. J. Mater. Chem. 2002, 12, 2255. (46) Dubois, G.; Corriu, R. J. P.; Rey
e, C.; Brandes, S.; Denat, F.; Guilard, R. Chem. Commun. 1999, 2283. (47) Dubois, G.; Rey
e, C.; Corriu, R. J. P.; Brandes, S.; Denat, F.; Guilard, R. Angew. Chem., Int. Ed. 2001, 40, 1087. (48) Corriu, R. J. P.; Embert, F.; Guari, Y.; Rey
e, C.; Guilard, R. Chem.;Eur. J. 2002, 8, 5732. (49) Brandes, S.; David, G.; Suspene, C.; Corriu, R. J. P.; Guilard, R. Chem.; Eur. J. 2007, 13, 3480. (50) Corriu, R. J. P.; Mehdi, A.; Rey
e, C.; Thieuleux, C. New J. Chem. 2003, 27, 905. (51) Corriu, R. J. P.; Mehdi, A.; Rey
e, C.; Thieuleux, C. Chem. Mater. 2004, 16, 159. (52) Corriu, R. J. P.; Mehdi, A.; Rey
e, C.; Thieuleux, C.; Frenkel, A.; Gibaud, A. New J. Chem. 2004, 28, 156. (53) Kassiba, A.; Makowska-Janusik, M.; Alauzun, J.; Kafroumi, W.; Mehdi, A.; Rey
e, C.; Corriu, R. J. P.; Gibaud, A. J. Phys. Chem. Solids 2006, 67, 875. (54) Veuthey, J. L.; Bagnoud, M. A.; Haerdi, W. Int. J. Environ. Anal. Chem. 1986, 26, 157. (55) Bagnoud, M. A.; Haerdi, W. Int. J. Environ. Anal. Chem. 1990, 38, 97. (56) Puranik, D. B.; Guo, Y.; Singh, A.; Morris, R. E.; Huang, A.; Salvucci, L.; Kamin, R.; David, V.; Chang, E. L. Energy Fuels 1998, 12, 792. (57) Gong, Y.; Lee, H. K. Anal. Chem. 2003, 75, 1348. (58) Sujandi, E. A. P.; Han, S.-C.; Han, D.-S.; Jin, M.-J.; Park, S.-E. J. Catal. 2006, 243, 410.

Langmuir 2009, 25(17), 9804–9813

Article

grafting or sol-gel synthesis)37-39,45-49,54-57 or cyclam-functionalized mesoporous solids (grafted or co-condensed with cyclam groups attached to the surface of mesochannels, or as periodic mesoporous organosilicas containing cyclam moieties in the mesopore walls).43,50-53,58,59 They have found applications mainly in catalysis,58,59 chromatography,54,55,57 or gas adsorption.44,45,49 Metal-loaded hybrid materials incorporating metal cations such as Cu(II), Co(II), or Eu(III) have also been prepared and characterized. It was suggested that their coordination chemistry in the solid differs from that in solution.47,48,52,53 Metalation can be performed before synthesis of the hybrid (i.e., on the cyclam-functionalized alkoxysilane precursor) or afterward by allowing the materials to react with solutions containing the metal ions. In these cases, however, rather harsh conditions (2-3 h of refluxing in ethanol with an excess of copper(II)) are required to ensure quantitative metalation (i.e., all cyclam groups complexed with metal ions, including those located within the mesopore walls).48,51 It was also possible to get bifunctionalized solids containing selectively one metal chelate on the surface of mesochannels and another one within the mesopore walls (e.g., Cu and Co).50 These approaches operate quite well to prepare supported metal chelate catalysts58,59 or gas adsorbents,44,45,49 but these previous investigations do not allow one to predict the behavior of cyclam-functionalized silica materials in aqueous media (i.e., in conditions for remediation38,39 or sensing60), where the reactivity of the immobilized active sites still needs to be elucidated. In a previous paper,43 we have reported on the preparation of silica samples functionalized with cyclam groups covalently attached to the silica surface via one, two, or four propyl arms and evaluated their long-term integrity in aqueous medium as well as their reactivity toward protons. Partial chemical degradation of the materials has been observed due to the basic properties of amine functions, but increasing the number of propyl arms (two or four anchorage points with respect to a single one) was found to enhance their stability in water. Meanwhile, significant restriction in protonation degree of cyclam groups was observed as a consequence of multiarm tethering. Such a reactivity decrease was also observed for copper(II) binding on the basis of some preliminary capacity measurements. In the present work, we have examined this binding process in more detail by providing dynamic characterization of the copper(II) uptake by multiarmed cyclam-functionalized silica gels (K60) and mesostructured silica (SBA-15) obtained by grafting cyclam precursors bearing either one, two, or four trialkoxysilyl groups. Various experimental parameters have been considered, including pH, metalto-ligand ratio, structure of the material, and number of tethering arms. Results have been discussed in terms of reactivity of immobilized ligands, accessibility and rate of access to the binding sites, chemical resistance of the adsorbents, and rate-determining step of the uptake process. Copper(II) has been selected as a model probe because of its well-known affinity for cyclam derivatives.61 The results obtained here are expected to be useful for other polyazamacrocycle-functionalized silica materials, as cyclic polyamines are known to form stable complexes with a large number of transition metals,31 the selectivity of which being tunable by judicious functionalization of the ligand scaffold.32 (59) Prasetyanto, E. A.; Park, S.-E. Bull. Korean Chem. Soc. 2008, 29, 1033. (60) Goubert-Renaudin, S.; Etienne, M.; Rousselin, Y.; Denat, F.; Lebeau, B.; Walcarius, A. Electroanalysis 2009, 21, 280. (61) Martell, A. E.; Smith, R. M.; Motekaitis, R. J. NIST Critically Selected Stability Constants of Metal Complexes Database, release 8.0; NIST Standard Reference Database No. 46; NIST: Gaithersburg, MD, 2004.

