Thin Films of Cross-Linked Metallo-Supramolecular Coordination

Oct 23, 2007 - For cross-linking degrees of 9.5% and below, soluble homogeneous networks are obtained. The molar mass of the networks is large and ...
0 downloads 0 Views 181KB Size
Download PDF
Langmuir 2007, 23, 12179-12184

12179

Thin Films of Cross-Linked Metallo-Supramolecular Coordination Polyelectrolytes Torsten K. Sievers,† Annika Vergin,† Helmuth Mo¨hwald,† and Dirk G. Kurth*,†,‡ Max Planck Institute of Colloids and Interfaces, Research Campus Golm, 14424 Potsdam, Germany, and National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan ReceiVed July 20, 2007. In Final Form: August 31, 2007 We report on the synthesis of a new tristerpyridine ligand, tris(2,2′:6′,2′′-terpyridinyl-4′-oxymethyl)ethane (tritpy), as well as its introduction into metal ion induced self-assembly of cross-linked metallo-supramolecular coordination polyelectrolytes (MEPE). For cross-linking degrees of 9.5% and below, soluble homogeneous networks are obtained. The molar mass of the networks is large and depends on the cross-linking degree. Due to the charges in the MEPE, the soluble networks are suitable for film formation on the basis of layer-by-layer self-assembly and to study the details of film growth. UV-vis spectroscopy, X-ray reflectivity, AFM, and ellipsometry show that the film growth is linear and continuous. The multilayers exhibit no inner structure and have a very low surface roughness. The thickness of the adsorbed layer of MEPE networks is in the range of 3 nm. The important point is that an influence of cross-linking is not seen in multilayers, which is the opposite of what is observed for the MEPE in solution. Our experiments did not reveal an influence of the preparation procedure on the adsorption process, e.g., increasing the layer thickness.

Introduction Metallo-supramolecular chemistry explores the interactions between metal-binding sites and metal ions to assemble complex functional structures through self-organization.1 The final metallosupramolecular structure forms from individual components through stepwise self-assembly algorithms, and it is a natural choice to use self-organization for the construction of functional materials in particular, since metal ions provide value-adding properties, including magnetic, reactive, electroactive and optical ones.2 A central goal in today’s research is therefore aimed at investigating self-assembly processes in order to gain an understanding of structure and properties of metallo-supramolecular functional materials.3 An avenue to combine, position, and orient molecules by self-assembly relies on macromolecular structures based on metal ion coordination.4 The resulting composite materials are potential candidates for the construction of supramolecular materials through multistep self-assembly.5-8 The combination of polytopic terpyridine ligands and transition metal ions leads to metallo-supramolecular coordination polymers9-11 or, if they are charged, polyelectrolytes5 (MEPE). According to the classification proposed by Rehahn, these assemblies are described as centered coordination polymers, because the coordinative bond is an integral part of the polymer * Corresponding author. Phone: +49-331-567-9211. Fax: +49-331-5679202. E-mail: [email protected]. † Max Planck Institute of Colloids and Interfaces. ‡ National Institute for Materials Science. (1) Muchado, V. G.; Baxter, P. N. W.; Lehn, J. M. J. Braz. Chem. Soc. 2001, 12 (4), 431-462. (2) Zhang, S. G. Nat. Nanotechnol. 2006, 1 (3), 169-170. (3) Ball, P. Nature 2001, 409 (6818), 413-416. (4) Hofmeier, H.; Schubert, U. S. Chem. Soc. ReV. 2004, 33 (6), 373-399. (5) Lehn, J. M. Tetrahedron 2006, 62 (9), 1919-1919. (6) Kurth, D. G.; Higuchi, M. Soft Matter 2006, 2 (11), 915-927. (7) Gianneschi, N. C.; Masar, M. S.; Mirkin, C. A. Acc. Chem. Res. 2005, 38 (11), 825-837. (8) Philp, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996, 35 (11), 1155-1196. (9) Andres, P. R.; Schubert, U. S. Macromol. Rapid Commun. 2004, 25, (15), 1371-1375. (10) Schmatloch, S.; van den Berg, A. M. J.; Alexeev, A. S.; Hofmeier, H.; Schubert, U. S. Macromolecules 2003, 36 (26), 9943-9949. (11) Constable, E. C.; Housecroft, C. E.; Smith, C. B. Inorg. Chem. Commun. 2003, 6 (8), 1011-1013.

