splittings of the order of 15 cm 1 , which lies perpendicular to the diagonal of the plot and can be traced to the coupling of the amino acids Thr-1 ± Ile34, Thr2 ± Ile33, and Cys3 ± Cys32. Finally, the close head-to-tail contact of the protein is visible as a cluster of strong level splittings in the range of 10 ± 15 cm 1 in the top right- and lower left-hand corner of the figure. We note that the two absorption lines originating from coupling may differ considerably in intensity, depending on the mutual orientation of the dipoles; this is also visible in Figure 2. Whereas potentially observable vibrational couplings in deuterated proteins are reduced to close D ± D van der Waals contacts and are thus rare,[20] the omnipresence of the peptide backbone CO group and its long-range dipole ± dipole interaction gives rise to numerous level splittings for 18O/16O and– slightly smaller in magnitude and not presented here–12C/13C isotope substitution patterns. The coupling strengths depend both on the distance between and on the mutual orientation of the dipoles [cf. Equation (1)], and thus permit the calculation of the upper limit of the interdipole distance. More important, protein secondary structure elements give rise to characteristic coupling patterns, and through-space interactions within a protein induce a large effect. Therefore, systematic isotope substitution provides a valuable tool to obtain structural information about proteins that are difficult to crystallize, for example, membrane proteins.
[8] K. J. Reynolds, X. D. Yao, C. Fenselau, J. Proteome Res. 2002, 1, 27 ± 33. [9] C.-Y. Shiau, M. F. Byford, R. T. Aplin, J. E. Baldwin, C. J. Schofield, Biochemistry 1997, 36, 8798 ± 8806. [10] J. Torres, P. D. Adams, I. T. Arkin, J. Mol. Biol. 2000, 300, 677 ± 685. [11] J. Torres, A. Kukol, J. N. Goodman, I. T. Arkin, Biopolymers 2001, 59, 396 ± 401. [12] A. Kukol, J. Torres, I. T. Arkin, J. Mol. Biol. 2002, 320, 1109 ± 1117. [13] W. H. Moore, S. Krimm, Proc. Nat. Acad. Sci. USA 1975, 72, 4933 ± 4935; S. Krimm, Y. Abe, Proc. Nat. Acad. Sci 1972, 69, 2788 ± 2792. [14] H. Torii, M. Tasumi, J. Chem. Phys. 1992, 96, 3379 ± 3387. [15] W. A. Hendrickson, M. M. Teeter, Nature 1981, 290, 107 ± 113. [16] Brookhaven protein data database (PDB) entry 1CRN. [17] A. M. J. J. Bonvin, J. A. C. Rullmann, R. M. J. N. Lamerichs, R. Boelens, R. Kaptein, Proteins 1993, 15, 385 ± 400. [18] M. Llina¬s, A. De Marco, J. T. J. Lecomte, Biochemistry 1980, 19, 1140 ± 1145. [19] We note that for isotope-substituted spectra, the Teller ± Redlich product rule [Equation (4)] holds, s Y Y (nÄ i* /nÄi)
mi =m i* (4) i
i
which relates the entire set of isotope-shifted wavenumbers to the entire set of masses, whereas no individual wavenumber shifts can be predicted solely on the basis of the isotope substitution pattern (the asterisk refers to the isotope-substituted spectrum, and the product runs over all nontranslational and nonrotational levels). See O. Redlich, Z. Phys. Chem. B 1935, 28, 371 ± 382. [20] J. Becker, F. Becher, A. Labahn, O. Hucke, T. Koslowski, J. Phys. Chem. B 2003, 107, 12878 ± 12883. [21] F. Siebert, W. M‰ntele, Eur. J. Biochem. 1983, 130, 565 ± 573. Received: July 10, 2003 [Z 902] Revised: November 10, 2003
Acknowledgements It is a pleasure to thank P. Gr‰ber, J. Becker, N. Utz, M. Rateitzak, O. Hucke and C. A. Zell for fruitful discussions. Keywords: biophysics ¥ dipole coupling ¥ IR spectroscopy ¥ isotope effects ¥ oxygen
