Preparation Parameter Development for Layer-by-Layer Assembly of

Sep 19, 2007 - ... Calum Dickinson , Fathima Laffir , Mikhail Vagin , and Timothy McCormac .... Subrata Ghosh , Satinder K. Sharma , Kenneth E. Gonsal...
0 downloads 0 Views 242KB Size
11120

Langmuir 2007, 23, 11120-11126

Preparation Parameter Development for Layer-by-Layer Assembly of Keggin-type Polyoxometalates Bin Wang,* Ritesh N. Vyas, and Shafi Shaik Department of Chemical Engineering, Lamar UniVersity, P.O. Box 10053, Beaumont, Texas 77710 ReceiVed June 18, 2007. In Final Form: July 31, 2007 Polyoxometalates possess many useful properties for electrochemical catalysis. These molecule-size clusters can be assembled into thin films through the layer-by-layer method. In this study, we determined a cluster concentration range within which layer-by-layer (LbL) films have been successfully fabricated. We also find the influence of salt added to the deposition solutions. In an attempt to understand the self-assembly process at the molecular level, thermodynamic arguments, derived from complexation between nanoscale particles and oppositely charged polyelectrolyte chains, have been employed to interpret the adsorption of polyoxometalate clusters onto a cationic polymer layer. The scaling results describe the contact mode between a polymer chain and a cluster. The assembly can be visualized with assistance by understanding the contact between the polymer chain and the cluster.

Introduction Polyoxometalates (POMs), a class of Angstrom-scale anionic clusters with much diversity in size, composition, and function, have been attracting increasing attention in recent years. Their interesting properties include the high stability of most of their redox states, the possibility of tuning their redox potentials by changing the heteroions and/or addenda ions without affecting their structure, the variability of the transition-metal cations which can be incorporated into the hetero-polyoxometalate structure, and the possibility of multiple electron transfer.1 These features make POMs attractive as redox catalysts for electrochemical processes. Application of POM-based materials often requires some methods to orient and integrate the clusters into the device architecture. Dip coating, the Langmuir-Blodgett technique, electrodeposition, and doping in conducting polymers have been used to form useful materials.2 Polymer matrices are excellent support to entrap POMs for applications in heterogeneous and electrochemical catalysis. Entrapment can be introduced during or after the polymer matrix formation process. Electrostatic layer-by-layer (LbL) assembly has been used to fabricate homogeneous, ultrathin films with thickness at the nanometer scale. Because most POM clusters are water soluble, they are ideal candidates for the LbL assembly technique. There has been increasing interest in constructing POM thin films through the LbL process. Many factors influence the preparation process of polyelectrolyte LbL films, including polymer type, molecular weight, concentration, deposition time, salt type and concentration, pH, and solvent composition.3 Among them, salt concentration has been shown to have the most predominant effect on the resultant film structure. These same factors may have profound effects on the ultrathin films containing POM clusters. Because POM clusters have vastly varied sizes (ranging from a few dozen atoms to a few hundred), the conditions employed in building POM-containing LbL films have been modified to distinct, individual patterns. A brief survey of the * To whom correspondence should be addressed. Phone: (409) 8807709. Fax: (409) 880-2197. E-mail: [email protected]. (1) Sadakane, M.; Steckhan, E. Chem. ReV. 1998, 98, 219-37. (2) Coronado, E.; Go´mez-Garcia, C. J. Chem. ReV. 1998, 98, 273-96. (3) (a) Schlenoff, J. B.; Ly, H.; Li, M. J. Am. Chem. Soc. 1998, 120, 7626-34. (b) Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153-60. (c) Schlenoff, J. B.; Dubas, S. T. Macromolecules 2001, 34, 592-8. (d) Jomaa, H. W.; Schlenoff, J. B. Macromolecules 2005, 38, 8473-80.

