Langmuir 2007, 23, 6997-7002
6997
Surface-Immobilized Pyridine-Functionalized γ-Cyclodextrin: Alkanethiol Co-adsorption-Induced Reorientation Colm T. Mallon,† Robert J. Forster,*,† Andrea McNally,† Elena Campagnoli,† Zoe Pikramenou,‡ and Tia E. Keyes*,† National Center for Sensor Research, School of Chemical Sciences, Dublin City UniVersity, Dublin 9, Ireland, and School of Chemistry, The UniVersity of Birmingham, Edgbaston, Birmingham, B152TT, U.K. ReceiVed January 24, 2007. In Final Form: April 12, 2007 Monolayers of di-6A,6B-deoxy-6-(4-pyridylmethyl)amino-γ-cyclodextrin (γ-CD-(py)2) have been formed on polycrystalline platinum electrodes and investigated using electrochemical and surface-enhanced Raman spectroscopy (SERS). The behavior of self-assembled monolayers of (γ-CD-(py)2) alone, (γ-CD-(py)2) backfilled with 1-nonanethiol, and 1-nonanethiol are reported. The potential dependence of the capacitance indicates that the film capacitance is higher for the backfilled CD layers than for 1-nonanethiol layers, most likely due to ion flux through the CD cavity. SERS spectra of the backfilled layer exhibit features associated with both pyridine-functionalized CD and alkane moieties. Investigations using [Fe(CN)6]4- as a solution-phase probe indicate that the backfilled CD-alkane thiol layer exhibits enhanced blocking properties compared to γ-CD-(py)2 films alone. Complete blocking was achieved by a combination of backfilling and insertion of a high-affinity guest 1-adamantylamine into the cavity. Significantly, an electroactive guest with high affinity for γ-CD, [Co(biptpy)2]2+, does not exhibit a redox response at the γ-CD-(py)2 layer but molecular recognition is turned on by backfilling the CD layer with 1-nonanethiol molecules. This switching on of the electrochemical activity suggests that the CD hosts are initially inaccessible but reorientate upon backfilling, exposing the CD opening to solution and permitting a supramolecular host-guest complex to form. The binding of [Co(biptpy)2]2+ to γ-CD in the backfilled monolayer depends on the bulk concentration of guest and is modeled by the Langmuir isotherm, yielding an association constant for the Co2+ state of 1.45 ( 0.46 × 105 M-1 and a limiting surface coverage 1.49 ( 0.25 × 10-11 mol cm-2. The surface coverage of the divalent state is higher than the trivalent state, reflecting the dynamic nature of the inclusion.
Introduction Building architecture for molecular electronics applications has traditionally relied on solution-phase approaches through covalent linkage of low-molecular-weight components, polymers, thin films, or colloids. However, this is limited by its synthetic complexity, lack of spatial control over individual molecular units in macromolecular systems, and ill-defined charge, energy, and mass transport pathways. An alternative and more promising tactic is to use nature’s self-assembly approach based on reversible H-bonding or hydrophobic-hydrophobic interactions to achieve a controlled and well-defined functional architecture.1 Host type molecules, such as cyclodextrins (CDs), crown ethers, and calixarenes, have attracted much attention as they have thermodynamically favorable interactions with specific guests and selectivity is achieved since the strength of this host-guest interaction depends on the size and polarity of both the host and the guest.2,3 Interfacial self-assembled monolayers (SAMs) offer the possibility of directional control of the assemblies out from the surface.4 Usually, the immobilization of CD layers onto solid supports occurs by self-assembly from solution; in the case of gold substrates, the CDs are typically modified with thiols5-9 or * To whom correspondence should be addressed. E-mail: tia.keyes@ dcu.ie (T.E.K.). † Dublin City University. ‡ The University of Birmingham. (1) Lehn, J. M. Angew. Chem., Int. Ed. 1990, 29, 1304. (2) Rekharaky, M. V.; Inoue, Y. Chem. ReV. 1998, 98, 1875. (3) Conn, M. M.; Rebek, J. Chem. ReV. 1997, 97, 1647. (4) Flink, S.; van Veggel, F. C. J. M.; Reinhoudt, D. N. AdV. Mater. 2000, 12, 1315. (5) Meada, Y.; Fukuda, T.; Yamamoto, H.; Kitano, H. Langmiur 1997, 13, 4187.