[1] a) J. K. Chin, R. Jimenez, F. E. Romesberg, J. Am. Chem. Soc. 2001, 123, 2426 ± 2427; b) J. K. Chin, R. Jimenez, F. E. Romesberg, J. Am. Chem. Soc. 2002, 124, 1846 ± 1847. [2] a) K. Ohno, S. Nomura, H. Yoshida, H. Matsuura, Spectrochim. Acta A 1999, 55, 2231 ± 2246; b) L. Grajcar, M. H. Baron, S. Becourarn, S. Czernecki, J. M. Valery, C. Reiss, Spectrochim. Acta A 1994, 50, 1015 ± 1022; c) R. A. Dluhy, R. Mendelsohn, H. L. Casal, H. H. Mantsch, Biochemistry 1983, 22, 1170 ± 1177; d) B. Pastrana, A. J. Mautone, R. Mendelsohn, Biochemistry 1991, 30, 10 058 ± 10 064; e) I. Echabe, M. A. Requero, F. M. Goni, J. L. R. Arrondo, A. Alonso, Eur. J. Biochem. 1995, 231, 199 ± 203; f) M. V. Fraile, G. LopezRodriguez, J. Gallego-Nicasio, P. Carmona, Biopolymers 2000, 57, 11 ± 18; g) R. Moritz, H. Fabian, U. Hahn, M. Diem, D. Naumann, J. Am. Chem. Soc. 2002, 124, 6259 ± 6264. [3] E. B. Wilson Jr., J. C. Decius, P. C. Cross, Molecular Vibrations; Dover Publications: New York, 1955. [4] a) J. Torres, I. T. Arkin, Biophys. J. 2002, 82, 1068 ± 1075; b) J. Torres, A. Kukol, I. T. Arkin, Biophys. J. 2000, 79, 3139 ± 3143. [5] a) S. Sonar, C. P. Lee, M. Coleman, N. Patel, X. M. Liu, T. Marti, H. G. Khorana, U. L. RajBhandary, K. J. Rothschild, Nat. Struct. Biol. 1994, 1, 512 ± 517; b) C. J. Noren, S. J. Anthony-Cahill, M. C. Griffith, P. G. Schultz, Science 1989, 244, 182 ± 188; c) D. R. Liu, T. J. Magliery, M. Pasternak, P. G. Schultz, Proc. Natl. Acad. Sci. USA 1997, 94, 10 092 ± 10 097. [6] V. Bergo, S. Mamaev, J. Olejnik, K. J. Rothschild, Biophys. J. 2003, 84, 960 ± 966. [7] X. Yao, A. Freas, J. Ramirez, P. A. Demirev, C. Fenselau, Anal. Chem. 2001, 73, 2836 ± 2842.
ChemPhysChem 2004, 5, 235 ± 239
DOI: 10.1002/cphc.200300930
Supramolecular Architectures of Electrostatic Self-Assembled Glucose Oxidase Enzyme Electrodes Ernesto J. Calvo*[a] and Alejandro Wolosiuk[b] Electrostatic self-assembly constitutes one of the most used and simplest methods for the preparation of supported organized supramolecular systems.[1] The incorporation of enzymes as building blocks in these structures results in a modified surface that has high selectivity toward molecules of biological interest. Applications of these systems range from the design and construction of nanoscale integrated circuits such as biochips and biosensors for genomic treatment to the development of biofuel cells.[2] In fact, wiring redox enzymes with electroactive polymers have integrated systems where biomolecular recog-
[a] Prof. E. J. Calvo INQUIMAE–Departamento de QuÌmica Inorga¬nica AnalÌtica y QuÌmica FÌsica Facultad de Ciencias Exactas y Naturales–Universidad de Buenos Aires Pabello¬n 2, Ciudad Universitaria, 1428 Buenos Aires (Argentina) Fax: ( 54) 11-4576-3341 E-mail:
[email protected] [b] Dr. A. Wolosiuk Department of Material Science and Engineering University of Illinois at Urbana Champaign, IL 61801 (USA)
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
235
nition can be related to an electrical signal in a single sensing device.[3] We report on the ™molecular wiring∫ efficiency and the fraction of ™wired∫ enzyme molecules of different molecular architectures of glucose oxidase (GOx, E.C. 1.1.3.4) in organized self-assembled nanostructures comprised of anionic enzyme layers alternating with layers of an osmium derivatized poly(allylamine) cationic polyelectrolyte (PAH-Os, Scheme 1), which act as redox relays.
