Simple Colorimetric Method for Quantification of Surface Carboxy

ARTICLE pubs.acs.org/ac

Simple Colorimetric Method for Quantification of Surface Carboxy Groups on Polymer Particles Andreas Hennig,*,† Angelika Hoffmann,† Heike Borcherding,‡ Thomas Thiele,‡ Uwe Schedler,‡ and Ute Resch-Genger*,† † ‡

BAM
Federal Institute for Materials Research and Testing, Richard-Willst€atter-Strasse 11, D-12489 Berlin, Germany PolyAn GmbH, Rudolf-Baschant-Strasse 2, D-13086 Berlin, Germany ABSTRACT: We present a novel, simple, and fast colorimetric method to quantify the total number of carboxy groups on polymer microparticle and nanoparticle surfaces. This method exploits that small divalent transition metal cations (M2þ = Ni2þ, Co2þ, Cd2þ) are efficiently bound to these surface functional groups, which allows their extraction by a single centrifugation step. Remaining M2þ in the supernatant is subsequently quantified spectrophotometrically after addition of the metal ion indicator pyrocatechol violet, for which Ni2þ was identified to be the most suitable transition metal cation. We demonstrate that the difference between added and detected M2þ is nicely correlated to the number of surface carboxy groups as determined by conductometry, thereby affording a validated measure for the trueness of this procedure. The variation coefficient of ∼5% found in reproducibility studies underlines the potential of this novel method that can find conceivable applications for the characterization of different types of poly(carboxylic acid)-functionalized materials, e.g., for quality control by manufacturers of such materials.

M

icrometer to nanometer-sized polymer particles are of great interest for a wide range of applications in materials and life sciences as well as for medical applications.1 Commonly, the surface of polymer particles is covered by an outer layer of functional groups, which confers colloidal stability to the particle and allows conjugation of peptides, proteins, antibodies, or DNA to the particle surface. The precise characterization of the nature and density of functional groups on polymer particle surfaces is therefore highly important. Standard techniques to quantify the number of surface functional groups include, for example, Fourier transform-infrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS), potentiometry, and conductometry. More recently, solid-state NMR and isothermal titration calorimetry have been suggested as well.2,3 Attractive alternatives are straightforward and sensitive colorimetric or fluorometric adsorption/desorption-based dye assays.2c,4,5 Therein, a dye of complementary charge is first adsorbed onto the particle surface. After extensive washing to remove unspecifically adsorbed dye, the dye is desorbed with a detergent and, finally, the amount of desorbed dye is quantified by absorption or fluorescence spectroscopy. This elegantly bypasses optical measurements with potentially scattering, absorbing, or emitting particles. A prominent example is the use of positively charged toluidine blue (TB) to quantify carboxy groups, although other cationic dyes have been used as well.2c,5 However, adsorption/desorption assays suffer from two shortcomings. First, they require tedious washing steps including repeated cycles of buffer addition, particle resuspension, centrifugation, and removal of washing solution, and second, it is arguable whether they quantitatively detect the number of surface functional groups. Deviations r 2011 American Chemical Society

from the expected value, which was, for example, obtained by potentiometry, have been rationalized by unspecific surface adsorption and the relatively large size of common dyes.2c,4 These limitations of conventional adsorption/desorptionbased dye assays encouraged us to replace the sterically demanding and bulky ionic dyes with small metal cations. We presumed a highly efficient binding of divalent and trivalent metal cations toward negatively charged polycarboxylate surfaces owing to the known tendency of polymers to minimize charge
charge repulsion by counterion binding, in particular at interfaces.6 Favorable multivalent interactions between metal cations and the high local concentration of carboxy groups on the surface may also contribute to efficient binding (similar to a chelate effect).7 Furthermore, the formation of highly ordered and stable metal
organic frameworks from carboxy-terminated surfaces has been recently demonstrated using layer-by-layer deposition of Cu2þ or Zn2þ.8 As a first proof-of-concept for our envisaged strategy, we demonstrate herein that addition of micrometer-sized and nanometersized particles covered with a poly(carboxylic acid) surface leads to an efficient extraction of different divalent transition metal cations (Cd2þ, Co2þ, Ni2þ) from solution. Excess metal cations in the supernatant are then quantified spectrophotometrically after a single centrifugation step by using a metal ion indicator that forms a strongly colored complex with these cations. The potential of this novel method for quantification of surface carboxy groups was evaluated by comparison with conductometry. Received: March 26, 2011 Accepted: May 10, 2011 Published: May 11, 2011 4970