sulfides,10 which act as a specific anchors to covalently tether the CDs to the surface. However, the surface coverage of the CDs is frequently less than that anticipated for a dense monolayer.11 The gaps or defects can be “backfilled” with an alkane thiol after first blocking the CD cavities with a suitable guest. This process yields tightly packed layers which can block reactions of solution-phase probes at the underlying electrode surface. This two-step adsorption method has been adopted by a number of groups.12-14 However, details concerning the structural consequences of mutual interactions of the alkane thiol, CD, and metal surface during the process of film formation remain incomplete. There are a few recent examples of co-adsorption induced orientational changes in monolayer structures, for example, thiol-induced orientational changes have been described for DNA,15 tetradecylmethyl viologen,16 and ferrocene.17 Re(6) Endo, H.; Nakaji-Hirabayashi, T.; Morokoshi, S.; Gemmei-Ide, M.; Kitano, H. Langmuir 2005, 21, 1314. (7) (a) Michalke, A.; Janshoff, A.; Steinem, C.; Henke, C.; Sieber, M.; Galla, H. J. Anal. Chem. 1999, 71, 2528. (b) Maeda, Y.; Kitano, H. J. Phys. Chem. B 1995, 99, 487. (8) Henke, C.; Steinem, C.; Janshoff, C.; Steffan, G.; Luftmann, H.; Sieber, M.; Galla, H. J. Anal. Chem. 1996, 68, 3158. (9) Hill, W.; Fallourd, V.; Klockow, D. J. Phys. Chem. B 1999, 103, 4707. (10) Beulen, M. W. J.; Bu¨gler, J.; de Jong, M. R.; Lammerick, B.; Huskens, J.; Scho¨nherr, H.; Vancso, G. J.; Boukamp, B. A.; Wieder, H.; Offenha¨user, A.; Knoll, W.; van Veggel, F. C. J. M.; Reinhoudt. D. N. Chem. Eur. J. 2000, 6, 1176. (11) Rojas, M. T.; Ko¨niger, R.; Stoddart, J. F.; Kaifer, A. E. J. Am. Chem. Soc. 1995, 117, 336. (12) Lee, J. Y.; Park, S. M. J. Phys. Chem. B 1998, 102, 9940. (13) Choi, S. J.; Choi, B. G.; Park, S. M. Anal. Chem. 2002, 74, 1998. (14) de Jong, M.; Huskens, J.; Reinhoudt, D. N. Chem. Eur. J. 2001, 7, 4164. (15) Arinaga, K. K.; Rant, U.; Tornow, M.; Fujita, S.; Abstreiter, G.; Yokoyama, N. N. Langmuir 2006, 22, 5560. (16) Reipa,V.; Laura Yeh, S.-M;. Monbouquette, H. G.; Vilker, V. L. Langmuir 1999, 15, 8126. (17) Lee, M.; Chung, C. Bull. Kor. Chem. Soc. 1999, 20, 132.
10.1021/la070212a CCC: $37.00 © 2007 American Chemical Society Published on Web 05/24/2007
6998 Langmuir, Vol. 23, No. 13, 2007
orientation in CD films induced by co-adsorption may significantly impact on the ability to form a host-guest complex. This work describes the adsorption of di-6A,6B-deoxy-6-(4pyridylmethyl)amino-γ-cyclodextrin (γ-CD-(py)2) on polycrystalline platinum electrodes. While much work has been carried out on thiol-functionalized R- and β-CD based monolayers, beyond the important work of Majda18 and Suzuki19 the properties of γ-CD monolayers remains relatively unexplored. Here, we demonstrate the binding of γ-CD difunctionalized with pyridine onto platinum and characterization of this monolayer using electrochemical and spectroscopic techniques including cyclic voltammetry and surface-enhanced Raman spectroscopy (SERS). These layers are backfilled with 1-nonanethiol to block defects in the films. Significantly, backfilling appears to dramatically affect the layer structure and the electrochemical data are consistent with CDs undergoing a reorientation during the backfilling process. This reorientation most likely involves breakup of face-to-face CD dimers switching on the monolayers ability to selectively bind a guest. For example, electroactive guest [Co(biptpy)2]2+ binds to the monolayer only after the CD layer has been backfilled. Significantly, the association constant for this interaction has been found from the Langmuir isotherm allowing the thermodynamics of binding in solution and at this oriented interface to be compared. Molecular recognition is a key element of research in fields as diverse as sensing and molecular electronics.1 Host-type molecules have attracted significant attention as they have thermodynamically favorable interactions with specific guests and selectivity can be achieved since the strength of this host-guest interaction depends on the size and polarity of both the host and the guest.2 Understanding what controls how SAMSs of host systems are oriented at an interface may be key to many of these applications. Experimental Section Materials. 1-Nonanethiol (95%), H2SO4 (99.999%), Na2SO4 (99%), ethanol (99.5%), potassium ferrocyanide(II) hydrate (99.99%) and 1-adamantylamine (97%) were obtained from Sigma-Aldrich. Acetonitrile (99.9%) was received from Fluka, and alumina powder was obtained from Buehler. All chemicals were used as received. Cobalt bisdiphenylterpyridine, [Co(biptpy)2]2+, was synthesized as reported previously.20 Synthesis of di-6A,6B-deoxy-p-toluenesulphonyl-γ-cyclodextrin, γ-CD-(OTs)2. To a solution of γ-CD (57.07 g, 44 mmol) in NaOH (1500 cm3, 0.4 mol dm-3), p-toluenesulphonyl chloride (54.62 g, 286.5 mmol) was added with vigorous stirring at 0 °C. The resulting solution was stirred for a further 3 h at the same temperature and then filtered by gravity to remove the excess of p-toluenesulphonyl chloride. The filtrate was neutralized with HCl (1 mol dm-3), and a white precipitate formed, which was collected by gravity filtration. The crude product was recrystallized from hot water and dried in vacuo for 4 h at 40 °C. (Yield: 15.75 g, 22%). C62H92O44S2· H2O MW ) 1622 g/mol. Molecular ion peak at m/z ) 1627.4 ([M+ + H]). 