electrode by the redox polymer self-contained mediator. We have shown that AFM reveals individual GOx molecules adsorbed onto discrete, positively charged cysteamine adsorbed on Au, while 2D protein aggregation is apparent when the enzyme is adsorbed on a PAH-Os polycation[3b]on the surface, with the enzyme crowding on the polyion strands at the surface. We explored several spatial configurations of the enzyme and the redox polymer as shown in Figure 2: The GOx monolayer can
Scheme 1. Molecular structure of poly(allylamine) modified with [Os(bpy)2ClpyCHO] .
Glucose oxidase is a homodimer of 186 kDa that catalyzes glucose oxidation via flavin adenine dinucleotide (FAD) reduction as shown in Figure 1. Since direct electron transfer from the prosthetic group to the electrode is hindered in GOx, the PAH/Os redox mediator ™wires∫ the enzyme and shuttles electrons from the FADH2 buried inside the protein structure to the electrode. The step-by-step electrostatic self-assembled adsorption of both components offers the possibility to restrict the deposition of an enzyme monolayer and to vary the spatial distribution of the enzyme and the Os molecular wire in a controlled way. Moreover, this organization in molecular dimensions brings the possibility to build different supramolecular architectures, in order to study the mechanisms of electrical signal generation from biomolecular recognition and to analyze how different enzyme architectures affect the ™wiring efficiency∫ by the redox polyelectrolyte and the fraction of ™wired∫ enzyme molecules. Here we focus on the effect of the two-dimensional spatial distribution of enzyme and redox mediator on the fraction of the enzyme molecules that are effectively ™wired∫ to the underlying
Figure 2. Different enzyme monolayers on electrostatically self-assembled electrodes: (a) PAH-Os/GOx, (b) cysteamine/GOx/PAH-Os, (c) PAH/GOx/PAH-Os, and (d) PAH-Os/GOx/PAH-Os.
be ™wired from underneath∫ when it is adsorbed on the PAH-Os (Figure 2 a); or ™wired from above∫ by PAH-Os when GOx is deposited on discrete, positive, short alkanethiol charged molecules like cysteamine (Figure 2 b); assembled over unmodified polymer, PAH, with PAH-Os on top ™wiring form above (Figure 2 c) or entrapped between two layers of PAHOs in an ∫above-below wiring™ fashion (Figure 2 d). It is worth mentioning that these schemes are only pictorial representations of the structures, as it has been shown that there is certain degree of layer interpenetration.[4] However, in these organized structures we can achieve a much better control over the position of the PAH-Os ∫wire™ respect to the GOx molecules than in a random hydrogel built with the same components. In the latter case, the random distribution of the PAH-Os around the enzyme prevents a systematic study of the vectorial redox charge transfer and propagation between the redox Figure 1. Reaction scheme of the electrochemical oxidation of b-D-glucose catalyzed by GOx mediator and the enzyme. and PAH-Os.