DOI: 10.1021/la900892q

9805

Article

Goubert-Renaudin et al.

Figure 1. Chemical structure of the silylated cyclam derivatives used in this work. Table 1. Some Physicochemical Characterization Data for Materials Used in This Work (Amount of Grafted Groups, N2 Adsorption-Desorption, Average Particle Size, Reactivity toward Protons) N2 adsorption-desorption material

cyclam content (mmol g-1)a

SBET (m2 g-1)

0.52 (8)

457 316

Vp (mL g-1)

particle size (μm)

protonable ligands (%)c

capacity for Cu(II) (mmol g-1)d

0.81 224 ge (2.0023), further indicates that the magnetic orbital in the ground state corresponds to dx2-y2, as expected for a nearly axial g tensor.72,73 Altogether, theses results are consistent with copper(II) coordination geometries possessing an axial symmetry, such as a square-planar arrangement of the nitrogen donors, a square pyramid with slight equatorial distortions, or a Jahn-Teller axially elongated octahedron. The g and the A hyperfine coupling constant values, which are also recapitulated in Table 2, are close to those found for copper complexes incorporating N-functionalized cyclam units.74,75 According to the ligand field theory, the g value increases and the A value decreases as the equatorial ligand field becomes weaker or as the axial ligand field becomes stronger due to a less efficient overlap between the dx2-y2 magnetic orbital and nitrogen atom orbitals of the ligand.70,75-77 Thus, more structural information about the coordination geometry can be extracted from the ESR data by considering the g /A ratio. Roughly speaking, g /A values of 110-120 cm have been assigned to planar or axially elongated octahedral complexes, while values in the 120-150 cm range are typical for square pyramidal complexes exhibiting slight to moderate distortions. Higher values are )

)

)

)

)