1H NMR (400 MHz, DMSO) HOTs-a,a′ 7.740 (d, 4H), HOTs-b,b′ 7.426 (d, 4H), OHGlu-2 + 3 5.891-5.722 (m, 16H’s), OHGlu-6 4.600-4.154 (m, 8H), HGlu-1 4.901-4.794 ppm (m, 8H). Synthesis of di-6A,6B-deoxy-6-(4-pyridylmethyl)amino-γ-cyclodextrin, γ-CD-(Py)2. γ-CD-(OTs)2 (4.0 g, 2.46 mmol) and 4-(aminomethyl)pyridine (16 g, 147.6 mmol) were reacted in DMF (40 cm3) in an inert atmosphere at 90 °C for 6 h. The reaction mixture was cooled and poured into acetone. The products precipitated from the solution. The precipitates were collected by filtration and (18) Chamberlain, R.; Slowinska, K.; Majda, M. Langmuir 2000, 16, 1388. (19) Suzuki, I.; Murakami, K.; Anzai, J.; Osa, T.; He, P.; Fang, Y. Mater. Sci. Eng. C 1998, 6, 19. (20) Alcock, N. W.; Barker, P. R.; Haider, J. M.; Hannon, M. J.; Painting, C. L.; Pikramenou, Z.; Plummer, E. A.; Rissanen K.; Saarenketo, P. J. Chem. Soc., Dalton Trans. 2000, 1447.
Mallon et al. washed with methanol to remove unreacted 4-(aminomethyl)pyridine. The solid was dissolved in a minimum of warm water. Keeping the solution at 5 °C overnight resulted in recrystallization. The product was recrystallized again from water and dried in vacuo at 40 °C for 4 h. (Yield 2.94 g, 78%). C60H92O38N4·4H2O MW ) 1548 g/mol. Molecular ion peak at m/z ) 1553.6 ([M+ + H]). 1H NMR (400 MHz, DMSO) Hpy-3,5 8.425 (d, 4H), HPy-2,67.309 (d, 4H), OHGlu-2 + 3 6.009-5.650 (m, 16H), OHGlu-6 4.664-4.509 (m, 8H), HGlu-1 4.892-4.828 ppm (m, 8H). Methods. Electrospray mass spectrometry in was conducted in positive-ion mode using a Bruker LC/MS Esquire series. Electrochemistry was performed in a standard three-electrode cell with platinum polycrystalline working electrode, (diameter 0.5 mm), Ag/AgCl reference electrode saturated with KCl, and a platinum mesh as a counter electrode at 23 ( 2 °C. All solutions were degassed with Argon, and a blanket of Argon was maintained over the solutions during experiments. Cyclic voltammetry and impedance measurements were preformed on a CH Instruments Model 600 electrochemical workstation. Raman spectroscopy was preformed on a Horiba Jobin Yvon HR800UV using an argon ion Laser (458 nm) as the exciting wavelength focused through a 10× objective lens on the electrode surface. Focusing was confirmed by using a video camera. Ten spectral acquisitions were accumulated, and each acquisition was 10 s in length. Electrodes were cleaned prior to use by polishing with alumina (0.05 µm) followed by sonication in Milli-Q water. The electrodes were electrochemically cleaned by scanning in 0.5 M H2SO4 between -0.25 and 1.45 V until stable voltammograms were obtained. Electrodes then underwent further scanning between -0.2 and 1.35 V followed by holding the potential at 0.3 V for 30 s to remove adsorbed hydrogen. The reference electrode was separated from the solution by a salt bridge filled with 0.5 M H2SO4 to prevent contamination of the working electrode with chloride ions. The area of the working electrodes was determined by integration of the oxide reduction peak taking the charge per square centimeter area as 420 µC for platinum.21 Monolayers of γ-CD-(py)2 were formed by immersing the working electrode into solutions containing 100 µM γ-CD-(py)2 overnight. Backfilling of the layers was preformed by subsequently immersing the γ-CD-(py)2 layers in a solution containing 2 mM 1-admantylamine and 0.2 mM 1-nonanethiol in ethanol. The adamantyl group has a strong affinity for the CD cavity22 and is expected to block the cavities while the alkanethiol fills the defects. Backfilled monolayers were thoroughly washed with Milli-Q water before use. Solutions of [Co(biptpy)2]2+ in the concentration range 1-40 µM were prepared in 90:10 H2O/ACN v/v, with 0.18 M Na2SO4 as the supporting electrolyte by dissolving the [Co(biptpy)2]2+ first in the acetonitrile. Electrodes were roughened for SERS by a modification of the method described by Tian and co-workers.23 The electrodes were cleaned as described above, and the potential was stepped between -0.2 and 2.4 V 20 000 times with a step length of 0.02 s in 0.5 M H2SO4. The potential was then held at 0 V for 400 s to fully reduce the surface oxide followed by further cycling in 0.5M H2SO4 between -0.2 and 1.35 V until the voltammogram was stable. The average surface roughness after this treatment was 7. Layer formation on these electrodes was identical to that described above.