backbone.12 The properties of the composites depend on the nature of the organic ligand and the metal ions. We can control whether the material is an oligomer, a macromolecule, a gel, a porous framework,13,14 or a nanocrystal15 by the judicious choice of metal ion and ligand, the solvent and the self-assembly conditions. Through the counterions or the substitution pattern of the ligand, it is also possible to adjust the solubility or thermoand lyotropic phase behavior.6,16-20 Finally, we can control the dynamic character of these macromolecules through the metal ion and ligand design, a useful property when it comes to engineering of smart materials that can respond to external stimuli. The dependency of the molar mass of these dynamic equilibrium polymers on the concentration and stoichiometry of ligand and metal ions has been described by Stuart and co-workers.21,22 Most importantly, the molar mass depends on the binding constants, the concentration, and the stoichiometry of ligand and metal ions. Metal ion coordination of neutral polytopic terpyridines results in positively charged macroions. In a second step, we can utilize electrostatic interaction of the positive MEPEs and negatively charged interfaces to fabricate well-defined thin films.18 This so-called electrostatic layer-by-layer self-assembly technique (LbL) has been pioneered by Decher et al.23-26 Since its (12) Rehahn, M. Acta Polym. 1998, 49 (5), 201. (13) Mueller, U.; Schubert, M.; Teich, F.; Puetter, H.; Schierle-Arndt, K.; Pastre, J. J. Mater. Chem. 2006, 16 (7), 626-636. (14) Kitagawa, S.; Noro, S.; Nakamura, T. Chem. Commun. 2006, 7, 701707. (15) Uemura, T.; Kitagawa, S. Chem. Lett. 2005, 34 (2), 132-137. (16) Kurth, D. G.; Lehmann, P.; Schu¨tte, M. Proc. Natl. Acad. Sci. U.S.A. 2000, 97 (11), 5704-5707. (17) Beck, J. B.; Ineman, J. M.; Rowan, S. J. Macromolecules 2005, 38 (12), 5060-5068. (18) Schu¨tte, M.; Kurth, D. G.; Linford, M. R.; Colfen, H.; Mohwald, H. Angew. Chem., Int. Ed. 1998, 37 (20), 2891-2893. (19) Kurth, D. G.; Meister, A.; Thunemann, A. F.; Forster, G. Langmuir 2003, 19 (10), 4055-4057. (20) Lehmann, P.; Kurth, D. G.; Brezesinski, G.; Symietz, C. Chem. Eur. J. 2001, 7 (8), 1646-1651. (21) Vermonden, T.; van der Gucht, J.; de Waard, P.; Marcelis, A. T. M.; Besseling, N. A. M.; Sudholter, E. J. R.; Fleer, G. J.; Stuart, M. A. C. Macromolecules 2003, 36 (19), 7035-7044. (22) van der Gucht, J. Equilibrum polymers in solution. Ph.D. thesis, Wageningen University, Wageningen, 2004. (23) Decher, G. Science 1997, 277 (5330), 1232-1237.

10.1021/la702199d CCC: $37.00 © 2007 American Chemical Society Published on Web 10/23/2007

12180 Langmuir, Vol. 23, No. 24, 2007

publication, it has attracted considerable attention and has recently led to commercial applications of composite systems of metal nanoparticles and polyelectrolytes that are now out on the market as Metal Rubber produced by Nanosonic Inc., Blacksburg, VA. Most recently, Tsukruk et al. reported the assembly of freely suspended nanomembranes over microfabricated cavities and their possible use as a thermo-optical array for applications like IR-microimagers.27 The fabrication of thin films with dynamic metallo-supramolecular polyelectrolytes offers opportunities to make stimuliresponsive layers28 as well as electrochromic windows.29,30 Through the ligand design and the assembly conditions it will be possible to control the permeability or other film properties. For certain applications, for instance coatings or electrochromic windows,30 it is desirable to control the amount of material that is deposited on the surface. Therefore, one of our interests is to study the self-assembly of three-dimensional MEPE networks on the surface as a means to control the amount deposited by adjusting the cross-linking degree. The adsorption of linear MEPE, as an example, assembled from iron(II) and 1,4-bis(2,2′:6′,2′′terpyridin-4′-yl)benzene (tpy-ph-tpy), on surfaces results in homogeneous thin layers. It was shown that the thickness of a deposited layer of MEPE is around 1.7 nm, which corresponds approximately to a molecular layer.18,31 To assemble MEPE networks, we use an approach from classical polymer chemistry, where linear chains are connected by a cross-linker to form networks. The modular nature of metal ion induced self-assembly allows the introduction of a polytopic ligand, which results in formation of networks. Therefore, we introduce, besides the previously used ligand tpy-ph-tpy, the flexible ligand 1,1,1-tris(2,2′:6′,2′′-terpyridinyl-4′-oxymethyl)ethane (tritpy) as crosslinker. We wish to explore the condition for the formation of soluble networks that are suitable for film formation and to study the details of LBL film growth.