moieties when increasing the number of anchorage points to the solid framework.43 This trend is also confirmed by measuring “maximal” capacity values (i.e., those obtained when contacting cyclam-functionalized materials with a large excess (>10 times) of Cu(II) over cyclam groups), as illustrated by the data collected in the last column of Table 1. These results also indicate that, contrary to metal ion binding from refluxing alcoholic solutions for which quantitative complex formation is possible,49,51 the same reaction performed in aqueous medium was always characterized by less-than-complete binding. This suggests that the accessibility to the coordination sites is not the limiting parameter but the complex formation itself. This is also sustained by the fact that the Cu(II) capacities for K60 and SBA-15 materials are not significantly different (Table 1) contrary to what was reported for other systems, for which much better accessibility was found with mesostructured adsorbent with respect to their nonordered homologues (e.g., Hg(II) binding to thiol-functionalized silica).9,10 It thus appears that an equilibrium establishes between solution-phase and cyclam-bonded Cu(II) species, indicating that immobilized cyclam-Cu(II) complexes are less stable than the same chelated metal cation in solution (log β110 = 28.1 for [Cu(cyclam)]2þ at 25
C and I = 0.1 M).61 This could be due to some conformational/flexibility differences between immobilized and solution-phase complexes, as previously suggested on the basis of UV-vis diffuse reflectance and electron spin resonance (ESR) spectroscopic measurements.47,49,53 Capacity decrease when passing from mono- to di- and to tetra-cyclam derivatives can be reasonably ascribed to more marked flexibility constraints when the macrocycle is covalently attached to the silica surface by more than one anchoring point. Interestingly, this trend follows the same order of decreasing stability constants observed upon N-methylation of the cyclam nitrogen atoms (log β110 =18.3 for [Cu(Me4cyclam)]2þ at 25
C and I = 0.1 M)61 due to the weaker ligand field stabilization energy (LFSE) provided by tertiary versus secondary amines. In addition, unfavorable electrostatic interactions in such a confined environment (the macrocyclic copper(II) complexes are positively charged) could also contribute to explain less-than-complete binding, as previously reported when forming charged species on the internal surfaces of porous organosilica materials.67,68 In order to gain structural information on the coordination scheme of the copper(II) centers embedded in the grafted cyclam derivatives but also to gauge the LFSE, the copper-loaded materials were further investigated by means of ESR spectroscopy. Since the electronic properties of the complexes are strictly dependent on the axial and equatorial donors, different spectral morphologies might be expected for the various modified silica gels. Indeed, the configuration adopted by the four nitrogen atoms in untethered cyclam complexes is strongly influenced by the presence of N-alkyl substituents, which leads to a decrease of the equatorial ligand field.69 The X-band ESR spectra of the copper(II) complexed materials recorded at 293 K are shown in Figure 3 together with the low-temperature frozen-solution spectrum pertaining to the reaction product of tetrasilylated cyclam with CuCl2. All spectra evidence magnetic anisotropy and exhibit parallel and perpendicular components typical for the S = 1/2 state of Cu(II). The coupling between the unpaired electron and the copper(II) nucleus (I = 3/2) produces the

Goubert-Renaudin et al.

(72) Hathaway, B. J.; Billing, D. E. Coord. Chem. Rev. 1970, 5, 143. (73) Pilbrow, J. R. Transition Ion Electron Paramagnetic Resonance, 1st ed.; Clarendon Press: Oxford, 1990. (74) Bucher, C.; Duval, E.; Barbe, J. M.; Verpeaux, J. N.; Amatore, C.; Guilard, R. C. R. Acad. Sci., Ser. IIc: Chim. 2000, 3, 211. (75) Dong, Y.; Lawrance, G. A.; Lindoy, L. F.; Turner, P. Dalton Trans. 2003, 1567. (76) Miyoshi, K.; Tanaka, H.; Kimura, E.; Tsuboyama, S. Inorg. Chim. Acta 1983, 78, 23. (77) Rybak-Akimova, E. V.; Nazarenko, A. Y.; Chen, L.; Krieger, P. W.; Herrera, A. M.; Tarasov, V. V.; Robinson, P. D. Inorg. Chim. Acta 2001, 324, 1.

Langmuir 2009, 25(17), 9804–9813

Goubert-Renaudin et al.

Article

Table 2. ESR Data for Copper(II) Loaded Mono-, Di-, and Tetrasilylated Materials and for the Tetrasilylated Cyclam Precursor in Frozen Solution )

g a

Cu(K60-mono)(NO3)2 Cu(K60-di)(NO3)2a Cu(K60-tetra)(NO3)2a Cu(SBA-15-tetra)(NO3)2a [Cu(tetrasilylated-cyclam)Cl]Clb

g^

)

material

)

)

A  10-4 (cm-1) g /A (cm)

2.18 2.18 2.18 2.18 2.24

2.06 191 114 2.06 190 114 2.06 190 114 2.06 190 114 2.04c 155 144 c 2.11 a Complexation was performed in water with Cu(NO3)2 at pH 4.75. The spectrum was recorded at 293 K. b The spectrum was recorded in a frozen CH2Cl2/toluene 2/1 v/v solution at 100 K. c The rhombicity of the complex 6 gy (g^ = gx = gy for complexes with a 4-fold symmetry axis). gives gx ¼

)

Langmuir 2009, 25(17), 9804–9813

)

)

)

)