dx.doi.org/10.1021/ac2007619 | Anal. Chem. 2011, 83, 4970–4974

Analytical Chemistry

ARTICLE

’ EXPERIMENTAL SECTION Materials. Buffers and transition metal perchlorate salts were from Alfa Aesar or Aldrich and used as received. Chemicals for conductometric titrations were from Merck or Roth or Aldrich. Poly(methyl methacrylate) (PMMA) microparticles with a covalently linked layer of varying amounts of poly(acrylic acid) (PAA) were individually prepared from PolyAn GmbH (Berlin, Germany) (diameter = 5, 6, and 11 μm) for this study or commercially available from Bangs Laboratories Inc. (diameter = 0.1 μm). Pyrocatechol violet (PV) was from Merck. Instrumentation. For absorption measurements, either a Varian Cary 5000 equipped with a temperature controller or a Bruins Instruments OMEGA 10 were used. Optical microscopy was carried out with an Olympus BX51 microscope equipped with a XC30 digital color camera. Conductivity measurements were performed with a WTW Cond 3310 conductivity meter equipped with a WTW TetraCon 325 temperature-corrected electrode. Procedures. In a standard procedure, varying amounts of the PMMA microparticle stock solution were incubated with 200 μM Ni2þ for ∼2 min in 10 mM Hepes, pH 7.5 (total volume 600 μL). After centrifugation for 1 min at 16 000 rcf (diameter = 5, 6, and 11 μm) or 30 min at 16 000 rcf (diameter = 0.1 μm), 500 μL of the supernatant was diluted to 1000 μL with 10 mM Hepes, pH 7.5 and a PV stock solution (freshly prepared on a daily basis) in 10 mM Hepes, pH 7.5 was added. The final concentration of PV was 40 μM. Absorption spectra were always recorded immediately after PV addition and mixing. The absorbance at 650 nm was plotted against the particle stock solution volume. Linear fitting of the initial linear decrease of this titration plot (see for example Figure 2b) gave the slope of the fitted line a, the y-intercept b, and the number of surface carboxy groups obtained by eq 1. This procedure was typically performed in triplicate to afford the amount of extracted Ni2þ and the respective standard deviation (∼5%). Conductometric titrations were performed as equilibrium titrations according to literature methods.9 In brief, varying amounts of microparticle stock solution were diluted with 15 mL of 0.3 mM KBr. KOH solution was used as the titrant and freshly prepared from standard concentrates of known titer. All data were acquired using base-into-acid titration with a volume increment of 5 μL of 0.02 to 0.1 M KOH/injection. Equilibrium state was reached by waiting for a stable conductivity value, and quantitative information was obtained by extrapolation to determine the titration end points.