Results and Discussion Capacitance Properties. As a monolayer assembles, it displaces solvent and ions at the electrode solution interface, typically causing a decrease in the interfacial capacitance. Figure 1 shows the capacitance versus potential curves for platinum electrodes modified with a γ-CD-(py)2 monolayer and following (21) Trasatti, S.; Petrii, O. A. Pure Appl. Chem. 1991, 63, 711. (22) Onclin, S.; Mulder, A.; Huskens, J.; Ravoo, B. J.; Reinhoudt, D. N. Langmuir 2004, 20, 5460. (23) Ren, B.; Huang, Q. J.; Cai, W. B.; Mao, B. W.; Liu, F. M.; Tian, Z. Q. J. Electroanal. Chem. 1996, 415, 175.
Surface-Immobilized Pyridine-Functionalized γ-Cyclodextrin
Figure 1. Capacitance-potential profiles for (a), bare, electrochemically cleaned platinum, (b) γ-CD-(py)2 monolayer, (c) backfilled CD layer after 24 h in sealing solution, and (d) 1-nonanethiol layer on platinum measured with AC impedance using a frequency of 520 Hz and amplitude of 5 mV. The 1-nonanethiol layer was formed on electrochemically cleaned platinum under identical conditions to the backfilling process. In each instance aqueous 0.2 M Na2SO4 was employed as electrolyte.
backfilling with 1-nonanethiol. In all cases the supporting electrolyte is 0.2 M Na2SO4. There is a small minimum in the capacitance curve for the bare electrode at ∼ -0.05 V which may be attributable to the potential of zero charge24 (PZC) and the order of magnitude of the capacitance is consistent with other studies.25 This feature is also present, although greatly suppressed, at the γ-CD-(py)2 monolayer but is lost at the 1-nonanethiol backfilled and 1-nonanethiol layers which display capacitance profiles which are almost independent of potential, which suggests a densely packed monolayer has formed. The drop in capacitance between curves (a) and (b) is consistent with formation of an adsorbed organic film at the electrode surface upon overnight exposure to a 100 µM solution of γ-CD-(py)2. The small error bars on the capacitance-potential profile for the γ-CD-(py)2 layer indicate reproducible layer formation. The capacitance values for the γ-CD-(py)2 layer are considerably higher than those found for classical alkanethiol layers (1-5 µF cm-2), indicating a more permeable film. This is likely to be a consequence of unmodified platinum between adjacent CDs and ion permeation through the CD cavity. These values are consistent with those found in thiol β/γ-CD monolayer studies.10,11 The potential-capacitance curves, Figure 1c and d, show the capacitance profiles for the γ-CD-(py)2layer and bare electrode, respectively, following 24 h exposure to the backfilling 1-nonanethiol/1-adamantylamine solution. The backfilled CD layer exhibits a higher capacitance (9.5 ( 0.4 µF cm-2) than the alkanethiol layer formed on the electrochemically cleaned platinum (6.3 ( 0.4 µF cm-2). This suggests that the alkanethiol molecules do not displace the pyridine bound CDs from the surface as this difference in capacitance must arise from the presence of the CD in the layer. Further capacitance monitoring of the backfilling process reveals that the CD layer capacitance continues to decrease slowly over a time scale of up to ∼100 h, to a final capacitance of 6-7 µF cm-2, while the alkanethiols final capacitance after the same period is 3-4 µF cm-2. Therefore, while further annealing of both layers with time appears to occur, the capacitance difference between the films is essentially constant. This further confirms that the pyridine bound CD are not lost from the surface during the backfilling process. (24) Weaver, M. J. Langmuir 1998, 14, 3932. (25) Walsh, D. A.; Keyes, T. E.; Forster, R. J. J. Phys. Chem. B 2004, 108, 2631.