Results and Discussion Polymer Network Synthesis. The polymer network (Scheme 1) is synthesized by metal ion induced self-assembly of the ligands tritpy and tpy-ph-tpy and iron(II) acetate in a 75 vol % acetic acid solution. Upon adding the colorless iron(II) acetate solution to the ligand solution, the mixture immediately becomes deep purple. Initially, we observe a colloidal suspension, which clears up after heating at 70 °C for 4 h. To specify the network, we use the cross-linking degree expressed in percent, which is defined as z ) nC/(nC + nL) × 100%, where nC is the molar amount of the tritopic ligand tritpy used for cross-linking and nL the molar amount of the linear ligand tpy-ph-tpy.32 To obtain a more detailed picture of a network, one usually refers to parameters such as cross-linking density, cross-linking index, and average length of the network chains.32 Typically, these parameter are determined through indirect approaches using mechanical, rubber elastic, and swelling (24) Decher, G.; Hong, J. D. Ber. Bunsen-Ges. Phys. Chem. 1991, 95 (11), 1430-1434. (25) Decher, G.; Hong, J. D.; Schmitt, J. Thin Solid Films 1992, 210 (1-2), 831-835. (26) Decher, G.; Lvov, Y.; Schmitt, J. Thin Solid Films 1994, 244 (1-2), 772-777. (27) Jiang, C.; McConney, M. E.; Singamaneni, S.; Merrick, E.; Chen, Y.; Zhao, J.; Zhang, L.; Tsukruk, V. V. Chem. Mater. 2006, 18 (11), 2632-2634. (28) Krass, H.; Papastavrou, G.; Kurth, D. G. Chem. Mater. 2003, 15 (1), 196-203. (29) Bernhard, S.; Goldsmith, J. I.; Takada, K.; Abruna, H. D. Inorg. Chem. 2003, 42 (14), 4389-4393. (30) Kurth, D. G.; Lopez, J. P.; Dong, W. F. Chem. Commun. 2005, 16, 21192121. (31) Kurth, D. G.; Osterhout, R. Langmuir 1999, 15, 4842-4846. (32) Elias, H.-G. Makromoleku¨le; Wiley-VCH, Weinheim, 1999.

SieVers et al. Scheme 1. Metal Ion Induced Self-Assembly of the Ligands tpy-ph-tpy and tripy Results in Cross-Linked Metallo-Supramolecular Coordination Polyelectrolytes (MEPE)a

a The following nomenclature is used: depending on the molar ratio of tripy (nC moles) to tpy-ph-tpy (nL moles), the cross-linking degree in percent, z, is indicated by a prefix according to z ) nC/(nC + nL) × 100% (z% MEPE).

properties of the networks in combination with thermodynamic models such as the Flory-Huggins theory.33-35 For the soluble, dynamic networks presented here, we use the following reasoning to describe the average length of a network chain (lC) expressed as the number of ditopic ligands between cross-linking points. For simplicity, we assume that all binding sites are occupied. The value of the variable (lC) depends on the cross-linking degree (χx), as well as the topicity of both the chain building (fL) and cross-linking (fC) ligands, which are the number of tpy units in the ligand (ditopic ) two tpy units; tritopic ) three tpy units). The normality of a ligand (N) is given by N ) n‚f, where n is the molar amount of ligand and f is the topicity of this ligand. To determine the average distance in repeat units between two cross-linking points (lC) in the network, we reason that lC ) NL/NC, where NL is the normality of a linear, ditopic ligand, and NC is the normality of a cross-linking, polytopic ligand, respectively. For the system presented herein, fL is 2 and fC is 3. Next we insert the cross-linking degree (χx) given by χx ) nC/(nL + nC) into the equation. Thus, we obtain the following equation to express the number of ligands between cross-linking points, lC:

lC )

fL 1 - χx ‚ fC χx

(1)

Equation 1 states that the average number of ligands between cross-linking points depends on the relative molar amounts of connecting and cross-linking ligands and the connectivity that is the topicity of the ligands, respectively (for a more detailed description, please see the Supporting Information). We synthesize coordination networks with different degrees of cross-linking. Between 1.5% and 9% cross-linking degree, the networks are soluble in water and acetic acid solutions (Table 1). For cross-linking degrees above 9%, the solutions become heterogeneous and eventually colored solid precipitates from solution. As can be seen from the above equation, the network (33) Flory, P. J.; Rehner, J. J. Chem. Phys. 1943, 11 (11), 521-526. (34) Frenkel, J. Acta Physicochim. URSS 1938, 9 (2), 235-250. (35) Frenkel, J. Rubber Chem. Technol. 1940, 13, 521.