’ RESULTS AND DISCUSSION To elaborate our strategy, we have selected pyrocatechol violet (PV), which is a classic indicator for transition metal cations and has more recently been used in indicator displacement assays.10 To identify a suitable metal ion, we first investigated the response of PV toward a broad range of metal cations (Naþ, Mg2þ, Ca2þ, Fe3þ, Co2þ, Ni2þ, Cu2þ, Zn2þ, Cd2þ, and Hg2þ) in 10 mM Hepes, pH 7.5. Under these conditions, addition of 10 mM Fe3þ, Cu2þ, and Zn2þ immediately gave a precipitate, presumably the corresponding hydroxide, while addition of Hg2þ gave a metallic gray precipitate within few minutes indicative of a redox reaction. Among the other cations, Co2þ, Ni2þ, and Cd2þ caused significant changes in the absorption spectrum of a 25 μM PV solution (Figure 1a
c). The presence of these transition metal ions was indicated by a bathochromic shift and an increase of the lowest energy absorption band, which was also apparent to the naked eye by a visible color change from green to blue. The appearance of

Figure 1. Transition metal cation titration (Ni2þ, Cd2þ, Co2þ) in the presence of 25 μM PV in 10 mM Hepes, pH 7.5. (a
c) Changes in the absorption spectrum upon addition of (a) NiCl2 (0
200 μM), (b) CdCl2 (0
200 μM), (c) CoCl2 (0
200 μM). (d) Concentration dependence of the absorbance A at 650 nm. Error bars indicate the standard deviation (n = 3). Solid lines represent a linear fit of the titration data (R > 0.99).

isosbestic points at ∼490 and 390 nm suggests the formation of a 1:1 complex between the metal ions and PV. Addition of common buffer additives such as up to 100 mM Naþ, Mg2þ, and Ca2þ did not result in changes in the PV absorption spectrum. Titration plots of the absorption at 650 nm versus Co2þ, Ni2þ, and Cd2þ concentration were nearly linear up to a concentration of 100 μM (R > 0.99, Figure 1b). Beyond that concentration, the titration plots were sufficiently curved to be fitted according to a 1:1 binding isotherm, which afforded association constants Ka of ∼103 to 104 M
1 (Table 1). Extrapolation of the fit gave the maximal change in absorbance ΔAmax at 650 nm, which was quite similar for all three cations. The change in absorbance at maximum cation concentration, where the plots were still linear, i.e., ΔA100 μM, was most pronounced for Ni2þ. Thus, the difference in ΔA100 μM among the three cations studied can be traced back to their different association constants. During our experiments, we observed a slow but continuous change in the absorption spectrum of the PV
M2þ complex, i.e., a decrease in absorbance at 650 nm with time. The relative decrease in absorbance is given in Table 1 as decomposition rate vdec. This change was negligible for Cd2þ, but care should be taken to measure the absorption spectra of PV solutions containing Ni2þ and Co2þ at best within 5 min after mixing of the reagents. These results prompted us to evaluate our novel surface group quantification strategy with Ni2þ. As micro- and nanoparticles, we have chosen uniform, nonporous particles consisting of 4971

dx.doi.org/10.1021/ac2007619 |Anal. Chem. 2011, 83, 4970–4974

Analytical Chemistry

ARTICLE

Table 1. Characterization of Pyrocatechol Violet
Metal Ion Complexationa ΔAmax

ΔA100 μM

vdec

(650 nm)b

(650 nm)c

(min
1)d

Ni2þ 8400 ( 900 0.66 ( 0.03 Co2þ 5500 ( 500 0.49 ( 0.06

0.28 ( 0.01 0.16 ( 0.02

(2.09 ( 0.02)  10
2 (1.89 ( 0.01)  10
2

Mnþ

Ka (M
1)

Cd2þ 1600 ( 200

0.56 ( 0.06 0.071 ( 0.005 (1.6 ( 0.4)  10
3

Values and standard deviation (n = 3) are given for 25 μM PV, pH 7.5. Maximal change in absorbance at saturation. c Change in absorbance at 100 μM Mnþ. d Decomposition rate given as change in normalized absorbance
d(A/A0)/dt (dimensionless) per minute as determined by linear fitting after 10 min. a b