Langmuir, Vol. 23, No. 13, 2007 6999
Figure 2. CV showing the response of bare, electrochemically cleaned platinum (a), γ-CD-(py)2 monolayer (b) and γ-CD-(py)2 monolayer backfilled with 1-nonanethiol (c) in 1 mM [Fe(CN)6]4as a solution-phase probe in 0.2 M Na2SO4 as the supporting electrolyte. (d) The backfilled layer under identical conditions as before but with 100 µL of 0.02 M 1-adamantylamine in ethanol added to the solution. All scan rates are 0.1 V s-1. Table 1. Peak-to-Peak Separations (∆Ep) and Current Ratios for Various Electrode Modifications on Platinum in 1 mM [Fe(CN)6]4- with 0.2 M Na2SO4 as the Supporting Electrolyte at Scan Rates of 0.1 V s-1 electrode modification
∆Ep (mV)
Iox/I0
Ired/I0
bare γ-CD-(py)2 backfilled CD layer backfilled + adamantly inj
67 ( 4 109 ( 7 236 ( 26 NM
NA 0.8 ( 0.09 0.49 ( 0.03 NM
NA 0.93 ( 0.10 0.35 ( 0.08 NM
Blocking Behavior. The ability of SAMs to block access of a redox-active solution-phase probe to the electrode surface can be used to estimate the porosity or defect density in the film. The blocking properties of all the monolayers presented were investigated using 1 mM [Fe(CN)6]4- in 0.2 M Na2SO4 as the solution-phase probe. Figure 2 illustrates the voltammetric response for the probe on (a) bare electrochemically cleaned platinum, (b) γ-CD-(py)2 monolayer, and (c) backfilled CD layers. The response on bare platinum is close to reversible with a peak to peak separation (∆Ep) of 67 mV (ideally 57 mV for a oneelectron transfer). The voltammetry of ferrocyanide is known to depend strongly on the nature of the electrode surface26,27 and is complicated by ion-pairing effects. However, it is apparent that the ∆Epvalues are larger and the currents are reduced for the modified electrodes. Notwithstanding the complex nature of ferrocyanide voltammetry, this behavior is consistent with a slower rate of heterogeneous electron transfer for the electrodes modified with γ-CD-(py)2. Figure 2 illustrates the CV obtained for the γ-CD-(py)2 and backfilled CD layers. The ∆Ep and current ratio values for these scans are summarized in Table 1. The γ-CD-(py)2 layer shows a ∆Ep value of 109 ( 7 mV and the current is reduced by ∼10-15% compared to the bare electrode, indicating the CD monolayer formation somewhat impedes access of the probe to the electrode surface. However, Table 1 and Figure 2c show that the backfilled layer displays much greater blocking than the γ-CD-(py)2layer, presumably due to the blocking of defects between the CD cavities. While the blocking properties of the layer increases significantly upon backfilling, complete blocking of the [Fe(CN)6]4- probe can be observed (26) Davies, T. J; Moore, R. R.; Banks, C. E.; Compton. R. G. J. Electroanal. Chem. 2004, 574, 123. (27) Davies, T. J.; Banks, C. E.; Compton, R. G. J. Solid State Electrochem. 2005, 9, 797.
7000 Langmuir, Vol. 23, No. 13, 2007
Figure 3. Raman spectra of 1-nonanethiol (a), backfilled CD layer (b), and γ-CD-(py)2 (c) monolayers on roughened platinum electrodes using 458 nm as the excitation wavelength.
Mallon et al.
Figure 4. CV showing the response of electrochemically cleaned, bare platinum, (grey), and γ-CD-(py)2 monolayer (black) in 13 µM Co(biptpy)22+ in 90:10 H2O/ACN with 0.18 M Na2SO4 as the supporting electrolyte. The scan rate is 20 V s-1.