Metallo-Supramolecular Coordination Polyelectrolytes

Figure 1. UV-vis-spectra of 3% MEPE (solid line) in water and of tpy-ph-tpy in 1% acetic acid (dotted line). The absorption bands are assigned as follows: 585 nm, MLCT band; 372 nm, metalcentered d-d transition; 320 nm, π-π* transition of MEPE; 287 nm, π-π* transition of ligand. Table 1. Average Distance (lC) in Number of Ligands between Two Cross-Linking Points as a Function of the Cross-Linking Degree in Percent, z, Based on Equation 1, and the Solubility of Cross-Linked MEPE z, %

lC

solubility

1.5 3 4.5 6 7.5 9

44 22 14 10 8 7

homogeneous solution in water, methanol and acetic acid

4 3 2

inhomogeneous suspensions in water, methanol and acetic acid

15 20 30

chain length (lC) decreases hyperbolically as a function of the cross-linking degree, χx. At a cross-linking of 9%, the number of ligands between two cross-linking points is approximately 7 (Table 1), which corresponds to a length of approximately 8 nm. It seems that the network becomes so dense that it is no longer soluble. Above this threshold, the network keeps becoming more compact, although at a slower rate (Table 1). In this range, the dimensions of the MEPE are comparable to those of metalorganic frameworks (MOFs).36,37 In contrast to MOFs, the binding constants of terpyridines and Fe(II) are very high, which results in macromolecular assemblies,18,21,38 and as a result, cross-linking causes precipitation. Therefore, we focus on soluble MEPE networks, which are suitable for film formation. Polymer Properties. Metal ion coordination by the terpyridine ligands is discernible by the characteristic color of iron(II)bisterpyridine complexes, which is attributed to the metal-toligand charge transfer (MLCT) band located at 585 nm. The absorption band of the pyridine groups shifts upon metal ion coordination from 287 to 320 nm (Figure 1). The occurrence of the MLCT band and the shift of the pyridine π-π* transitions confirm metal ion coordination and formation of MEPE. We note that MEPE is insoluble in most organic solvents but soluble in water and aqueous acetic acid. The ligands are soluble in acidic aqueous solution, presumably as protonated species, and organic solvents but not in water. To quantify the molar mass or hydrodynamic radii, we employ analytical ultracentrifugation (AUC). These experiments are (36) Rowsell, J. L. C.; Yaghi, O. M. Angew. Chem., Int. Ed. 2005, 44 (30), 4670-4679. (37) Sudik, A. C.; Cote, A. P.; Wong-Foy, A. G.; O’Keeffe, M.; Yaghi, O. M. Angew. Chem., Int. Ed. 2006, 45 (16), 2528-2533. (38) Holyer, R. H.; Hubbard, C. D.; Kettle, S. F. A.; Wilkins, R. G. Inorg. Chem. 1966, 5 (4), 622-625.

Langmuir, Vol. 23, No. 24, 2007 12181

Figure 2. UV-vis spectra of PEI/PSS(3%MEPE/PSS)n on quartz as a function of the number of layers. The MLCT band at 594 nm indicates the incorporation of MEPE in the multilayer structure. The UV-vis spectra taken after deposition of the MEPE cycle are omitted for clarity. The insert shows the absorption maxima at 290, 327, and 594 nm as a function of the number of layers. The linear increase indicates that equal amounts are deposited in each adsorption cycle.