)

indicative of considerable distortions or octahedral complexes with a strong axial ligand field. While the materials isolated after metalation have different colors (violet for Cu(II)-cyclam-mono, yellowish for Cu(II)-cyclam-di, and light brown for Cu(II)-cyclam-tetra), the corresponding ESR parameters collected in Table 2 are strikingly similar, suggesting that the equatorial ligand field strength remains almost unaffected by N-alkylation and grafting. With A and g /A values close for all materials to 190  10-4 cm-1 and 114 cm, respectively, either a nearly square planar or more probably an axially elongated octahedral coordination geometry with slight in plane distortion and a weak axial ligand field strength can be inferred. The ESR spectra are thus characteristic of hexacoordinated Cu(II) complexes with a strong equatorial ligand field, that is short Cu-N in plane and long Cu-OH2 axial bonds. Interestingly, the ESR data for the metalated materials (Figure 3 and Table 2) do not corroborate with those found for the copper(II) model complex formed with the tetrasilylated cyclam precursor. The corresponding spectrum recorded in frozen solution points to a distorted square pyramidal coordination geometry with a chlorine atom in axial position, in agreement with the electronic absorption data.49 Analysis of the ESR spectrum gives A and g / A values close to 155  10-4 cm-1 and 144 cm, respectively, which are typical of a square pyramidal coordination geometry of the copper(II) ion with rather strong rhombic distortions. Thus, it is clear that the in-plane ligand field strength is stronger for the grafted macrocycles than for the tetrasilylated cyclam precursor. Since the ligand field strength of grafted cyclam seems not to be influenced by the number of anchoring chains nor by the nature of the mesoporous support (K60 versus SBA-15), it follows that the capacity decrease observed when passing from mono- to di- and to tetrasilylated cyclam derivatives cannot be ascribed to differences in the apparent stabilities of the surface complexes, but perhaps to lower amounts of polyamines exhibiting a favorable conformation for complex formation with copper(II) cations. pH Influence on the Chemical Resistance of the Adsorbents. An important and unusual feature in Figure 2 is the peculiar behavior of K60-mono (curve a) compared to that of K60-di and K60-tetra (curves b and c), which indicates an apparent “resolubilization” or “desorption” of some Cu(II) that was previously accumulated on K60-mono during the initial stage. This leaching was not observed with K60-di and K60-tetra in the conditions of the experiment, for which a continuous increase in Cu(II) uptake tending to level off at longer time was obtained due to the expected progressive binding of metal ions to cyclam moieties (see curves b and c). The origin of Cu(II) leaching from K60-mono has to be found in the lower long-term integrity of this adsorbent with respect to multiarmed cyclam silica materials. Indeed, the level and rate of degradation of

Figure 4. Influence of pH on copper(II) uptake by K60-mono as a function of time ((A) two-days range; (B) expanded view for shorter times): pH = 1.5 (a), 3.2 (b), 4.2 (c), and 5.2 (d). Experimental conditions as in Figure 2 except that pH values have been adjusted by adding suitable amounts of HNO3 to 0.1 M sodium acetate solutions.

cyclam-functionalized silica gels was found to increase when passing from K60-tetra to K60-di and to K60-mono as a result of the enhanced hydrolysis tendency of grafted cyclam groups when decreasing the number of tethering arms.43 This could be ascribed to the amine-assisted alkaline hydrolysis of siloxane bonds, as otherwise reported for other polysiloxane-immobilized amine ligands.30 Figure 4A shows that Cu(II) leaching from K60mono becomes more and more important as the pH of the medium increases, a trend that agrees well with previous degradation studies.43 For instance, it was hardly noticeable at pH 3.2 (curve b), more marked at pH 4.2 (curve c), and even more at pH 5.2 (curve d). According to Figure 4A, almost no Cu(II) uptake could be observed after 50 h at pH 1.5, whereas more than 99% Cu(II) should be coordinated to cyclam at this pH on the basis of thermodynamic speciation calculations performed for the Cu(II)-cyclam system in homogeneous solution under similar conditions (see the Supporting Information, Figure S 1 and Table S 1). This is in agreement with the exceedingly slow binding process at pH 1.5, as it takes more than 3 weeks to fully form [Cu(cyclam)]2þ using the same metal and ligand concentrations (see the Supporting Information, Figure S 2). Moreover, it might also support the aforementioned hypothesis of less stable complexes once the macrocycle has been immobilized on silica in comparison to [Cu(cyclam)]2þ in the homogeneous aqueous phase. The Cu(II) leaching process was also monitored by UV-vis spectroscopy (Figure 5). All starting Cu(NO3)2 solutions were bluish. Just after suspending cyclam-functionalized silica particles (i.e., 5 min), the solution was characterized by an absorption band near 770 nm corresponding to [Cu(H2O)5]2þ cations.44 The intensity of this band was found to slightly decrease upon further Cu(II) uptake. A more impressive decrease of this signal was observed after some hours of reaction with K60-mono, concomitantly to the growing of a new absorption band at 520 nm (see curves b-d in Figure 5A), with the solution turning from blue to pink, which is the characteristic color of the [Cu(cyclam)]2þ complex.75 The appearance of macrocyclic Cu(II) complexes in solution can be related to the leaching of cyclam groups from K60-mono due to degradation of the material in aqueous medium.43 Such solution-phase Cu(II)-cyclam species appeared more slowly and in a lesser extent in the case of the K60-di adsorbent (i.e., the same signal is obtained after 24 h of reaction with K60-di but only after 3 h with K60-mono; compare curves d in Figure 5B and c in Figure 5A), while they were not detectable at all after 24 h reaction with K60-tetra. These data indicate that K60-di offers the best compromise between sufficiently high binding capacity for copper and acceptable stability, at DOI: 10.1021/la900892q