literature reveals a few major types of conditions that have been successfully applied. Some methods use relatively concentrated POM solutions (g5 mM) during the deposition, while others apply less concentrated solutions (∼1 mM). When the more concentrated POM solutions are used, the solutions are usually kept at low pH (∼2-3).4 Accordingly, one variation is to wash the deposited slides with dilute HCl solution.5 When the POM is used in dilute solutions, both lower (∼1-3.5)6 and slightly higher (∼5-6)7 pH ranges have been applied. Variable salt concentrations have been used during the deposition as well.8 There are some reports that apply electrochemistry to deposit POM clusters onto an electrode surface.9 However, the nature of electrodes used in electrochemistry limits the application scope of the LbL method. The LbL method has been employed to provide functional structures with applications such as electrochromic thin films, photovoltaics, ionic-conducting systems, cell templating, and drug delivery systems. Nonlithographic patterning has been incorporated to afford two- and three-dimensional structures on many types of substrates.10 Some large-scale techniques such as roll-to-roll processing have also become available for LbL construction. It is thus more interesting to use a dip coating fashion to produce LbL structures. We attempt to find a set of generic fabrication parameters that applies to most (4) (a) Jiang, M.; Wang, E.; Wei, G.; Xu, L.; Li, Z. J. Colloid Interface Sci. 2004, 275, 596-600. (b) Zhang, G.; Dong, X.; Yang, W.; Yao, J. Thin Solid Films 2006, 496, 533-8. (c) Gu, N.; Wei, D.; Niu, L.; Ivaska, A. Electrochim. Acta 2006, 51, 6038-44. (5) (a) Moriguchi, I.; Fendler, J. H. Chem. Mater. 1998, 10, 2205-11. (b) Liu, S.; Tang, Z.; Wang, Z.; Peng, Z.; Wang, E.; Dong, S. J. Mater. Chem. 2000, 10, 2727-33. (c) Feng, Y.; Han, Z.; Peng, J.; Lu, J.; Xue, B.; Li, L.; Ma, H.; Wang, E. Mater. Lett. 2006, 60, 1588-93. (6) (a) Ichinose, I.; Tagawa, H.; Mizuki, S.; Lvov, Y.; Kunitake, T. Langmuir 1998, 14, 187-92. (b) Wang, Y.; Guo, C.; Chen, Y.; Hu, C.; Yu, W. J. Colloid Interface Sci. 2003, 264, 176-83. (c) Ma, H.; Peng, J.; Han, Z.; Feng, Y.; Wang, E. Thin Solid Films 2004, 446, 161-6. (7) (a) Liu, S.; Kurth, D. G.; Volkmer, D. Chem. Commun. 2002, 976-7. (b) Liu, S.; Kurth, D. G.; Bredenko¨tter, B.; Volkmer, D. J. Am. Chem. Soc. 2002, 124, 12279-87. (c) Wang, Y.; Wang, X.; Hu, C. J. Colloid Interface Sci. 2002, 249, 307-15. (8) (a) Liu, S.; Kurth, D. G.; Mo¨hwald, H.; Volkmer, D. AdV. Mater. 2002, 14, 225-8. (b) Liu, S.; Mo¨hwald, H.; Volkmer, D.; Kurth, D. G. Langmuir 2006, 22, 1949-51. (c) Wang, Y.; Wang, X.; Hu, C.; Shi, C. J. Mater. Chem. 2002, 12, 703-7. (9) (a) Kulesza, P. J.; Chojak, M.; Miecznikowski, K.; Lewera, A.; Malik, M. A.; Kuhn, A. Electrochem. Commun. 2002, 4, 510-5. (b) Kulesza, P. J.; Chojak, M.; Karnicka, K.; Miecznikowski, K.; Palys, B.; Lewera, A.; Wieckowski, A. Chem. Mater. 2004, 16, 4128-34. (10) Park, J.; Hammond, P. T. AdV. Mater. 2004, 16, 520-5.

10.1021/la701789n CCC: $37.00 © 2007 American Chemical Society Published on Web 09/19/2007

Assembly of Keggin-type Polyoxometalates

Figure 1. Typical polymer configurations obtained at a sphere. (a) Polymer chain contacts the sphere at one point. (b) Polymer chain wraps around the sphere over a finite length. The structures of PDDA and Keggin polyoxometalate are also shown.

POM clusters. In this study, we use a strong polyelectrolyte poly(diallyl dimethyl ammonium chloride), PDDA (Figure 1), as the cationic polymer. We choose two Keggin-type POM clusters (Figure 1) to study their layer-by-layer deposition behavior. Because of the profound importance of salt concentration in LbL structures, we use salt concentration as a variable during the deposition process. The understanding acquired from this study can be easily applied to other more specific materials and LbL requirements.

Background The deposition mechanism from a polyanion/polycation pair to form layer-by-layer ultrathin films has been studied extensively by various groups. We use a recent contribution from Schlenoff’s group to understand some fundamental aspects.3 The driving force for the buildup of LbL films is the release of counterions, similar to polyelectrolyte complexes formation. Because the ions are well hydrated, release of waters of hydration causes hydrophobicity between polyelectrolyte segments in the polymer complex. Electrostatic neutrality is maintained by combined polyion pairing and counterions. The authors coined the term “intrinsic” charge compensation to describe the situation when a positive polymer charge is balanced by a negative polymer charge. The term “extrinsic” denotes the polymer charges balanced by salt counterions. In most LbL ultrathin films formed from two oppositely charged polyelectrolytes, intrinsic ion pairing overwhelmingly dominates the bulk structure. A polyelectrolyte coil is swollen in aqueous solutions due to the internal electrostatic repulsion between its charged segments. When salt is added, this repulsion is screened, the polymer coil shrinks, and polyelectrolyte chains adapt loose conformations including more loops and tails. Thus, thin layers are formed from low ionic strength solutions when polyelectrolyte chains adsorb onto the surface as coiled trains. If the ionic strength of the deposition solutions is high enough, adaptation of the dangling loop mode by more polymer segments increases the thickness of each layer. However, even in the relatively flat, coiled structures, polymer interpenetration is a universal phenomenon. Penetration of one polyion into the bulk structure has been observed to decay exponentially and can extend to two to three bilayers. The interaction between polyelectrolytes and charged surface is depicted by different interpretations. Mayes and co-workers developed a “sticker” theory describing polyelectrolyte adsorption onto charged surfaces.11 Adsorption may be portrayed as anchoring a monomer site onto a matching discrete (“sticker”) site on the surface which has an average dimensionless sticker density. When the average (11) (a) Park, S. Y.; Barrett, C. J.; Rubner, M. F.; Mayes, A. M. Macromolecules 2001, 34, 3384-8. (b) Park, S. Y.; Rubner, M. F.; Mayes, A. M. Langmuir 2002, 18, 9600-4.