carried out in a 0.1 M solution of potassium acetate in order to screen the electrostatic repulsion of the positively charged network, which interferes with sedimentation.39,40 Under these conditions, we observe that MEPE networks with a cross-linking degree above 3% coagulate, become highly viscous, and form heterogeneous solutions that cannot be analyzed quantitatively by AUC. For the clear solutions of 1.5% and 3% MEPE we measured a molar mass of 246 and 174 kDa, respectively. The increase in the viscosity of MEPE solutions with a cross-linking degree above 3% indicates the formation of even higher molar masses. The solubility dependence on the ionic strength demonstrates the polyionic character of the network. Multilayer Thin Films. Our results indicate that metal ion coordination results in positively charged networks of high molar masses. Therefore, we address the question whether cross-linked MEPE are suitable for film formation based on the LbL technique. For that purpose, the cross-linked MEPE is utilized as cationic and sodium poly(styrene-4-sulfonate) as anionic species in the electrostatic layer-by-layer assembly. Film formation is monitored by UV-vis spectroscopy. Quartz slides are used as substrate. The spectra, taken after each deposition cycle, are shown in Figure 2. First, a base layer of PEI/PSS is adsorbed on the substrate to have well defined starting conditions for the layer build-up. The spectrum of the base layer is the bottom trace in Figure 2. After each deposition step, a spectrum was recorded as shown. The UV-vis spectra taken after deposition of MEPE are omitted for clarity. The inset shows the absorption as a function of the number of double layers (MEPE/PSS) for absorption maxima at 290, 327, and 594 nm. As one can see, film growth is linear and the build-up is regular. A comparison of the UV-vis spectra of the multilayer and 3% MEPE in solution reveals a red shift of all peaks (287 f 290, 320 f 327, 585 f 594 nm). A similar effect has been reported for the case of multilayers of linear MEPE containing iron(II) and tpy-ph-tpy.31 The shift is caused by an environmental effect and can be modeled with effective medium theory.41,43 (39) Schachman, H. K. Ultracentrifugation in Biochemistry; Academic: New York, 1959. (40) Fujita, H. Foundations of Analytical Utracentrifugation. Wiley: New York, 1975. (41) Norskov, J. K. Phys. ReV. B: Condens. Matter 1982, 26 (6), 2875-2885.

12182 Langmuir, Vol. 23, No. 24, 2007

SieVers et al.

Figure 4. AFM images of a silicon surface covered with a layer prepared with 3% MEPE (on the left) and 9% MEPE (on the right). The surface roughness is 0.15 nm (3% MEPE) and 0.2 nm (9% MEPE), respectively. Figure 3. X-ray reflection data for multilayer films with the following composition: (a) triangles, PEI/PSS(3%MEPE/PSS)8 (thickness 26.3 nm); (b) circles, PEI/PSS(9%MEPE/PSS)8 (thickness 26.7 nm), q ) scattering vector. The well-pronounced Kiessig fringes indicate a very uniform thickness. The upper curve is shifted along the ordinate for clarity. The films are prepared on silicon wafers.

Multilayer Thin Film Analysis. To analyze the film, we use X-ray reflectivity (XRR), optical ellipsometry, and atomic force microscopy (AFM). First, we address the results obtained from XRR and ellipsometry together and compare them afterward with the data gathered by AFM. Figure 3 shows the experimental XRR curves for two different multilayer films. The triangles resemble a film with the composition PEI/PSS(3%MEPE/PSS)8 and the circles PEI/PSS(9%MEPE/PSS)8, respectively. The upper curve is shifted for clarity. The well-pronounced Kiessig fringes indicate a very low surface roughness, which is confirmed by the fitting routine Parrat32. We determine a surface roughness of approximately 0.3 nm and an overall thickness of 26.3 nm for PEI/PSS(3%MEPE/PSS)8 and 26.7 nm for PEI/PSS(9%MEPE/ PSS)8 that is approximately 3 nm per layer pair. The curves shown in Figure 3 do not show any Bragg peaks, indicating the absence of an inner structure or stratification of the films. The steeper decay at 9% MEPE film indicates a higher roughness on the nanometer scale. These values are confirmed by optical ellipsometry. To compute a thickness from the ellipsometric data, we use the complex refractive index at 633 nm (n ) 1.5-i 0.07), which was determined previously from UV-vis spectroscopy.31 A variation of the real part of the refractive index from 1.45 to 1.55 does not significantly affect the computed film thickness.31,44 We measure the thickness for each deposited layer, by averaging the results of measurements at five different, well-separated spots on the sample. On average, we obtained 2.8 nm for one MEPE layer and 1.5 nm for the PSS layer, which is 3.3 nm per layer pair. The cross-linking degree has no influence on the layer thickness. The PSS layer thickness is in agreement with results published elsewhere.18,31 With AFM, we investigate the morphology of the films. A height image of a 3% MEPE (left) and a 9% MEPE (right) layer deposited directly on silicon is shown in Figure 4. Interestingly, we observe a very smooth surface; there is no evidence from AFM images that the films consist of networks or globular structures, as one may expect. Apparently, the MEPE network adsorbs flat on the surface. The surface roughness is determined by investigating the roughest area of a height image and is found to be 0.15 nm (3% MEPE) and 0.2 nm (9% MEPE), respectively, which is in agreement with the roughness determined from X-ray reflectance (∼0.3 nm). Also, we measure the film thickness by (42) Norskov, J. K.; Lang, N. D. Phys. ReV. B: Condens. Matter 1980, 21 (6), 2131-2136. (43) Stott, M. J.; Zaremba, E. Phys. ReV. B: Condens. Matter 1980, 22 (4), 1564-1583. (44) Kurth, D. G.; Bein, T. Langmuir 1995, 11 (8), 3061-3067.