poly(methyl methacrylate) (PMMA) with a covalently linked outer layer of poly(acrylic acid) (PAA). Overall, 18 different particle batches varying in diameter and in the number of carboxy groups were investigated. The particles were selected to cover a mean diameter range between 0.1 and 11 μm (as given by the supplier or as determined by optical microscopy) and a PAA range of 10
7000 μmol/g (as determined by conductometry). To underline the broad applicability of our novel approach, some of the microparticles were also encoded with fluorescent dyes encapsulated in the microparticle core such as employed for common microparticle-based assay platforms. In a typical assay, varying amounts of the PMMA microparticle stock solution were mixed with 200 μM Ni2þ and centrifuged to afford the supernatant. After 2-fold dilution and addition of PV, absorption spectra were recorded. These measurements revealed the disappearance of the absorption band at 650 nm with increasing amounts of microparticle stock solution during incubation (Figure 2a), indicative of the decrease of supernatant concentration of Ni2þ. The absorption spectrum at highest microparticle concentrations (>100 μL in Figure 2) coincided with the absorption spectrum of PV in the absence of Ni2þ, which clearly demonstrates that Ni2þ was extracted to such an extent that its concentration was below the detection limit of the metal ion indicator. A plot of the absorbance at 650 nm against the volume of microparticle stock solution gave a linear decrease until a plateau was reached (Figure 2b) suggesting a linear relation between the amount of microparticles and remaining Ni2þ in solution. Control experiments with unfunctionalized PMMA particles gave no decrease in absorbance at 650 nm, which excludes the possibility of unspecific adsorption between metal ions and the PMMA core. Longer incubation times of particles and Ni2þ had no influence on the outcome of the titration, suggesting that ion exchange on the particle surface is instantaneous. The amount of microparticle stock solution to fully extract Ni2þ from the solution is given by the intersection between the initial linear decrease in the titration plot and the absorbance of PV in the absence of Ni2þ (see Figure 2b). Presuming that this value is correlated to the number of surface carboxy groups, one can derive the following equation surface carboxy groups ðμmol=gÞ ¼

n½M2þ Va wðAPV
bÞ

ð1Þ

where V is the volume and [M2þ] the metal ion concentration during particle/M2þ incubation, a the slope, b the y-intercept of the initial linear decrease, APV the absorbance of PV in the absence of M2þ, w the mass concentration (in milligrams/milliliter) of the

Figure 2. Determination of carboxy surface concentrations of PMMA/ PAA polymer core-shell microparticles. (a) 0
120 μL of a microparticle suspension (diameter = 6 μm, 240 μmol/g surface PAA by conductometry, 37.7 mg/mL) were incubated in 10 mM Hepes, pH 7.5, 200 μM Ni2þ. Absorption spectra were taken after centrifugation, 2-fold dilution, and addition of 40 μM PV. (b) Plot of the absorbance at 650 nm versus the volume of microparticle suspension during incubation. Single titrations were performed and evaluated individually to afford the amount of extracted Ni2þ per amount of beads (cf. Experimental Section for details).

particle stock solution, and n a stoichiometry factor indicating the number of surface carboxy groups per metal cation. To ascertain the stoichiometry factor n, we plotted the amount of surface carboxy groups determined by conductometry against the amount of Ni2þ extracted by the microparticles as determined with PV (Figure 3). Linear regression revealed a very good correlation (R > 0.98), which underlines the principle suitability of our assay. The stoichiometry factor obtained by this linear regression was 2.65 ( 0.12. This clear deviation from an integer indicates that no well-defined transition metal complexes are formed with PAA on the microparticle surface. To further verify this conjecture, the influence of different transition metal cations on the surface binding stoichiometry was tested by repeating titrations with Co2þ and Cd2þ. The identical results obtained for all three cations suggest that the interaction between the transition metal cations and the microparticle surface is purely electrostatic. Selected surface carboxy group determinations were also repeated at pH 8.0, which gave identical results and suggests a complete deprotonation of surface carboxy groups at pH 7.5. The precision of the assay was determined on the basis of six replicates with two arbitrarily selected microparticle preparations. The variation coefficients of these two series of measurements were 2.8% and 4.3%, which demonstrates the excellent reproducibility of our method. The trueness can be judged by inspection of Figure 3. The closer data points are to the fitted line, the better both methods are correlated. For example, the data point around 250 μmol/g extracted Ni2þ translates into a surface carboxy concentration of 650 μmol/g considering a stoichiometry factor of n = 2.65. While this is only in moderate agreement with the conductometry value of 1000 μmol/g, other values are much better correlated, for example, 479 and 506 μmol/g, 102 and 86 μmol/g, or 3253 and 3210 μmol/g as determined by conductometry and by our method, respectively. Compared with other surface adsorption-based colorimetric and fluorometric assays, the method disclosed herein is faster and more reliable. Excessive repeated washing cycles are not necessary, and the supernatant can be directly analyzed after a single centrifugation step. Such a direct measurement of the 4972