after addition of 100 µL of 0.02 M (1 mM in cell solution) 1-adamantylamine in ethanol (Figure 2d). As the adamantyl group is a high-affinity guest for CD this observation suggests that the cavities in the backfilled layer are capable of inclusion and have not been filled by the alkanethiols. Furthermore, suppression of current signal after the cavities are filled suggests that the space between cavities has been successfully blocked by the alkanethiols and that the backfilled layer is relatively defect free. Raman Studies of γ-CD-(py)2 and Backfilled Layers. Pyridine has been the subject of several SERS investigations on platinum23,28,29 and displays characteristic in plane vibration modes at ∼1000, 1200, and 1600 cm-1. Figure 3c shows the SERS spectra of γ-CD-(py)2 adsorbed on electrochemically roughened platinum and has bands at 1035, 1209, 1550, and 1606 cm-1 which are attributed to the adsorbed pyridine linkers and confirm adsorption of the γ-CD-(py)2 via its pyridine linkers. The spectrum also shows a band at 3066 cm-1 which is assigned to the aromatic C-H stretching in the pyridine. The spectrum for 1-nonanethiol shown in Figure 3a exhibits intense C-H stretch modes between 2850 and 2946 cm-1 and a band at 1537-1624 cm-1 attributed to CH3/CH2 deformation modes.30 The spectrum of the backfilled layer shown in Figure 3b contains contributions from both pyridine and alkanethiol moieties. The wavelength of the pyridine features are slightly shifted at 1025, 1204, 1541, and 1610 cm-1 but clearly confirm the persistence of γ-CD-(py)2in the backfilled layer. The alkanethiol nonaromatic C-H stretching can also be seen at 2850-2935 cm-1 as in the pure 1-nonanethiol layer, verifying that the alkyl thiol simply backfills, without displacing the γ-CD-(py)2layer. Taken together with the capacitance and electrochemical data, these results clearly indicate that the CDs functionalized with pyridine represent a flexible and stable approach to monolayer formation. Effect of Backfilling on CD Orientation. Figure 4 shows the voltammetric response of bare platinum and γ-CD-(py)2 monolayer in the presence of electroactive guest complex [Co(biptpy)2]2+ (13µM) in 90:10 v/v H2O/ACN with 0.18 M Na2SO4 as the supporting electrolyte. The complex displays broad, illdefined signals at bare platinum at 110 mV. However, the voltammetric response of the γ-CD-(py)2 monolayer shows that this signal is completely blocked by the CD film. This is surprising as the diphenyl moiety is expected to associate effectively with
the CD cavity. The inhibition of the [Co(biptpy)2]2+ signal suggests that the hydrophobic CD cavity is oriented in the film in such a way that its secondary face is not fully exposed to the solution. Weisser and co-workers have previously considered the possibility of a hexagonal close-packed or brick-packed model to describe the orientation of the CD tori on the electrode surface,31 and it would appear that the brick packed model best describes the CD orientation in the γ-CD-(py)2 monolayer, as this model sees the CD cavities facing each other which would prevent penetration into the cavity. As the CDs presented here possess only two surface linkers, they are anticipated to have significant orientational freedom compared with those in which all the sugars are bound. Hydrogen bonding between CDs is a tentative, but likely, explanation for the adoption of this brick-packing model in the γ-CD-(py)2 layer. In the crystalline state, cyclodextrins are well known to pack through intermolecular glucose hydroxyl H-bond interactions which may be mediated by water molecules.32 Head-to-head and head-to-tail type intermolecular H-bonding interactions have been particularly noted for γ-CD.33 Figure 5 shows the cyclic voltammetry of 13 µM [Co(biptpy)2]2+ in 90:10 H2O/ACN on platinum electrodes which are modified with backfilled γ-CD-(py)2 (a) and 1-nonanethiol (b) layers in 90:10 H2O/ACN with 0.18 M Na2SO4 as the supporting electrolyte. The response at the backfilled layer at 220 mV is consistent with a surface confined signal for [Co(biptpy)2]2+ as the peak current scales linearly with scan rate and plots of log peak current versus log scan rate have slopes of 1. The peak to peak separation is 40 mV which is higher than that expected for an ideal surface confined signal (0 mV) but significantly lower than that expected for a solution-phase species (57 mV). Remarkably, unlike the electrode modified only with γ-CD-(py)2, in the mixed SAM the cyclodextrin host becomes accessible. The backfilling process stimulates binding of the [Co(biptpy)2]2+ to the monolayer. The higher redox potential of the [Co(biptpy)2]2/3+ couple at the backfilled CD layer compared to bare platinum is consistent with a more nonpolar microenvironment at the electrode surface as the complex is harder to oxidize. Figure 4 also shows that the [Co(biptpy)2]2+ interacts with the 1-nonanethiol layer, presumably via hydrophobic interactions. However, the redox process mediated through the
(28) Bilmes, M. A.; Rubim, J. C.; Otto, A.; Arvia, A. J. Chem. Phys. Lett. 1989, 159, 89. (29) Cai. W.; She, C.; Ren, B.; Yao, J.; Tian, Z.; Tian, Z. Q. J. Chem. Soc., Faraday Trans. 1998, 94, 3127. (30) Bryant, M. A.; Pemberton, J. E. J. Am. Chem. Soc 1991, 113, 8284.