AFM. Possible methods to determine the film thickness are either scratching the film and measuring the depth of the depression or extracting the thickness from an edge of the film. To prepare a well-defined edge, a drop of MEPE is deposited on a silicon wafer. After 15 min, the droplet is removed by an argon stream and the sample is rinsed in water for 5 min. Scanning the edge of the resulting film at different positions by AFM reveals a thickness of 2.4 ( 0.7 nm for 3% MEPE and 4.5 ( 2.9 nm for 9% MEPE. The values and the standard deviation of the data are determined from three different AFM height images. For each image we take an averaged cross section of an area of the edge of the droplet. While these results qualitatively confirm the data obtained by XRR and ellipsometry, we note an increase in the thickness; however, the standard deviation of the thickness in different samples varies quite strongly. This result might be caused by preparation. Cross-section analysis reveals that material is predominantly deposited at the edge of the droplet (data not shown). Apparently this is caused by evaporation of water. Due to this mechnism, we observe a thickness of 100-200 nm relative to the substrate in very close proximity to the edge of the droplet. The larger the distance to the edge, the smaller the thickness of the film becomes, until it reaches a constant value of ca. 3 nm. Besides this decreasing of the thickness, a higher roughness of the film is observed in the area of the border of the droplet. These two observations lead to the differences in the standard deviation, which is thus making the observed small influence of the crosslinking degree on the film thickness questionable. Notably, we do not detect a significant influence of the crosslinking degree on the layer thickness. Even the difference in the thickness of films prepared with cross-linked MEPE and linear MEPE is not remarkable. For a linear MEPE multilayer, a thickness of 1.5-2 nm has been reported.18,31 Next, we address the question of whether the preparation procedure has an influence on the adsorption process, e.g., the layer thickness. In the case of the most studied polyelectrolyte combination, PSS/PAH, the variables that affect LBL film growth are well-documented.45 These variables are salt concentration, which influences the film thickness; adsorption time, which has to be sufficient to facilitate the formation of a complete layer; and pH value, which defines the polyelectrolyte charge density, to mention the most prominent parameters. We reason that these variables also apply to the MEPE/PSS layer growth. In general, the main parameters are solvent type, temperature, and time, in addition to cross-linking degree, concentration, ionic strength, and pH in each step. In order to understand how we can control the deposited amount of material in each step, we investigate the influence of the experimental conditions on the film build-up. We analyze the influence of time, solvent type, and temperature in the adsorption and the rinsing step, respectively. Additionally, (45) von Klitzing, R. Phys. Chem. Chem. Phys. 2006, 8 (43), 5012-5033.

Metallo-Supramolecular Coordination Polyelectrolytes

Figure 5. UV-vis spectra of the multilayer before and after rinsing. The solid line corresponds to PEI/PSS(3%MEPE/PSS)3, the dashed line to PEI/(PSS/3%MEPE)4 before rinsing, and the dotted line to PEI/(PSS/3%MEPE)4 after rinsing for 10 min with water at room temperature. A decrease of the absorption of approximately 10% is observed upon rinsing the sample. The dotted-dashed line corresponds to a PEI/PSS base layer.