dx.doi.org/10.1021/ac2007619 |Anal. Chem. 2011, 83, 4970–4974

Analytical Chemistry

ARTICLE

transferability of binding stoichiometries also applies to the frequent but sometimes indiscriminate use of dye-based assays.11

Figure 3. Correlation of surface carboxy groups of a variety of PMMA micro- and nanoparticles (diameter = 0.1 to 11 μm, 10
1200 μmol/g surface PAA, some particles were encoded with encapsulated fluorescent dyes) as determined by conductometry (n = 3) versus the amount of Ni2þ extracted by microparticles as determined by the Ni2þ/PV assay (n g 3). Error bars indicate the standard deviation. The inset shows the whole range of particles investigated (10
7000 μmol/g surface PAA). Linear regressions (broken lines), which were performed individually for both presented data ranges, revealed a very good correlation (R > 0.98) and suggested that 2.65 ( 0.12 carboxy groups bind per Ni2þ.

supernatant is not possible with other adsorption/desorption-based dye assays, which require a large excess of dye during the adsorption step, such that the difference in dye concentration before and after adsorption is commonly too small to be reliably quantified.2c Regarding the assay reproducibility, for example, the common TB assay is much less precise with a variation coefficient of ∼20%.5 Furthermore, the binding stoichiometry of the transition metal-based PV assay appears to be more robust than the binding stoichiometry reported for dye-based adsorption/desorption assays. For example, quantification of PAA on poly(vinylidene fluoride) films by a TB assay and comparison with results from potentiometric titrations suggested that only the uppermost layer of PAA was detected by the TB assay, presumably for steric reasons.4 Such a size effect is likely to be much smaller or even absent for the use of small M2þ as surface binding species. Interestingly, the same authors briefly reported that adsorption of Cu2þ to the polymer surface, subsequent desorption with EDTA, and quantification of the Cu
EDTA complex gave results, which were in good agreement with the potentiometric titration, yet this strategy was not further pursued.4 Obviously, the absence of well-defined complexes is not necessarily detrimental for an analytical application. Our results cover a surface concentration of nearly 3 orders of magnitude with an excellent correlation. Accordingly, this seems to be one of the best calibrated colorimetric detection methods for PAA on surfaces. Most likely, this robust procedure can also be transferred to chemically similar poly(carboxylic acids) on surfaces. Surprisingly, stoichiometries reported for binding of Cd2þ to poly(N-isopropyl acrylamide-acrylic acid-2-hydroxyethyl acrylate) microgel particles were in excellent agreement with our findings (∼2.6 ( 0.5 versus 2.65 ( 0.12).1g Clearly, it remains to be clarified to which extent this agreement is coincidental or whether the binding stoichiometry is generally transferable to polymer particle surfaces with alternating, periodical, and statistical copolymers or even to surface carboxy-modified inorganic particles such as silica particles or quantum dots, and work into that direction is underway. Noteworthy, the issue of general