(31) Weisser, M.; Nelles, G.; Wohlfart, P.; Wenz, G.; Mittler-Neher, S. J. Phys. Chem. 1996, 100, 17893. (32) (a) Saenger, W.; Steiner, T. Acta Crystallogr. 1998, A54, 798 (b). Saenger, W, J. Incl. Phen. Macrocycl. Chem. 1984, 2, 445. (33) Saenger, W.; Steiner, T. Acta Crystallogr. 1998, B54, 450.
Surface-Immobilized Pyridine-Functionalized γ-Cyclodextrin
Figure 5. CV showing the response of backfilled CD (a) and 1-nonanethiol (b) layers in 13 µM [Co(biptpy)2]2+ in 90:10 H2O/ ACN with 0.18 M Na2SO4 as the supporting electrolyte. All scan rates are 20 V s-1.
Figure 6. Schematic illustrating possible orientation of the γ-CD(py)2 molecule on the platinum surface before (top) and after (bottom) backfilling with 1-nonanethiol. [Co(biptpy)2]2+ is shown including in the backfilled layer.
alkanethiol SAM occurs at significantly higher redox potential (∼100 mV) than that at the backfilled CD layer. This difference in potential between the alkanethiol and CD layers is likely to arise from the different micro-environments experienced by the guest. The more hydrophobic nature of the alkanethiol film is expected to make oxidation of the Co2+ center thermodynamically more difficult. Unlike the mixed CD containing SAM, reproducibility of the redox signal for the [Co(biptpy)2]2+/3+ couple at the 1-nonanethiol layer is extremely poor and interaction of the complex appears to depend on individual layer defects common
Langmuir, Vol. 23, No. 13, 2007 7001
Figure 7. Variation of surface coverage of [Co(biptpy)2]2+/3+ with bulk concentration. The solid lines are the best fits to the Langmuir isotherm. The inset shows the linearized Langmuir isotherm.
in alkanethiol layers formed on polycrystalline electrodes. The oxidation potential of the [Co(biptpy)2]2+, however, is never lower than that shown in Figure 5. In marked contrast, the backfilled CD layer displays excellent reproducibility, suggesting the presence of stable, reproducible binding site within the layer that is independent of film defects or electrode roughness. As no inclusion could be observed at the γ-CD-(py)2-only layer, it seems that the backfilling process alters the orientation of the cyclodextrin cavity within the layer. Previous work on CD layers in our group has suggested the reorientation of the CD in β-CD-(py) layers over time on gold as monitored by SERS spectroscopy which was attributed to the formation of intermolecular interactions between CD cavities.34 It is tempting to speculate that because of the good length match of the 1-nonanethiol to the γ-CD-(py)2 molecule, assembly of the alkanethiol molecules around the CD cavities forces the γ-CD(py)2 into an upright orientation while increasing the rigidity of the film and potentially disrupting any intermolecular interactions between the γ-CD-(py)2 (Figure 6) making the CD host more amenable to inclusion with bulky solution-phase guests. Alterations to SAM orientation have been noted previously for co-adsorption of alkyl thiol with DNA, tetradecylmethyl viologen, and ferrocene-terminated thiol; however, this is to our knowledge the first report of orientational changes in hostguest-type SAMs. Adsorption Isotherm for [Co(biptpy)2]2+ Inclusion. As inclusion into CD cavities is an equilibrium process, the number of cavities occupied is expected to vary with the concentration of guest in solution. The Langmuir isotherm has been used to determine the association constant (Kass) in several CD monolayer systems.5,6,11 A scan rate of 20 V s-1 was chosen to evaluate the [Co(biptpy)2]2+ interaction with the layer as this scan rate effectively excludes solution phase interference at the low concentrations employed in this study as peak currents are proportional to the root scan rate for solution-phase species but are directly proportional for surface bound species. Therefore, at high scan rates the surface bound signal dominates the response. Figure 7 shows the dependence of surface coverage of Cobalt complex included into the backfilled layer as a function of bulk concentration for Co2+ and Co3+ forms. The solid lines are the best fits of the Langmuir isotherm (eq 1) applied to the data, and the inset of Figure 6 shows the linear form of the data according to eq 2: (34) McNally, A. Ph.D. Thesis, Dublin Institute of Technology, 2005.
7002 Langmuir, Vol. 23, No. 13, 2007
Mallon et al.