9809

Article

Goubert-Renaudin et al.

Figure 5. UV-vis spectra of the supernatant solutions after various reaction times of 2  10-4 M Cu(NO3)2 (in 0.1 M acetate buffer at pH 4.75) with (A) K60-mono, (B) K60-di, and (C) K60-tetra (solid-to-solution ratio adjusted to keep cyclam/Cu(II) ratio equal to 2); t = (a) 0, (b) 1, (c) 3, and (d) 24 h. Scheme 1. Illustration of Cu(II) Binding to Cyclam-Mono-Grafted Silica and Subsequent Evolution while Aging: Free Macrocycle Leaching and Subsequent Decomplexation of Less Stable Cu(II)-Grafted Cyclam and Formation of More Stable Cu(II)-Cyclam Complexes in Solution

least for reasonably short reaction times (approximately some hours). At this stage, one cannot state unambiguously on the mechanism leading to solution-phase Cu(II)-cyclam complexes, but “simple” leaching of the whole metal complexes from the solid seems to be unlikely as the lone electron pairs of the nitrogen atoms are involved in Cu(II) binding and are therefore no longer available for amine-assisted hydrolysis of siloxane bonds (i.e., the process responsible for degradation of amine-functionalized silica,30 via nucleophilic activation78). A more likely explanation would be the progressive leaching of nonmetalated siloxypropylcyclam moieties in solution (according to data in Figure 2, more than one-half of them remain unbound to Cu(II) cations and are thus expected to leach in solution as a result of amine-assisted hydrolysis), which are thus likely to shift the equilibrium between solid-phase-immobilized Cu(II)-cyclam complexes and more stable solution-phase complexes due to greater binding strength as discussed above. The overall uptake/release process (Scheme 1) thus involves Cu(II) binding to cyclam-functionalized materials (prevailing at shorter contact times) concomitantly with possible free macrocycle leaching (in the hierarchy K60-mono > K60-di > K60-tetra) inducing Cu(II) leaching back from the solid into the solution (via decomplexation of less stable Cu(II)-grafted cyclam and formation of more stable Cu(II)-cyclam complexes in solution, with this process becoming more and more prevalent upon increasing equilibration time and pH). (78) Corriu, R. J. P.; Dabosi, G.; Martineau, M. J. Organomet. Chem. 1978, 154, 33.

9810 DOI: 10.1021/la900892q

Figure 6 illustrates the dependence of the Cu(II) uptake kinetics as a function of the solid-to-solution ratio (adsorbent mass per solution volume) for the K60-di material, that shows the best compromise between stability and complexation capacity. As expected, increasing the adsorbent content results in higher uptake yields but this beneficial effect is not linear (see inset in Figure 6). Distribution ratios (Kd) were calculated from the data displayed in Figure 6 according to eq 1, where C0 is the Cu(II) concentration of the solution prior to extraction, C is the Cu(II) concentration after extraction, V is the solution volume (mL), and m is the mass of adsorbent added to the medium (g). The numerical values of Kd are high and increase concomitantly with the solid-to-solution ratio: 1620, 2930, and 3580 mL g-1, respectively, for cyclam/Cu(II) ratios of 1:1, 2:1, and 3:1. Interestingly, they are of the same order of magnitude, yet 2-3 times higher than those reported for the extraction of uranyl ions with related silica-gel-bound tetraazamacrocycles.38  Kd ¼

 C0 - C V C m

ð1Þ

Complete Cu(II) removal from the solution was however never observed (max. 86% removal for a 3-fold excess of cyclam over Cu(II) in the medium). In addition, increasing too much the adsorbent mass per solution volume, resulting in higher amounts of unreacted cyclam groups in the suspension, led to some degradation of the material and some Cu(II) leaching (yet in a very low amount) from the adsorbent. This starts to be visible Langmuir 2009, 25(17), 9804–9813

Goubert-Renaudin et al.