236
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org
ChemPhysChem 2004, 5, 235 ± 239
In a recent report we have shown the importance of electronhopping diffusion in these organized layer-by-layer enzyme multilayers, by shifting a single active GOx layer from the bottom to the top layer in a multilayer structure of inactive FAD-free GOx apoenzyme.[5] Herein, we focus the attention on the effect of supramolecular architecture on the fraction of wired enzyme molecules in the structure. Considering the set of reactions depicted in Figure 1 and neglecting diffusion of soluble substrate in the ultrathin enzyme film, the expression for the catalytic b-D-glucose oxidation current density is given by Equation (1):[3a]
Icat
2Fk cat G enz k cat KM 1 kOs
III S
(1)
The Michaelis constant for b-D-glucose is KM (k 1 kcat)/k1; kcat and k are the enzyme turnover (glucose oxidation) and the bimolecular FADH2 reoxidation rate constant respectively and Genz is the surface concentration of the total active enzyme wired by the Os polymer, while [S] is the concentration of b-D-glucose. F is the Faraday constant. From the glucose electrocatalytic oxidation current density it is possible to obtain the quantities k[Os] and Genz , by assuming kcat 700 s 1 and KM 25 mM, the enzyme parameters in solution.[6] The figure of merit for the molecular wiring efficiency is the FADH2 bimolecular reoxidation rate constant, k, which can readily be obtained from k[Os] once the volume concentration of mobile polymer-bound osmium sites colliding with the active enzyme FADH2 is evaluated. For this purpose we measured the electrical charge density, q, consumed during the oxidation of Os(II) to Os(III) in the absence of b-D-glucose and the ellipsometric film thickness, df, which led to the osmium site concentration, [Os] q/(FAdf).[7] The fraction of molecularly wired enzyme in the different supramolecular structures was evaluated by comparing Genz, the surface concentration of the total active enzyme wired by the Os polymer, with the total mass of enzyme deposited in each layer, which was obtained from the frequency shift in a quartz crystal microbalance (QCM) during electrostatic adsorption of the enzyme at the osmium polyelectrolyte-functionalized layer.[3] Inspection of Equation (1) shows that the catalytic current alone is not a good indicator of either the wiring efficiency or the fraction of wired enzyme, since the flux of electrons in the external circuit depends on the FADH2 oxidation rate, k, the redox polymer concentration, [Os], and on the concentration Genz of the wired enzyme molecules, which is a small fraction of the total number of active enzyme molecules on the surface.[8] Nevertheless, the strategy described above allowed us to separate these factors by using a suitable kinetic analysis of Equation (1). The voltammetric response of the PAH-Os adsorbed in the absence of glucose is shown in Figure 3 a and is characterized by the reversible oxidation of the surface-confined osmium groups. The formal oxidation potential, Eo', measured from the cyclic voltammogram, was 0.27 V vs. saturated calomel electrode (SCE) ChemPhysChem 2004, 5, 235 ± 239 www.chemphyschem.org
Figure 3. (a) Cyclic voltammetry at 50 mV s 1 in the absence of glucose for a PAH/GOx/PAH-Os electrode in 0.1 M Tris(hydroxymethyl)aminomethane (Tris) buffer, pH 7.5 and 0.2 M KNO3 . (b) Dependence of the steady-state glucose catalytic oxidation current density on potential for the same electrode in 0.1 M Tris buffer, pH 7.5, 0.2 M KNO3 and increasing glucose concentration (2.5, 7.5, 25 and 60 mM) at 5 mV s 1.