Langmuir, Vol. 23, No. 22, 2007 11121

distance between stickers lies between that of the segment size and the adsorbing chain dimension, supermonolayers can be deposited. The adsorbed coil can be seen as creating a higher effective polymer concentration in the near-surface region. Multilayer stability is largely governed by polymer charge density and solution ionic strength, with high charge densities and low ionic strengths favoring multilayer formation. The incremental layer thickness scales linearly with salt concentration, with fixed concentration of preadsorbed layer and the existence of a loop region next to the inner bulk of the adsorbing layer. Nanoparticles with dimensions in the 1-100 nm range can be assembled with polyelectrolytes through the layer-by-layer method.12 Both ionic and van der Waals interactions have significant contributions to depositing nanoparticles, most waterbased and highly charged, into LbL films. A combination of strong electrostatic attraction and firm binding to the surface polyelectrolyte layer makes the adsorbed nanoparticle layer thermodynamically more preferable than the dissolved state. Yet, polyoxometalate clusters are structurally distinctive compared to nanoparticles such as CdSe, TiO2, SiO2, and Au. Many nanoparticles often do not have narrowly distributed sizes and charge densities from batch to batch. This is in stark contrast to polyoxometalates, which are discrete, molecular clusters with fixed sizes and charges. This unique structural characteristic makes it plausible to scale the LbL deposition process with POM clusters. Trivalent ions seem to be limiting cases to make LbL structures.13 Therefore, Keggin-type POM clusters PMo12O403and SiMo12O404- are chosen here. We use both PMo12/SiMo12 and PMo123-/SiMo124- to indicate these molecular ions. Deposition of POM via the LbL method is treated analogous to complex formation by a spherical particle and a polyelectrolyte of the opposite sign.14 The precise nature of the complex depends on such parameters as the charge of the sphere Z, the sphere radius D, the linear charge density of the polymer τ, the ionic strength (parametrized by the screening length κ-1), and the flexibility of the polymer (measured by an intrinsic persistence length l0). The electrostatic repulsion between monomers is determined by the Debye-Hu¨ckel potential in which the Bjerrum length lB measures the separation at which the interaction between two elementary charges is on the order of the thermal energy kT. lB takes a value of ∼0.7 nm in water.14 The screening length is determined by the salt concentration c: κ2 ) 8πlBc for monovalent salt. κ-1 for a 0.1 M monovalent salt solution is about 1 nm.14 At dilute conditions, e.g., 0.1 mM (of the electrolyte concentration from the complex formation components), the reciprocal of square root of the screening length becomes ∼33 nm. The bending energy of the polymer and electrostatic interactions between the monomers (repulsive) and between the polymer and the sphere (attractive) comprise the complexation free energy. The attractive electrostatic interaction between the polymer and the sphere can be monitored not only by varying the polymer line-charge density and the sphere charge but also by changing the ionic strength of the solution, which is the variable that we employ in the experimental design. For a very small sphere (D , κ-1), a weakly deformed polymer is in contact with the sphere at one point (Figure 1a). For intermediate sphere charge and rather low salt concentration, a bent polymer touches the sphere over a finite segment of its length (Figure 1b). For large sphere charge and intermediate salt (12) Kotov, N. A. In Multilayer thin films: Sequential assembly of nanocomposite materials; Decher, G., Schlenoff, J. B., Eds.; Wiley-VCH: Weinheim, Germany, 2003; Chapter 8. (13) Bertrand, P.; Jonas, A.; Laschewsky, A.; Legras, R. Macromol. Rapid Commun. 2000, 21, 319-48. (14) Netz, R. R.; Joanny, J. F. Macromolecules 1999, 32, 9026-40.

11122 Langmuir, Vol. 23, No. 22, 2007

Wang et al.