Figure 6. Influence of cyclam/Cu(II) ratio on copper(II) uptake by K60-di as a function of time: cyclam/Cu(II) ratios were (a) 1, (b) 2, and (c) 3. Experimental conditions: 100 mL of a starting solution containing 2  10-4 M Cu(NO3)2 in 0.1 M acetate buffer (pH 4.75) to which selected amounts of K60-di were added: (a) 57.1 mg, (b) 114.3 mg, and (c) 177.4 mg. Inset: Corresponding variation of maximal Cu(II) content in the solid versus equilibrium Cu(II) concentration in solution.

in curve c in Figure 6 for equilibration times longer than 48 h. Adsorbent degradation and Cu(II) leaching was faster when using increasing amounts of K60-mono (instead of K60-di) for which higher capacity values (all between 65 and 75% Cu(II) uptake) were observed after 8, 5, 3, and 1 h equilibration, respectively, for cyclam/Cu(II) ratios of 1, 2, 3, and 5 (see Figure S 3A in the Supporting Information). No Cu(II) release was observed with K60-tetra (see Figure S 3B in the Supporting Information), but this material was characterized by lower capacities, as discussed above. It is noteworthy that capacities measured on such materials keeping their chemical integrity (K60-di and K60-tetra) were independent of pH in the range between pH 3 and pH 6. Relation between the Mesostructural Order of the Adsorbent and the Copper(II) Uptake Kinetics. The resort to ordered mesoporous organosilica materials is now established as an efficient way to improve the behavior of organic-inorganic hybrids by enhancing accessibility to active centers and/or speeding up transport processes in the porous structure.10,11,24,79 One already knows from results in Table 1 that accessibility to the binding sites in cyclam-functionalized silica is not the determining factor influencing the maximum binding capacity, as approximately the same values were observed for K60 and SBA-15 materials. Now, one would like to know if dynamics (and consequently kinetics) of the uptake process can be affected by the texture of the adsorbents. Surprisingly, only little advantage of long-range organization of mesopore channels in SBA-15-mono was evidenced with respect to K60-mono (Figure 7A), no difference in Cu(II) uptake was observed when passing from K60-di to SBA-15-di (Figure 7B), and even more amazing was the behavior of samples modified with tetrasilylated cyclam since the amount of extracted Cu(II) on SBA-15-tetra was twice lower than that on K60-tetra (Figure 7C). These results indicate that structural order of the adsorbent does not provide any benefit to the kinetics of the uptake process. As enhancing the mass-transfer rates of the targeted species to cyclam binding sites via the use of mesostructured adsorbents does not improve the speed of Cu(II) uptake, the (79) Galarneau, A.; Iapichella, J.; Brunel, D.; Fajula, F.; Bayram-Hahn, Z.; Unger, K.; Puy, G.; Demesmay, C.; Rocca, J.-L. J. Sep. Sci. 2006, 29, 844.