we verify whether the addition of salt to the rinsing solvent helps reducing the amount of MEPE dissolved in the rinsing step. The precise experimental conditions are given in Table 1 in the Supporting Information. The analysis of the surface coverage is carried out by UV-vis spectroscopy due to the strong MLCT band of MEPE. We note that the variation of these parameters, e.g., changing the adsorption time from 10 to 45 min, has very little effect on the deposited amount. A representative set of data is given in Figure 5. The UV-vis spectra are recorded from the same sample. The solid line corresponds to a PEI/PSS(3%MEPE/PSS)3 multilayer. The next spectrum is taken directly after adsorption of an additional layer of 3% MEPE (dashed line), deposited from 10-3 M aqueous MEPE solution at room temperature, and an adsorption time of 15 min. We observe an increase of the absorption at 594 nm from 0.096 to 0.12 (25%). After rinsing the sample in water for 10 min at room temperature (spectra represented by the dotted line), we see a decrease of absorption of approximately 10% from 0.12 to 0.11. This result leads to the conclusion that already in the adsorption step itself the amount of material deposited is limited and not remarkably higher than in the case of the linear MEPE.31 The results for all experiments carried out to investigate the influence of the experimental conditions are similar regarding the deposited amount determined from the adsorption data. Interestingly, rinsing removes only a small fraction of the adsorbed material, which is mostly loosely bound, physisorbed material. Also, the films show no dissolution if exposed to water after rinsing. The important point is that, despite the 3D nature of the network in solution, the amount of material deposited in the adsorption step is apparently finite. A possible explanation for this observation is either that charge compensation at the interface limits the growth, that decomposition/disassembly of the MEPE occurs at the interface, or that the MEPE network is not entangled enough at the interface to enable the irreversible adsorption of a thick layer.

Conclusion Iron(II) ion induced self-assembly of the ligands tritpy and tpy-ph-tpy results in cross-linked MEPE networks. For crosslinking degrees of 9% and below, water-soluble, homogeneous polymer networks are obtained. The solution of these polymer networks shows UV-vis spectra similar to the one known for linear MEPE, which indicates the formation of metal ion coordination bonds.18,19 The molar mass of the networks is large as far as we can determine it in solution and depends on the cross-linking degree. Soluble cross-linked MEPEs are suitable

Langmuir, Vol. 23, No. 24, 2007 12183

for build-up of multilayers by the layer-by-layer-technique. UVvis spectroscopy, X-ray reflectivity, AFM, and ellipsometry show that the film growth is linear and continuous. The multilayers exhibit no inner structure and have a very low surface roughness. The thickness of the adsorbed layer of MEPE networks is comparable to the one of linear MEPE. The important point is that an influence of cross-linking is not seen in the multilayer system, which is the opposite of what is observed for the MEPE in solution. The question, which has to be addressed next, is whether this effect is a principle one that is valid for cross-linked networks or whether it relies on the dynamic nature of the MEPE network, which may disassemble upon adsorption to the surface. A theoretical description indicates that, compared to the volume phase, the interfacial region is depleted of polymers.22 Regarding the concentration dependency for MEPE, this means that the molar mass of MEPE decreases during the approach to the surface.22 Experimental Section General Remarks. The chemicals including anhydrous DMSO were purchased from Sigma-Aldrich and used without further purification. All other solvents were obtained from Merck and used without purification. Only Millipore water, obtained from a Purelab filter system, with a conductivity of 0.055 µS/cm was used. UV-vis spectra were measured on a Varian Cary50 Conc spectrophotometer. UV-vis spectra of films were recorded on quartz substrate. NMR spectra were recorded on a Bruker DMX400. Mass spectra were recorded on a Varian MAT 771 and 112S as FAB-MS. X-rayreflectivity was measured on a Stoe 2Θ/Θ instrument. The measurements were made on a silicon wafer with a thickness of 675 µm obtained from Silchem. The reflectivity spectra were analyzed by applying the standard fitting routine Parratt 32.46 AFM measurements were carried out on a multimode AFM from Veeco Instruments in tapping mode on silicon wafers. Elemental analysis was done with a VarioEL CHNO by Elementar Analysesysteme GmbH. Melting points were determined with a VEB Waegetechnik Radebeul melting point microscope. Silicon and quartz wafers were cleaned with chloroform, acetone, ethanol, water, and hot peroxomonosulfuric acid (30% hydrogen peroxide/concentrated sulfuric acid in a volume ratio of 1/3, also called piranha solution). 1,1,1-Tris(2,2′:6′,2′′-terpyridinyl-4′-oxymethyl)ethane (tritpy). 1,1,1-Tris(hydroxymethyl)ethane (131.5 mg, 1.09 mmol) was suspended in 40 mL of dried dimethyl sulfoxide (DMSO) containing powdered potassium hydroxide (921 mg, 16.42mmol). The suspension was stirred for 10 min at room temperature. 4′-Chloro-2,2′: 6′,2′′-terpyridine (1172 mg, 4.36mmol) was added, and the mixture was stirred for 72 h at 70 °C. The color changed from orange to red-brown. The reaction mixture was stirred for an additional 24 h at room temperature, and 50 mL of water was added. A pale yellow precipitate was obtained by filtration of the reaction mixture. The solid residue was washed with 20 mL of water, 20 mL of methanol, and 5 mL of diethyl ether, to give a white powder, which was dried in vacuo. Yield: 448 mg (0.55 mmol), 50.3%. Mp: 215 °C. 1H NMR (400 MHz, CDCl3, δ): 8.67 (d, J3 ) 4.46 Hz, H6, H6′′, 6H), 8.57 (d, J3 ) 7.39 Hz, H3, H3′′, 6H), 8.07 (s, H3′, H5′, 2H), 7.80 (t, J3 ) 7.71 Hz, H4, H4′′, 6H), 7.28 (m, H5, H5′′, 6H), 4.47 (s, H7, 2H), 1.48 ppm (s, H9, 1H). 13C NMR (100 MHz, CDCl3, δ): 166.85 (C4′), 157.12 (C2′, C6′), 155.96 (C2, C2′′), 148.98 (C6, C6′′), 136.66 (C3′, C5′), 123.71 (C3, C3′′), 121.26 (C4, C4′′), 107.36 (C5, C5′′), 68.13 (C7), 40.52 (C8), 17.32 ppm (C9). Anal. Calcd: C, 73.78; H, 4.83; N, 15.49. Found: C, 72.14; H, 4.53; N, 15.04. MS-FAB m/z (relative intensities): [M + K]+ ) 852 (5), [M + Na]+ ) 836 (55); [M + H]+ ) 814 (70), [C16H12N3O]+ ) 262 (15); [C15H10N3]+ ) 232 (18), [C5H4N]+ ) 78 (45). (46) Braun, C. Parrat32 or The ReflectiVity Tool; Hahn-Meitner-Institut: Berlin, 1997-1999.