and the difference between the anodic and cathodic peak potential (DEp) was typically smaller than 10 mV, at scan rates between 5 mV s 1 to 1000 mV s 1. Thus, the surface coverage of redox sites can be determined from the integration of the anodic or cathodic waves at low scan rates. On the other hand, Figure 3 b shows typical catalytic current ± voltage curves for an electrode coated with PAH/GOx/PAH/Os in solutions of varied glucose concentrations. As the potential increases the Os(II) within the film is oxidized to Os(III) and we observe a catalytic current for the enzyme substrate oxidation as described by Equation (1). If E Eo', all osmium sites in the polyelectrolyte film are oxidized and the electrocatalytic current for the oxidation of b-Dglucose depends on the concentration, as shown in Figure 4. The best fit to Equation (1) is shown by the continuous line and the resulting values of k[Os] and Genz are compared in Table 1 for the different molecular architectures, that is, spatial distribution of enzyme and redox polyelectrolyte mediator: a) PAH-Os/GOx; b) cysteamine/GOx/PAH-Os; c) PAH/GOx/PAH-Os and d) PAH-Os/ GOx/PAH-Os, as depicted in Figure 2. The relevant difference is the highest fraction of wired enzyme for system labeled as (c), where almost 30 % of the enzyme molecules present at the surface are electrically wired to the electrode, unlike system (a) where only 1 % of the GOx molecules contribute to the catalytic
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
237
Table 1. Wiring efficiency and amount of enzyme wired for different molecular architectures.[a] Experiment
k[Os] [s 1]
[Os] [mM]
k [103 M
a) MPS/PAH-Os/GOx b) cysteamine/GOx/PAH-Os c) MPS/PAH/GOx/PAH-Os d) MPS/(PAH-Os)2/(GOx)1
590 846 2481 1142
124 96 278 183
4.7 8.8 8.9 6.3
1
s 1]
Genz [10 0.12 0.31 0.41 0.74
12
mol ¥ cm 2]
(1.1 %) (6.6 %) (29.3 %) (6.6 %)
[Os]/[GOx] 3.2 10.2 99.3 8.7
[a] The values in parentheses represent the percentage of enzyme wired relative to the enzyme detected by QCM.
Figure 4. Typical catalytic current response for b-D-glucose concentration for self-assembled nanostructured thin films based on different architectures: (a) PAH/Os/GOx, (b) cysteamine/GOx/PAH-Os, (c) PAH/GOx/PAH-Os, (d) (PAH-Os)2/ (GOx)1 . All measurements were performed in 0.1 M Tris buffer of pH 7.5, 0.2 M KNO3 .
current by collision with the mediator. The structures labeled as (b) and (d) show a similar lower fraction of wired enzyme; around 6 %. This contrasts to the high fraction of the enzyme that can be oxidized by a soluble mediator such as ferrocene ± methanol. Notice the remarkable increase in the catalytic response for a GOx monolayer when it is wired from above and below (Figure 4 d) with respect to the PAH-Os underneath the enzyme layer (Figure 4 a). Charge reversal and overcompensation in the surface layer is achieved by loops and tails that can wave out into the electrolyte with a population of soluble counter-ions. These loops and tails are expected to have enhanced degrees of freedom as compared to inner polyion pairs.[9] The surface polyion loops and tails are less involved in electrostatic interactions than polyion trains and thus have more flexibility by segmental motion as molecular wires of the redox enzyme. It should be noted that the role of the osmium poly(allylamine) polycation is twofold: Firstly, it acts as the glue to hold the GOx building blocks together in the nanostructure, and secondly it also functions as a molecular wire, connecting the enzyme FADH2 to the underlying electrode. Close examination of the last two columns in Table 1 shows that the larger the ratio of the osmium to the total enzyme concentration (as detected by
238
QCM), the larger the fraction of wired enzyme. Therefore the larger the ratio of Os to total GOx concentration, the greater the excess of flexible osmium polymer segments that can collide with the enzyme and hence the greater the fraction of Os that can be positioned at the enzyme surface with the proper distance to the FADH2 moiety. Conversely, if there is a shortage of polymer relative to the enzyme (cases a, b and d) then most of the polymer segments will be electrostatically bound to the enzyme and the less chance they have to interact with another enzyme molecule. The bimolecular FADH2 oxidation rate constants, k in Table 1, are very similar in all cases and result only slightly higher for case (b) and (c), where GOx was adsorbed on osmium free cysteamine or unmodified poly(allylamine). It is interesting to note that the values of k in Table 1 are comparable to that reported for the intermolecular oxidation in solution of GOx modified covalently with the same osmium complex, k 8.3 103 M 1 s 1.[10a] However, in all cases k values are much lower than the value for the bimolecular oxidation rate of the soluble enzyme with soluble osmium complex, approximately k 3.7 105 M 1 s 1, where free diffusion collision of the osmium complex with the enzyme can take place.[10a] On the other hand, intramolecular oxidation of GOx molecules with tethered Os complexes has been shown to be less efficient, with k 0.2 to 0.4 s 1.[10b] In conclusion, unlike the bimolecular FADH2 oxidation rate constant or ™wiring efficiency∫, which remains almost unchanged for the different monolayer architectures studied, the fraction of enzyme molecules that are effectively wired by the redox-active polyelectrolyte mediator strongly depends on the ratio of [Os] to [GOx]. Moreover, because the system is now immobilized and integrated in a supramolecular architecture, the relative amount of polymer-bound redox centers with respect to the GOx molecules limits the enzyme-electrode response. The results reported here provide new insights into the rational design of molecular recognition devices.