Langmuir 2009, 25(17), 9804–9813

Article

rate-determining step is not diffusion but should be related to the complex formation itself. Steric/conformational constraints in the macrocycle are also likely to occur upon immobilization in a confined environment. For example, in tetrasilylated materials, the macrocycle attached to mesopore channels in SBA-15 is flattened against the silica surface, whereas it is more randomly distributed in the nonordered pore structure of K60,43 which could contribute to explain lower reactivity of SBA-15-tetra in comparison to K60-tetra with respect to both protonation (Table 1) and Cu(II) binding (Figure 7C). Actually, the rate of Cu(II) complexation by cyclam-type ligands in aqueous solution is known to be strongly pH-dependent, being faster at higher pH and slower at lower pH values due to protonation of the nitrogen donor atoms.33 Moreover, the amount of monoprotonated ligand at a given pH, which is the reactive species according to mechanistic studies (see the Supporting Information), is expected to decrease in the order mono- > di- > tetra-substituted cyclam due to the intrinsically lower basicity of tertiary versus secondary amines. This effect contributes to the observed kinetic trend. It is also the case for Cu(II) binding to cyclam-grafted silica gels, as illustrated in Figure 4B, which shows dramatic speeding up of the uptake by rising the pH (at the beginning of the experiment to avoid the deleterious effect of chemical degradation of the adsorbent). For example, after 1 h equilibration, 10, 30, and 49% of Cu(II) was extracted by K60-mono from 0.2 mM solutions at pH values of 3.2, 4.2, and 5.2, respectively. Interestingly, the kinetic curves of Figure 4B and the simulated concentration profiles for [Cu(cyclam)]2þ complex formation in homogeneous solutions follow very similar trends at the various investigated pH values (see the Supporting Information, Figure S 3). Even though the kinetic data pertaining to grafted and unbound cyclam units are not strictly comparable, their great similarity supports the fact that pH can be viewed as a parameter likely to tune the rate of formation of Cu(II)-cyclam complexes immobilized in the family of adsorbents studied here. Experiments as those depicted in Figure 7A have been repeated at two other pH values, namely, 3.2 and 5.7 (Figure 8), corresponding to situations where the binding process was, respectively, slow and fast. At pH 3.2, no significant differences between ordered and nonordered materials can be drawn (compare curves a and b in Figure 8), since the ratedetermining step is the slow complexation reaction that governs the whole uptake process (Cu(II) diffusion to binding sites and complex formation with cyclam moieties). These data also show that when working in conditions where K60-mono and SBA-15mono are not degradated (i.e., at pH 3.2), the shape of the uptake curves and their binding capacities became similar (but not the kinetics, of course) as for K60-di and SBA-15-di materials (compare curves a and b in Figure 8A with part B in Figure 7 and curve b in Figure 2). The situation was found to be very different at pH 5.7, where the use of SBA-15-mono (curve c in Figure 8) allows to significantly speed up Cu(II) removal in comparison to K60-mono (curve d in Figure 8). In this case, the limiting parameter appears to be mass transport to the cyclam sites instead of complex formation kinetics. This is clearly visible at the beginning of the experiment (part B of Figure 8), as longer equilibration times resulted in significant degradation of cyclammono materials as discussed above. Corresponding data obtained for K60-di and SBA-15-di materials (see Figure S 4 in the Supporting Information) show that the beneficial effect of the ordered mesoporous material was almost no more visible in this case because the rate of [Cu(cyclam)]2þ complex formation was decreasing when increasing the number of alkyl arms. One can thus conclude that the superiority of mesostructured DOI: 10.1021/la900892q

9811

Article

Goubert-Renaudin et al.

Figure 7. Influence of structural order of the adsorbent on copper(II) uptake by silica samples grafted with (A) mono-, (B) di-, and (C) tetrasilylated cyclam derivatives, expressed as a function of time. Comparison is given between (a) SBA-15 and (b) K60 materials. Experimental conditions: 100 mL of a starting solution containing 2  10-4 M Cu(NO3)2 in 0.1 M acetate buffer (pH 4.75) to which selected amounts of adsorbents were added to reach cyclam/Cu(II) ratio equal to 2: (A) (a) 71.9 mg of SBA-15-mono, (b) 66.7 mg of K60-mono; (B) (a) 76.6 mg of SBA-15-di, (b) 114.3 mg of K60-di; (C) (a) 119.0 mg of SBA-15-tetra, (b) 181.8 mg of K60-tetra.

Figure 8. Combined effect of pH and structural order of the cyclam-mono adsorbents on Cu(II) uptake, expressed as a function of time ((A) two-days range; (B) expanded view for shorter times), at pH 3.2 (a, b) and 5.7 (c, d) using either SBA-15-mono (a, c) or K60-mono (b, d). Other conditions as in Figures 2 and 4.