Table 1. Preparation Parameters Used in Layer-by-Layer Construction of (POM|PDDA)10 Films: 1-6, PMO12; 7-12, SiMo12 sample

POM (mM)

pH of POM

pH of PDDA

NaCl (M)

1 2 3 4 5 6 7 8 9 10 11 12

5 1 0.1 5 1 0.1 5 1 0.1 5 1 0.1

2.2 2.6 3.3 2.2 2.6 3.3 2.2 2.7 3.4 2.2 2.7 3.4

2.2 2.6 3.3 2.2 2.6 3.3 2.2 2.7 3.4 2.2 2.7 3.4

0 0 0 0.1 0.1 0.1 0 0 0 0.1 0.1 0.1

concentration, the polymer completely wraps around the sphere.14 The size of POM clusters makes them fall in the range of small sphere, meaning there will be no wrap around by polyelectrolyte chains over a cluster. The transition from one point contact to a finite segment length can be estimated by comparing two sphere charges Zt and Z. For a small sphere charge Z the polymer curvature is smaller than the sphere curvature 1/D. The sphere charge Zt, at which the polymer and the sphere curvature are equal, determines the touching transition14

Figure 2. UV-vis absorption of LbL buildup for (PMo12|PDDA)10, sample 1. (Inset) Absorption growth versus number of bilayers at wavelengths 215 and 310 nm.

[ ( )]

Zt ≈ τD hl 0 ln

1 hl 0 D2κ2

1/2

For Z < Zt, the polymer contacts the sphere only at one point; for Z > Zt, it touches the sphere over a finite length. The intrinsic persistence length l0 of PDDA has a value of 10.0 nm, and the linear charge density is τ ≈ 0.73 nm-1.15 The diameter of PMo12/ SiMo12 clusters is 1.1 nm.16 When the LbL deposition is performed with no added salt, the ionic strength is from the POM electrolyte (0.1 mM) and its counterion (0.3 mM for PMo123- and 0.4 mM for SiMo124-). Thus, the calculated values are Zt ≈ 2.59 for PMo123- and Zt ≈ 2.48 for SiMo124-, where screening length value is taken as a function of the ionic strength: κ2 ) 8πlBc. When the LbL deposition is carried out with added salt (0.1 M NaCl), the intrinsic persistence length has to be estimated with additional information. A recent publication estimates that persistence length values are very close at ionic strengths between ∼0.3 and ∼30 mM.17 Thereafter, we can safely state that the Zt values are close to ∼2.5 for conditions both with and without salt added during the deposition, which are smaller than Z ) 3 for PMo123- or 4 for SiMo124-. The scaling results suggest that PDDA may have contact with PMo123-/SiMo124- over a finite length during the LbL fabrication. The above scaling helps predict some features of the layerby-layer preparation of POM-containing films. First, we can intuitively compare the PDDA persistence length versus the POM radius. The POM radius is much smaller than the PDDA persistence length, i.e., 0.55 vs 10 nm. It is thus very unlikely to bend the polymer chain to a significant degree around a POM sphere. Even with added salt, which “softens” the polymer chains, the difference still remains too large to overcome. Plus, PDDA has a charge density of ∼0.73 nm-1, so a half-sphere wrap of a cluster by a PDDA chain will bring ∼2.4 positive charges to the cluster. There is no apparent short-range interaction between a POM cluster and a PDDA chain, and the LbL deposition of (15) Pallandre, A.; Moussa, A.; Nysten, B.; Jonas, A. M. AdV. Mater. 2006, 18, 481-6. (16) Klemperer, W. G.; Wall, C. G. Chem. ReV. 1998, 98, 297-306. (17) Hsiao, P. Y. Macromolecules 2006, 39, 7125-37.

Figure 3. UV-vis absorption of LbL buildup for (SiMo12|PDDA)10, sample 11. (Inset) Absorption growth versus number of bilayers at wavelengths 204 and 305 nm.

POM/PDDA is viewed to predominantly rely on long-range electrostatic interaction. For this type of LbL films, surface charge overcompensation is the key to sequential deposition of oppositely charged components. Thus, a half-sphere wrap by PDDA renders a PMo12 cluster slightly negatively charged, which is still attractive toward a PDDA layer. From this reasoning we estimate that in a POM|PDDA multilayer structure the polymer chain in contact with a POM cluster is likely less than a half-wrap, i.e., two repeat units. Yet, the deduced polymer chain/cluster interfacial mode does not exclude any one point contact configuration.

Results UV-Visible Absorption Analysis. Absorption from O f Mo ligand-to-metal charge transfer in POM clusters is used to monitor the layer-by-layer growth of POM|PDDA films on quartz slides. Like a typical LbL process, we observe monotonic increment for both PMo12|PDDA and SiMo12|PDDA film buildup. These films were prepared under conditions with two primary variables, POM concentration and solution ionic strength. The preparation conditions are summarized in Table 1. During preparation UVvis absorption was recorded after each bilayer deposition. Two UV-visspectra,representingsamplenumber1for(PMo12|PDDA)10 and sample number 11 for (SiMo12|PDDA)10, are shown in Figures 2 and 3, respectively. UV-vis spectra for the remaining samples can be found in the Supporting Information. From the UV-vis spectra we observe monotonous growth of absorption versus bilayers deposited, regardless of POM cluster concentration or solution ionic strength. Some spectra show more even growth than others. However, we attribute the less uniform growth to