’ CONCLUSIONS With our metal ion-based assay, we have introduced a fast, simple, and robust colorimetric method for the quantification of surface carboxy groups. In contrast to other more time-consuming adsorption/desorption-based dye assays, this method requires just a single centrifugation step. An excellent linear correlation with results from conductometry over nearly 3 orders of magnitude of surface concentrations underlines its wide applicability and the low variation coefficient of ∼5% demonstrates its excellent reproducibility. Overall, we believe that such a simple and reliable method is highly useful, e.g., for particle characterization and quality control during particle production and functionalization. This may also pave the way for the future development of high-throughput methods in materials synthesis and analysis. ’ AUTHOR INFORMATION Corresponding Author

*Fax, þ49 30 8104 5005; phone, þ49 30 8104 5962; e-mail, [email protected] (A.H.). Fax, þ49 30 8104 1157; phone, þ49 30 8104 1134; e-mail, [email protected] (U.R.-G.).

’ ACKNOWLEDGMENT Financial support from the Federal Ministry of Economics and Technology (Grant Number BMWI VI A2-17/03) is gratefully acknowledged. ’ REFERENCES (1) (a) Baumes, J. M.; Gassensmith, J. J.; Giblin, J.; Lee, J.-J.; White, A. G.; Culligan, W. J.; Leevy, W. M.; Kuno, M.; Smith, B. D. Nat. Chem. 2010, 2, 1025–1030. (b) Abdelrahman, A. I.; Dai, S.; Thickett, S. C.; Ornatsky, O.; Bandura, D.; Baranov, V.; Winnik, M. A. J. Am. Chem. Soc. 2009, 131, 15276–15283. (c) Barner, L. Adv. Mater. 2009, 21, 2547–2553. (d) de Silva, A. P.; James, M. R.; McKinney, B. O. F.; Pears, D. A.; Weir, S. M. Nat. Mater. 2006, 5, 787–790. (e) Lu, Y.; Mei, Y.; Drechsler, M.; Ballauf, M. Angew. Chem., Int. Ed. 2006, 45, 813–816. (f) Jin, H.-J.; Choi, H. J.; Yoon, S. H.; Myung, S. J.; Shim, S. E. Chem. Mater. 2005, 17, 4034–4037. (g) Zhang, J.; Xu, S.; Kumacheva, E. J. Am. Chem. Soc. 2004, 126, 7908–7914. (h) Han, M.; Gao, X.; Su, J. Z.; Nie, S. Nat. Biotechnol. 2001, 19, 631–635. (2) (a) Gaborieau, M.; Nebhani, L.; Graf, R.; Barner, L.; BarnerKowollik, C. Macromolecules 2010, 43, 3868–3875. (b) Goldmann, A. S.; Walther, A.; Nebhani, L.; Joso, R.; Ernst, D.; Loos, K.; Barner-Kowollik, C.; Barner, L.; M€uller, A. H. E. Macromolecules 2009, 42, 3707–3714. (c) Goddard, J. M.; Hotchkiss, J. H. Prog. Polym. Sci. 2007, 32, 698–725. (d) Okubo, M.; Suzuki, T.; Tsuda, N. Colloid Polym. Sci. 2006, 284, 1319–1323. (e) Qi, K.; Ma, Q.; Remsen, E. E.; Clark, C. G., Jr.; Wooley, K. L. J. Am. Chem. Soc. 2004, 126, 6599–6607. (f) Hoare, T.; Pelton, R. Langmuir 2004, 20, 2123–2133. (g) Li, P.; Xu, J.; Wang, Q.; Wu, C. Langmuir 2000, 16, 4141–4147. (3) (a) Panella, B.; Vargas, A.; Ferri, D.; Baiker, A. Chem. Mater. 2009, 21, 4316–4322. (b) Huang, S.; Joso, R.; Fuchs, A.; Barner, L.; Smith, S. V. Chem. Mater. 2008, 20, 5375–5380. (c) Xing, Y.; Dementev, N.; Borguet, E. Curr. Opin. Solid State Mater. Sci. 2007, 11, 86–91. (d) Qi, K.; Ma, Q.; Remsen, E. E.; Clark, C. G., Jr.; Wooley, K. L. J. Am. Chem. Soc. 2004, 126, 6599–6607. (4) Clochard, M. C.; Begue, J.; Lafon, A.; Caldemaison, D.; Bittencourt, C.; Pireaux, J. J.; Betz, N. Polymer 2004, 45, 8683–8694. 4973