Figure 9. Influence of pH and adsorbent structure on the voltammetric response of cyclam-mono-grafted silica-based carbon paste electrodes: (A) K60-mono at pH 3.7; (B) SBA-15-mono at pH 3.7; (C) K60-mono at pH 6.4; (D) SBA-15-mono at pH 6.4. Preconcentration was carried out at open circuit for 5 min in a 1.0  10-6 M Cu(II) solution, and square-wave voltammograms were recorded after medium exchange to Cu(II)-free HNO3 (3 M) medium after 60 s electrolysis at a potential of -0.5 V. 9812 DOI: 10.1021/la900892q

materials over the nonordered ones from the kinetic point of view is only valid when the considered chemical reaction (complexation here, but this could be catalytic reaction or other recognition event) is not too slow in order to benefit from fast mass transport in uniform mesopore channels of monodisperse size. Among the wide range of applications of silica-based organicinorganic hybrid materials, the electroanalytical ones are clearly identified to be controlled by mass transport processes to active centers.23,24 For this reason, electrodes modified with ordered mesoporous hybrids were often much more sensitive than those based on nonordered homologues.11 Figure 9 demonstrates that this could also be the case for preconcentration electroanalysis of Cu(II) at cyclam-grafted silica-modified carbon paste electrodes (i.e., Cu(II) accumulation at open circuit followed by voltammetric detection), provided the pH of the preconcentration medium is high enough to ensure fast Cu(II) binding by immobilized cyclam units (see parts C and D in Figure 9). The more intense peak current observed for the electrodes prepared with SBA-15-mono is due to more effective Cu(II) preconcentration with respect to K60-mono as a result of faster access to the binding sites. This advantage was no longer observed at lower pH (see parts A and B in Figure 9) or when using cyclam-di or cyclamtetra derivatives instead of cyclam-mono, where the rate determining step in the preconcentration event was the Cu(II)-cyclam complex formation. The performance of the method can thus be tuned by a careful adjustment of experimental conditions that can be optimally controlled by a good knowledge of the reactivity of the adsorbent in solution. This was precisely the aim of this work.

Conclusion The silylation degree of cyclam-grafted silica adsorbents and pH are critical parameters affecting their reactivity toward Cu(II) binding. Indeed, both the binding capacities and uptake rates were found to decrease upon increasing the number of alkyl arms linking cyclam moieties to the silica framework, as a result of improved chemical resistance of these groups. Less-thancomplete metal ion uptake was systematically observed, most probably because of the lower stability of immobilized complexes relative to those in solution. As only partial loading could be achieved after 2 days of contact with the solution, the remaining uncomplexed macrocycles, especially those grafted by only one arm (cyclam-mono), were likely to leach out of the solid as a consequence of their basic properties and this chemical degradation led to significant Cu(II) desorption. Increasing pH led to significant increase of the rate of [Cu(cyclam)]2þ complex formation, but this was at the cost of loss of the chemical Langmuir 2009, 25(17), 9804–9813

Goubert-Renaudin et al.

integrity of the cyclam-mono adsorbent. The best performance was obtained with cyclam-di-grafted materials because they offer the best compromise between rather high chemical resistance while keeping a good uptake capacity. This work also provides the first direct evidence of Cu(II) leaching from polysiloxane-immobilized ligands in the course of the sorption experiment, as a consequence of the low chemical resistance of the material (i.e., K60-mono) in aqueous medium. The speed of the copper binding process was mainly controlled by the rate of complex formation (especially restrictive at low pH) rather than by mass transport restriction in the porous solids, as pointed out by comparing the reactivity of disordered silica gel and well-ordered mesoporous silica. Only in conditions where Cu(II)-cyclam complex formation was fast (at higher pH values), one can observe a little advantage of mesostructured adsorbents which are known to impart fast mass transport properties. This can be exploited to enhance the sensitivity of cyclam-grafted silica-modified electrodes with respect to Cu(II) detection.

Langmuir 2009, 25(17), 9804–9813

Article

Acknowledgment. This work was supported by the French National Research Agency (Project No. NT05-3 41602 “Mesoporelect”), the Centre National de la Recherche Scientifique (CNRS), the Ministere de l’Enseignement Sup
erieur et de la Recherche, and the Conseil R
egional de Bourgogne. We are grateful to Yoann Rousselin for providing cyclam-functionalized alkoxysilane precursors, to J
er^ome Cortot for ICP-AES analyses, and to Laure Michelin for X-ray fluorescence measurements. Supporting Information Available: Distribution diagram of copper(II) in the presence of 2 equivalents of cyclam, simulated time-dependent concentration profiles for [Cu(cyclam)]2þ complex formation in solution at various relevant pH values, Cu(II) uptake curves for K60-mono and K60-tetra showing the influence of cyclam/Cu(II) ratio as a function of time, and Cu(II) uptake curves for K60-di and SBA-15-di at various pH values. This material is available free of charge via the Internet at http://pubs.acs.org.

DOI: 10.1021/la900892q

9813