12184 Langmuir, Vol. 23, No. 24, 2007 1,4-Bis(2,2′:6′,2′′-terpyridin-4′-yl)benzene (tpy-ph-tpy) was synthesized according to a literature procedure starting with 2-acetylpyridine and benzene-1,4-dialdehyde (terephthalaldehyde).47,48 In addition to the published procedure, the obtained product was recrystallized from glacial acetic acid three times and washed with ethanol, and the solid was dried in vacuo. Cross-Linked Metallo-Supramolecular Coordination Polyelectrolyte (MEPE). For the synthesis of cross-linked MEPE, a general method was established, as described below. To obtain different degrees of cross-linking, the molar ratio of tpy-ph-tpy, tritpy, and Fe(OAc)2 was adjusted according to the following principle: nL moles of tpy-ph-tpy, nC moles of tritpy, (nL + 1.5nC) moles of Fe(OAc)2. The following nomenclature is used: Depending on the molar ratio of nL moles of tpy-ph-tpy and nC moles of tripy, the cross-linking degree is added as prefix according to z ) nC/(nC + nL) × 100% (z% MEPE). Tpy-ph-tpy (155 mg, 0.287mmol) and tritpy (311 mg, 0.014mmol) were dissolved in 25 mL of 75% acetic acid. A solution of iron(II) (47) Constable, E. C.; Thompson, A. M. W. C. J. Chem. Soc.-Dalton Trans. 1992, 24, 3467-3475. (48) Kro¨hnke, F. Synthesis 1976, 1, 1-24.

SieVers et al. acetate (53 mg, 0.307mmol) in 25 mL of 75% acetic acid was added, while stirring vigorously. The reaction mixture turned immediately from yellow to dark purple and was kept at 70 °C for 4 h. The solvent was allowed to evaporate and the obtained solid was dissolved in water. After filtration, the solvent was evaporated to give a purple, metal-like, shiny solid. Yield: 213 mg (0.291mmol monomer), 97.1% (based on monomer). UV-vis (water): λmax, nm () ) 287 (39 071), 320, (51 723), 372 (12 648), 585 (30 022).

Acknowledgment. The authors thank C. Stolle for her help with the preparation of chemicals, A. Vo¨lkel and H. Co¨lfen for their assistance with AUC measurements, and A. Heilig for AFM measurements. This work was supported by the collaborative research center 448 (SFB 448) of the German research society (DFG). Supporting Information Available: A detailed derivation of a formula to calculate the average distance in monomers between two cross-linking points and all experimental parameters for the investigation of the influence of preparation on the films. This material is available free of charge via the Internet at http://pubs.acs.org. LA702199D