Assembly of Keggin-type Polyoxometalates

Langmuir, Vol. 23, No. 22, 2007 11123

Table 2. UV-vis Absorption, Molar Surface Coverage (×10-10 mol‚cm-2), and Monolayer Equivalent Values for (POM|PDDA)10 Films: 1-6, PMO12, A215; 7-12, SiMo12, A204 sample

absorption

molar coverage

equivalent coverage

1 2 3 4 5 6 7 8 9 10 11 12

0.301 0.283 0.260 1.044 0.702 0.310 0.353 0.318 0.255 1.093 0.737 0.451

1.00 0.94 0.87 3.48 2.34 1.03 0.9 0.8 0.7 2.9 1.9 1.2

0.8 0.8 0.7 2.8 1.9 0.8 0.9 0.9 0.8 0.7 2.9 1.9

Figure 4. Cyclic voltammogram of LbL buildup for (PMo12|PDDA)10, sample 1. (Inset) Second anodic peak current versus number of bilayers deposited.

experimental irregularities because the enormous amount of work was performed manually. The important findings here are that there is no apparent exponential growth. In aqueous solution the molar absorption coefficient for SiMo12 is ∈204 ) 7.6 × 104 M-1‚cm-1; for PMo12, ∈215 ) 7.5 × 104 M-1‚cm-1. Surface coverage per layer is calculated according to the equation5a,7b

Γ)

N AA λ 2m∈λ

where NA is Avogadro’s constant, Aλ is the absorbance, ∈λ is the molar absorption coefficient (M-1‚cm-1), and m is the number of layers. For example, sample 1, (PMo12|PDDA)10 prepared from 5 mM PMo12 solution without salt, has an absorption A215 ) 0.301. The LbL buildup is two sided, so the molar surface coverage is Γ ) 1.00 × 10-10 mol‚cm-2. One can assume a close geometric packing of POM clusters to calculate the coverage of POM clusters on a surface. We use data from scanning tunneling microscopy observation of Keggin clusters deposited on a gold surface to estimate a monolayer coverage at 1.25 × 10-10 mol‚cm-2.16 This estimation takes into account the probable equilibrium separation of clusters, not the collision diameter value detected by X-ray crystallography. The UV-vis absorption, molar surface coverage, and monolayer equivalent values are presented in Table 2. Cyclic Voltammetry Analysis. The electrochemistry of POM clusters embedded in PDDA matrix is demonstrated through cyclic voltammetry (CV). All CV curves were recorded using Ag/AgCl as the reference with a scan rate of 50 mV‚s-1. Two CV curves, representing sample 1 for (PMo12|PDDA)10 and sample 11 for (SiMo12|PDDA)10, are shown in Figures 4 and 5, respectively. CV spectra for the remaining samples can be found in the Supporting Information. The scan range is from -0.2 to 0.7 V for both PMo12|PDDA and SiMo12|PDDA films on indium tin oxide (ITO) coated glass slides (2 cm2). The ITO slides were cleaned according to a mild procedure.18 As a result we observed different base curves for the cleaned ITO. The base curves were later found to influence the curve shape of deposited POM|PDDA films, i.e., the upward or downward shifts of waves. Overall, construction of LbL films of POM clusters is evident by studying the sequential cyclic voltammograms. The continuous growth of the film after every deposition cycle indicates a monotonous increment by adding a bilayer of POM|PDDA. Here, we focus the analysis of CV curves on the redox potentials and number of transferred electrons associated with each wave. The redox curves for PMo12 appear as oxidations at -0.05, 0.18, and 0.37 (18) Fou, A. C.; Onitsuka, O.; Ferreira, M.; Rubner, M. F.; Hsieh, B. R. J. Appl. Phys. 1996, 79, 7501-9.

Figure5. CyclicvoltammogramofLbLbuildupfor(SiMo12|PDDA)10, sample 11. (Inset) Second anodic peak current versus number of bilayers deposited.

V. On the reduction side, the second and third reduction waves seem to merge yet are still distinguishable. For SiMo12, the oxidation waves are 0.02, 0.18, and 0.32 V. And it is the first and second waves on the reduction side that appear to merge. Overall, the first and second redox systems resemble 2-electron courses, as observed in many similar studies. The third redox couple for both PMo12|PDDA and SiMo12|PDDA films seems to depend on individual sample. This third redox couple falls into two categories, either a clear 2-electron course (Figure 4) or more than 2-electron course (Figure 5). All samples illustrate the LbL growth of POM|PDDA bilayers with the insets showing the monotonic increase of number of POM layers versus POM coverage. Surface coverage from cyclic voltammetry is calculated according to the equation7b,8a

Γ)

4ipRT n2F2VA

ip is the second anodic peak current (amperes), n is the number of electrons transferred (2 in this case), V is the scan rate (V‚s-1), A is the geometric area of the electrode (cm2), the gas constant R ) 8.314 J‚K-1‚mol-1, the temperature is 298 K, and Faraday’s constant is F ) 96485 C‚mol-1. An example is demonstrated with sample 1, (PMo12|PDDA)10 made from 5 mM PMo12 solution, whereas ip ) 18.6 × 10-6 A, Γ ) 0.96 × 10-10 (mol‚cm-2). Like the UV-vis calculation, we obtain monolayer equivalent coverage for POM deposition, and the results are presented in Table 3. To analyze the redox behavior of the POM|PDDA films, CV curves were obtained when the scan rate was varied from 10 to 100 mV‚s-1. Two representative curves are shown in Figures 6 and 7, while the remaining ones are presented in the Supporting Information. In Figure 6 the plot of the second anodic peak current versus the scan rate appears as a linear line, while in

11124 Langmuir, Vol. 23, No. 22, 2007

Wang et al.