dx.doi.org/10.1021/ac2007619 |Anal. Chem. 2011, 83, 4970–4974

Analytical Chemistry

ARTICLE

(5) (a) Uchida, E.; Uyama, Y.; Ikada, Y. Langmuir 1993, 9, 1121–1124. (b) Ivanov, V. B.; Behnisch, J.; Holl€ander, A.; Mehdorn, F.; Zimmermann, H. Surf. Interface Anal. 1996, 24, 257–262. (c) Kang, E. T.; Tan, K. L.; Kato, K.; Uyama, Y.; Ikada, Y. Macromolecules 1996, 29, 6872–6879. (d) R€odiger, S.; Ruhland, M.; Schmidt, C.; Schr€oder, C.; Grossmann, K.; B€ohm, A.; Nitschke, J.; Berger, I.; Schimke, I.; Schierack, P. Anal. Chem. 2011, 83, 3379–3385. (6) (a) Hennig, A.; Gabriel, G. J.; Tew, G. N.; Matile, S. J. Am. Chem. Soc. 2008, 130, 10338–10344. (b) Pantos, A.; Tsogas, I.; Paleos, C. M. Biochim. Biophys. Acta, Biomembr. 2008, 1778, 811–823. (c) Sakai, N.; Matile, S. J. Am. Chem. Soc. 2003, 125, 14348–14356. (d) Ariga, K.; Kunitake, T. Acc. Chem. Res. 1998, 31, 371–378. (7) (a) Major, R. C.; Zhu, X.-Y. J. Am. Chem. Soc. 2003, 125, 8454–8455. (b) Doyle, E. L.; Hunter, C. A.; Phillips, H. C.; Webb, S. J.; Williams, N. H. J. Am. Chem. Soc. 2003, 125, 4593–4599. (8) (a) Shekhah, O.; Wang, H.; Strunskus, T.; Cyganik, P.; Zacher, D.; Fischer, R.; W€oll, C. Langmuir 2007, 23, 7440–7442. (b) Shekhah, O.; Wang, H.; Kowarik, S.; Schreiber, F.; Paulus, M.; Tolan, M.; Sternemann, C.; Evers, F.; Zacher, D.; Fischer, R. A.; W€oll, C. J. Am. Chem. Soc. 2007, 129, 15118–15119. (9) Lorenz, O.; Breidenich, N.; Denter, U.; Reinm€oller, K.-H.; Rose, G. Angew. Makromol. Chem. 1982, 103, 159–185. (10) (a) Butterfiled, S. M.; Hennig, A.; Matile, S. Org. Biomol. Chem. 2009, 7, 1784–1792. (b) Hanshaw, R. G.; O’Neil, E. J.; Foley, M.; Carpenter, R. T.; Smith, B. D. J. Mater. Chem. 2005, 15, 2707–2713. (c) Zhu, L.; Zhong, Z.; Anslyn, E. V. J. Am. Chem. Soc. 2005, 127, 4260–4269. (d) Zhu, L.; Anslyn, E. V. J. Am. Chem. Soc. 2004, 126, 3676–3677. (11) Davydoff, W.; Linow, K.-J.; Philipp, B. Acta Polym. 1991, 42, 646–650.

4974

dx.doi.org/10.1021/ac2007619 |Anal. Chem. 2011, 83, 4970–4974