Experimental Section Sodium 3-mercapto-1-propane sulfonate (MPS) and cysteamine were purchased from Aldrich and Sigma, respectively, and used without further purification. Gold flags (0.5 1.0 cm2) used as working electrodes were cleaned as previously reported and cycled in 2 M H2SO4 between 0.1 and 1.6 V at 0.1 V s 1 to check for surface contamination. Electrochemically active surfaces were calculated from the reduction peak of gold oxide.[3] For electrode modification, the gold electrode was first immersed in an aqueous solution of 0.02 M MPS in 0.01 M H2SO4 or 0.02 M
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org
ChemPhysChem 2004, 5, 235 ± 239
cysteamine ethanolic solution for 2 h. The Au samples were then removed from the adsorption solution and thoroughly washed in distilled water. The poly(allylamine) osmium derivative was adsorbed on the modified gold electrode from 0.4 % w/v aqueous solutions for 10 min. GOx was adsorbed from a 1 mM aqueous solution. After adsorption the modified electrodes were thoroughly rinsed with distilled water. The redox polymer Os(bpy)2ClPyNHpoly(allylamine) (PAH-Os) was synthesized as previously reported.[10] Glucose solutions were prepared from a stock solution equilibrated in the anomers, and kinetics herein refer to the total glucose concentration. All solutions were prepared in Milli-Q(Millipore¾)water and stored at 4 8C. Enzyme catalysis experiments were carried out under N2 atmosphere. A standard three-electrode electrochemical cell was employed with a home-built operational amplifier potentiostat. The reference electrode was a saturated calomel electrode, and all potentials herein are quoted with respect to this reference; a platinum auxiliary electrode was employed.
Acknowledgements This work was supported by the Argentine Research Council (CONICET) and Motorola Semiconductor Sector (Arizona (USA)). E.J.C. acknowledges a Guggenheim Fellowship 2000 ± 2001. Keywords: electrochemistry ¥ enzyme catalysis ¥ monolayers ¥ nanostructures ¥ sensors
[1] a) G. Decher, J.-D. Hong, Ber. Bunsen-Ges. 1991, 95, 1430; b) Y. Lvov in Protein Architecture: Interfacing molecular assembly and immobilization biotechnology (Eds. Y. Lvov, H. Mˆhwald), Marcel Dekker, New York, 2000, pp. 1 ± 396. [2] a) R. Rajagopalan, A. Heller in Molecular Electronics (Eds.: J. Jortner, M. Ratner), Blackwell Science, Oxford, 1997, pp. 241 ± 254; b) I. Willner, E. Katz, Angew. Chem. 2000, 112, 1230; Angew. Chem. Int. Ed. 2000 l39, 1180, and references therein; c) E. Katz, I. Willner, A. B. Kotlyar, J. Electroanal. Chem. 1999, 479, 64 ± 68; d) S. C. Barton, H. H. Kim, G. Bingaimin, Y. Zhang, A. Heller, J. Am. Chem. Soc. 2001, 123, 5802 ± 5803. [3] a) J. Hodak, R. Etchenique, E. J. Calvo, K. Singhal, P. N. Bartlett, Langmuir 1997, 13, 2708 ± 2716; b) E. J. Calvo, R. Etchenique, L. Pietrasanta, A. Wolosiuk, C. Danilowicz, Anal. Chem. 2001, 73, 1161 ± 1168; c) E. J. Calvo, F. Battaglini, C. Danilowicz, A. Wolosiuk, M. Otero, Faraday Discuss. 2000, 116, 47 ± 65. [4] a) G. Decher, Science 1997, 277, 1232 ± 1237; b) M. Lˆsche, J. Schmitt, G. Decher, W. G. Bouwman, K. Kjaer, Macromolecules 1998, 31, 8893 ± 8906. [5] E. J. Calvo, C. Danilowicz, A. Wolosiuk, J. Am. Chem. Soc. 2002, 124, 2452 ± 2453. [6] R. Wilson, A. P. F. Turner, Biosens. Bioelectron. 1992, 7, 165. [7] E. S. Forzani, M. Otero, M. A. Perez, M. L. Teijelo, E. J. Calvo, Langmuir 2002, 18, 4020 ± 4029. [8] The total number of active enzyme molecules in the structure can be assessed by using a soluble mediator such as ferrocene metanol with a lower redox potential than the osmium complex. [9] K. Lowack, C. A. Helm, Macromolecules 1998, 31, 823 ± 833. [10] a) C. Danilowicz, E. Corto¬n, F. Battaglini, J. Electroanal. Chem. 1998, 445, 89 ± 94; b) F. Battaglini, P. N. Bartlett, J. H. Wang, Anal. Chem. 2000, 72, 502 ± 509.