Table 3. Second Anodic Charge (×10-6 C‚cm-2), Molar Surface Coverage (×10-10 mol‚cm-2), and Monolayer Equivalent Values for (POM|PDDA)10 Films: 1-6, PMO12; 7-12, SiMo12 sample

charge

molar coverage

equivalent coverage

1 2 3 4 5 6 7 8 9 10 11 12

18.62 15.47 8.17 50.83 35.34 13.91 20.05 18.90 13.15 54.51 44.31 27.79

0.96 0.84 0.42 2.63 1.80 0.72 1.03 0.97 0.68 2.82 2.29 1.44

0.8 0.7 0.4 2.1 1.5 0.6 0.9 0.8 0.6 2.4 1.9 1.2

Figure 6. Influence of scan rate on redox curves of (PMo12|PDDA)10, sample 6. (Inset) Second anodic current versus scan rate.

Figure 7 linearity is obtained when the peak current is plotted against the square root of the scan rate.

Discussion LbL Film Structural Conformation Interpreted from UVvis Absorption Values. One important result obtained from this study is to establish a set of fabrication parameters for successful LbL assembly of POM films. We are able to determine two important parameters in such a process, POM concentration and solution ionic strength. We find that POM can be adsorbed onto LbL assemblies from a concentration range from 0.1 to 5 mM. We even tried a POM concentration of 0.01 mM. At this very low concentration roughly half-monolayer adsorption was observed (data not shown) for both low and high solution ionic strengths. The other important parameter is the solution ionic strength. We use two situations, without added salt and with added salt (0.1 M). We think the ionic strength has a more profound effect on the microstructures of the formed films. When the deposition was performed without added salt, the ionic strength of a solution was at a few millimolar scale originated from the counterions and HCl added to adjust the pH. Under this condition, POM clusters adsorb at submonolayer coverage. This observation is in good agreement with some recent studies7,9 that claim an about-monolayer POM deposition within LbL structures. The type of clusters has a limited effect within experimental error. SiMo124- clusters exhibited ∼15% more coverage in two occasions than did PMo123-, while in the third case the two ions showed almost identical coverage. Trivalent and tetravalent particles both can displace monovalent chloride counterions and thus can be used in the layer-by-layer deposition. The cluster concentration shows a limited effect on the deposition process. When the cluster concentration was increased from 0.1 to 5 mM, a 50-fold increase, the coverage merely increased by ∼15-30% for both anionic species. High component concentration generally enhances the adsorption kinetics. A complete kinetics study will be carried out in the future. The quartz slides are coated with poly(ethylene imine) (PEI), a widely used basic, branched polyelectrolyte to coat freshly treated silica-type surfaces. For the first layer we can consider POM anionic clusters as stickers when they deposit on the PEIcoated surface. The electrostatic force is strong enough to attach POM clusters onto the polyelectrolyte layer, replacing some chloride anions balancing PEI charges. This process is kinetic controlled, and higher POM concentration would produce more deposition. Under the experimental conditions, the lowest POM concentration is at least close to the threshold concentration level for saturated adsorption. Now the surface is an about-monolayer of POM clusters embedded in PEI. Followed by deposition with PDDA, the surface charge reverses from overcompensation by

Figure 7. Influence of scan rate on redox curves of (SiMo12|PDDA)10, sample 8. (Inset) Second anodic current versus square root of the scan rate.

PDDA. Here, the PDDA layer conformation is determined by the deposition solution. If the solution ionic strength is low, then PDDA will adapt a rather flat, coiled train conformation. The number of loops (and tails) extruding into the adjacent solution phase is not extensive. When this surface is exposed to POM solution, POM clusters come in contact with and attach onto the PDDA surface. Repeating this process produces a somewhat stratified structure (Figure 8a). When NaCl is added to the deposition solutions, the situation is more complicated. For depositions from a low POM concentration, the adsorption is still at the about-monolayer level. However, when POM concentration is above 1 mM, multilayer structures start to form. The process can be explained by a sequence of procedures. First, high ionic strength screens PDDA chains strongly so that the polymers adapt to an extended configuration in solution with substantial loops (and tails) dangling into the solution phase. After adsorption onto the surface, PDDA chains still remain in the loose configuration. This loose configuration provides room for absorbing extra clusters. Yet, the extra free volume is not easily accessed by POM clusters. More particles are available from high POM concentrations to enter the extra free volume. Thus, higher POM adsorption is observed. The structure formed from a high ionic strength solution is depicted in Figure 8b. The difference between a high POM coverage and a low POM coverage is reflected in the density of the clusters dispersed in the polymer matrix. Multilayer POM deposition has been reported for molybdenum oxide clusters (Mo8O264-),6a which are close to POM clusters in our study. In the case of deposition from low POM concentration solutions (samples 6 and 12), the about-monolayer adsorption values similar to those of samples 1-3 and 7-9 are considered a coincidence. For LbL films built from polyelectrolytes with high ionic strength, exponential layer thickness growth has been observed.19

Assembly of Keggin-type Polyoxometalates

Figure 8. Proposed POM|PDDA multilayer microstructures under different preparation conditions. (a) Under low ionic strength, polymer chains are compact and flat and POM clusters adapt a stratified structure. (b) Under high ionic strength, polymer chains dangle around and POM clusters disperse relatively loosely in the polymer matrix. Legends: gray slab, substrate; brown chains, PEI; polyhedrons, POM; blue chains, PDDA.

The phenomenon is explained by a so-called “in-and-out” model. In a film constituted with cationic and anionic polyelectrolytes, polymer chains exist both tightly complexed to the oppositely charged polyelectrolytes and loosely dispersed in the matrix. This is similar to Schlenoff’s term of intrinsic and extrinsic charge balance. The extrinsic polymer chains can diffuse in and out of the film. During rinsing, some extrinsic polymer chains diffuse to the film/solution interface and become intrinsically complexed by the incoming, oppositely charged polyelectrolytes. Now this polymer complex becomes the new top layer of the LbL film. The amount of extrinsic polymer chains is proportional to the film thickness, thus making the buildup growth exponential. In our study, lack of exponential growth indicates two aspects of the nature of the POM|PDDA film. One aspect is that POM clusters are not capable of constructing a network that permits diffusion of PDDA chains. The other point is that POM clusters themselves cannot diffuse inside the PDDA network either. The latter aspect is in agreement with an early study that has entrapped POMs into polyaniline matrix.20a The entrapment was conducted by electrochemically cycling a polyaniline-coated electrode in POM dissolved in sulfuric acid solution. During the cycling the authors found that POMs effectively replaced sulfate ions and formed films stable in dilute sulfuric acid for a long period. In a subsequent study,20b in which poly(4-vinyl pyridine) was used as the polymer matrix, the size of anions is considered the (19) (a) Lavalle, Ph.; Gergely, C.; Cuisinier, F. J. G.; Decher, G.; Schaaf, P.; Voegel, J. C.; Picart, C. Macromolecules 2002, 35, 4458-65. (b) Richert, L.; Lavalle, Ph.; Payan, E.; Shu, X. Z.; Prestwich, G. D.; Stoltz, J. F.; Schaaf, P.; Voegel, J. C.; Picart, C. Langmuir 2004, 20, 448-58. (c) Lavalle, Ph.; Vivet, V.; Jessel, N.; Decher, G.; Voegel, J. C.; Mesini, Ph. J.; Schaaf, P. Macromolecules 2004, 37, 1159-62. (20) (a) Keita, B.; Bouaziz, D.; Nadjo, L. J. Electroanal. Chem. 1988, 255, 303-13. (b) Keita, B.; Essaadi, K.; Nadjo, L. J. Electroanal. Chem. 1989, 259, 127-46.

Langmuir, Vol. 23, No. 22, 2007 11125

determining factor for such stability. Smaller Fe(CN)44- ions cannot be entrapped in the polymer film firmly. LbL Film Structural Interpretation According to Cyclic Voltammetry. By plotting the current against the bilayers deposited, formation of POM films apparently follows a linear growth layer-by-layer pattern. Many POM films built via the LbL technique have shown this type of linear growth when plotting the current increase at the redox potentials.5a,9a In our study, there is a general trend of redox waves moving to higher potentials, which is probably due to uncompensated resistance. The cyclic voltammograms fall into two general categories: those show typical three 2-electron redox courses and those have a high current on the third reduction wave. The behavior may be related to the microstructures of the films. For PMo12 and SiMo12 clusters in the LbL films constructed from low ionic strength solutions, CV curves demonstrate typical three 2-electron redox courses within the scan potential range. By referring to Figure 8a, this observation is conveniently explained by redox centers (POM clusters) dispersed in a uniform, predominantly 2-dimensional thin film structure. For films obtained from deposition solutions with added salt, CV curves show increased current for the third redox pair, except sample 5. Capacitance currents have been recorded in thin films of POM clusters complexed with multivalent cations.21 In those films POM clusters in the outer layer do not have access to the electrode surface. As a result, charge injected during reduction is trapped inside the matrix. In the current study, samples made from higher solution ionic strength adopt the more porous structure depicted in Figure 8b. The porous structure supplies extra free volume for capacitance charge buildup, which may be in the form of electrolyte ions. Thus, we observe higher current values for the third redox pairs for those samples. In contrast, more uniform structure assembled from low solution ionic strength does not leave enough free volume for charge buildup. There is no apparent explanation for sample 5 built from salt-added solution that does not show the increased thirdpair current (cf. Supporting Information). The charge transport mechanism and kinetics of thin films have been interpreted via the study of varied scan rate. Generally speaking, two types of film charge transport mechanisms have been proposed.22 If the current versus the scan rate plot is linear (cf. Figure 6), then the charge transport involves the film itself, a so-called surface-confined model. If linearity is obtained from current versus the square root of the scan rate (cf. Figure 7), then the charge transport is termed diffusion-limited. We observe both types of charge transport through the films. However, we want to point out another mechanism involved in interpreting the charge transport phenomena. LbL assembly can be considered the progressive reduction of the active area of an electrode. The electrode is unevenly covered by the bilayers. The behavior is called the capillary membrane model.23 Thereafter, the existence of pinholes alters the charge transport behavior observed by experiments and renders the varied scan rate experimentally less interpretable. This is a tradeoff feature of ultrathin films formed via the LbL assembly. Experimental Section Materials and Apparatus. Polyoxometalates H4SiMo12O40‚nH2O and H4PMo12O40‚nH2O (Sigma-Aldrich) were of reagent grade and used without further purification. Poly(ethylene imine) (PEI; Mw (21) Ingersoll, D.; Kulesza, P. J.; Faulkner, L. R. J. Electrochem. Soc. 1994, 141, 140-7. (22) Crespilho, F. N.; Zucolotto, V.; Brett, C. M. A.; Oliveira, O. N., Jr.; Nart, F. C. J. Phys. Chem. B 2006, 110, 17478-83. (23) Barreira, S. V. P.; Garcı´a-Morales, V.; Pereira, C. M.; Manzanares, J. A.; Silva, F. J. Phys. Chem. B 2004, 108, 17973-82.