Energy Transfer from Dye ± Zeolite L Antenna Crystals to Bulk Silicon Stefan Huber and Gion Calzaferri*[a] The possibility of radiationless electronic excitation energy transfer due to a dipole ± dipole interaction from an excited molecule to a nearby semiconductor was first proposed by Dexter.[1] Different attempts followed to observe this experimentally by measuring the photophysical properties of monolayers or sub-monolayers of fluorescing dyes on semiconductor surfaces separated by a spacer layer thus preventing direct electronic contact.[2±6] However, to obtain appreciable electronic excitation energy transfer to the semiconductor, the dyes on it should absorb most of the light, which is not achievable by a monolayer. Multilayers cannot be used because of the very efficient quenching of the electronic excitation. We have therefore devised a new type of host ± guest material that enables us to circumvent this problem. The material consists of zeolite L crystals containing organized dye molecules that behave as monomers. As a host we used zeolite L, a crystalline aluminosilicate with hexagonal symmetry, which consists of one-dimensional channels running through the whole crystal. The latter are ideal for incorporating organic dye molecules. Details of these host ± guest materials, which allow light harvesting within the volume of the crystals and fast excitation energy transport to stopcock molecules–acceptor dyes located at the external zeolite L surface–have recently been reported. We also showed that these stopcock molecules attached selectively to the base of the zeolite L crystals because of their shape, and due to the fact that the base and coat of the crystals have distinctly different chemical properties.[7±9] Here, we report the first successful experiments on excitation energy transfer from dyes inside the zeolite L channels to dyes covalently bound on its external surface and further through a thin layer of silicon dioxide, preventing electron transfer, to a silicon semiconductor. The system is explained in Figure 1 where we show a single channel of such an acceptor,donor-zeolite L composite placed on a semiconductor. We used N,N'-bis(2,6-dimethylphenyl) perylene3,4,9,10-tetracarboxylic diimide (DXP, 1)
as the donor encapsulated in zeolite L, and ATTO680 as the acceptor bound to the surface of the crystals because both are strongly luminescent, stable dyes with matching absorption and fluorescence spectra. ATTO680 is an aminoreactive dye whose structure has not yet been published by the supplier (Fluka). The
Received: August 4, 2003 [Z 930]
[a] S. Huber, Prof. Dr. G. Calzaferri Department of Chemistry and Biochemistry, University of Bern Freiestrasse 3, 3012 Bern (Switzerland) Fax: ( 41) 31-631-39-94 E-mail:
[email protected]
ChemPhysChem 2004, 5, 239 ± 242
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
239