11126 Langmuir, Vol. 23, No. 22, 2007 60 000) and poly(diallyl dimethyl ammonium chloride) (PDDA; Mw 250 000) were purchased from Aldrich, but their aqueous solutions were centrifuged for 20 min before use. The water used for all experiments was purified by a Barnsted Nanopure II purification system. The resistivity was about 18 MΩ/cm. The solutions used to prepare multilayer assemblies were PEI solution (10 mM, pH 8) and PDDA solution (10 mM). Other solution parameters are explained in Table 1. UV-vis absorption spectra were recorded using a Varian Cary 50 Bio UV-Visible Spectrophotometer equipped with a neon lamp. Electrochemical measurements were conducted on a Voltalab 10 PGZ 100 instrument. A conventional three-electrode setup was used. The working electrode was an indium tin oxide (ITO) coated glass substrate modified with self-assembled films, while a platinum wire was used as the counter electrode. An Ag/AgCl (3 M KCl) reference electrode was used for all measurements. Substrates. The POM|PDDA multilayer films were deposited onto the following substrates: quartz slides (SPI supplies 25 × 76 × 1 mm) for UV-vis and ITO slides (Sigma-Aldrich, 80-100 Ω) for cyclic voltammetry. The quartz slides were cleaned by soaking them overnight in a base bath (10% KOH in isopropyl alcohol) followed by immersion in a solution containing 3 parts H2SO4 (96 wt % aqueous solution) and 1 part H2O2 (30 wt % aqueous solution) at 80 °C for 1 h and finally rinsing with copious amounts of water. The ITO slides were cleaned by soaking them overnight in a similar base bath followed by rinsing with copious amounts of water and then sonication in water for 10 min. Film Preparation. First, a PEI precursor film was deposited onto a clean substrate by dipping the substrate in PEI solution for 20 min followed by consecutive dipping in three nanopure water baths for 3 min each. The PEI-coated substrate was then dipped in POM (5 min) and PDDA solution (5 min) in an alternating fashion for 10 consecutive times; between each dipping the substrate was exposed to three consecutive nanopure water baths.

Wang et al.

Conclusion We first use scaling to gain insight into the interaction between polyoxometalates and polyelectrolytes during the layer-by-layer process. According to the estimation, PDDA chains may wrap around POM clusters with a finite length. We then establish a set of parameters for streamlined preparation of ultrathin LbL films. LbL deposition is confirmed via UV-vis absorption and cyclic voltammetry measurements. The microstructures of the constructed LbL films can be interpreted from the experimental results. The most important parameter in such self-assembly is the solution ionic strength. At low ionic strength, polyoxometalate clusters tend to form a monolayer between two polyelectrolyte layers. The difference in cluster concentration does not affect the deposition significantly. At high ionic strength, polyelectrolyte chains extend vertically, forming a 3-dimentional structure that accommodates incoming clusters, and multilayer formation becomes likely. Under this condition, the POM cluster concentration plays a more significant role in forming the final film configuration. The multilayer, 3-D structures demonstrate electrochemical capacitance charge buildup, which is absent from the monolayer structures. Acknowledgment. We thank Lamar University for financial support. Supporting Information Available: Figures of UV-vis absorption, cyclic voltammetry, and variable scan rate voltammetry of (POM|PDDA) films. This material is available free of charge via the Internet at http://pubs.acs.org. LA701789N