Γ ) βC Γmax - Γ
(1)
C 1 C ) + Γ Γmax βΓmax
(2)
Where Γ is the surface coverage in mol cm-2, Γmax is the saturation surface coverage and β is the association constant for the guest-CD interaction, Kass. The good correlation of this plot indicates that the isotherm is a good model for the inclusion process, and the values found for Γmax and β are shown in Table 2. The saturation surface coverage of [Co(biptpy)2]2+/3+ ranges between 1.49 ( 0.25 × 10-11 and 1.03 ( 0.12 × 10-11 mol cm-2, ∼20% of that expected if a close-packed monolayer of γ-CD-(py)2 were formed (7 × 10-11 mol cm-2) and every cavity were filled by [Co(biptpy)2]2+/3+. However, the surface coverage found is consistent with other reports of guests included in CD monolayers.11,35,36 Given the large association constants, this low apparent surface coverage may arise either because dense monolayers of CD are not formed or not all of the cavities reorient upon backfilling with alkane thiol. Assuming all the available cavities are filled, an estimate of the area of occupation for the CD was calculated from the Γmax value for [Co(biptpy)2]2+ of 1114 Å. While the nature of the packing at the molecular level is not known, the average separation between CDs at the backfilled layer is estimated from the molecular diameter of γ-CD to be ∼18 Å. Significantly, the association constant for interfacial host-guest formation appears to be almost an order of magnitude larger than that found in solution. For example, the solutionphase association constant for an osmium-terpy-diphenyl complex with permethylated β-CD has been reported as 2 × 104 M-1.37 Both higher10,11,38 and lower39 binding constants have been found previously for immobilized CDs. Therefore, it appears that subtle changes in the CD orientation can impact on the thermodynamics of host-guest association. It is also perhaps important to note that there may be a significant interfacial electric field present in the case of interfacial host-guest formation depending on the redox potential of the guest and the ionic strength of the electrolyte solution. The free energy of adsorption can be calculated from the β parameter of the Langmuir isotherm according to eq 4:
∆Gads ) -RT ln β
(3)
where R is the gas constant and T is the temperature in K. The values calculated for [Co(biptpy)2]2+/3+ are shown in Table 2 and indicate that the binding strength of the cobalt complex into the CD cavity is comparable to that found for strongly bound (35) Wang, Y.; Kaifer, A. E. J. Phys. Chem. B 1998, 102, 9922. (36) Lahav, M.; Ranjit, K. T.; Katz, E.; Willner, I. Chem. Commun. 1997, 259. (37) Haider, J. M.; Chavarot, M.; Weidner, S.; Sadler, I.; Williams, R. M.; De Cola, L.; Pikramenou, Z. Inorg. Chem. 2001, 40, 3912. (38) Kitano, H.; Taira, Y.; Yamamoto, H. Anal. Chem. 2000, 72, 2976. (39) Kitano, H.; Taira, Y. Langmuir 2002, 18, 5835.
Table 2. Saturation Surface Coverages (Γmax), Association Constants (Kass), and Free Energy of Adsorption (∆Gads) for Oxidized and Reduced Forms of the CD Guest [Co(biptpy)2]2/3+ as Determined from the Langmuir Isotherm complex
Γmax (mol cm-2)
Kass ) β (M-1)
-∆Gads (kJ mol-1)
[Co(biptpy)2]2+ [Co(biptpy)2]3+
1.49 ( 0.25 × 10-11 1.03 ( 0.12 × 10-11
1.45 ( 0.46 × 105 2.67 ( 0.45 × 105
29.4 ( 0.7 30.9 ( 0.4
guests such as adamantyl derivatives with β/γ-CD and pyrene with β-CD in solution.2 The trivalent form of the cobalt complex is expected to exit the cavity as its bulk concentration is expected to be very low. The lower surface saturation coverage for the oxidized guest (1.03 × 10-11 mol cm-2) is consistent with this expectation and confirms the dynamic nature of the inclusion. However, a significant portion of the trivalent state remains in the cavities even on the tens of seconds time scale. This result suggests that the rate of host-guest formation is significantly larger than that for dissociation.
Conclusion SAMs of γ-cyclodextrins functionalized with two aminomethyl pyridine groups on platinum electrodes are reported for the first time. Electrochemical studies revealed that the layers are permeable to solution-phase probes and that backfilling of the defects in the layer with 1-nonanethiol greatly increased the blocking ability of the SAM. SERS spectra of backfilled layers showed both pyridine and alkanethiol bands, indicating that backfilling with alkanethiols did not displace the pyridine moieties of the CDs. The exposure of the SAM to a solution of inclusion complex [Co(biptpy)2]2+ was monitored electrochemically. Whereas no electrochemical signal was observed at the pure γ-CD-(py)2 SAM, a surface-confined and highly reproducible redox process associated with the [Co(biptpy)2]2+/3+ couple was observed at the backfilled SAM. This suggested that the backfilling process disrupts intermolecular interactions in the layer, causing orientational changes at the CD which permitted inclusion of the solution-phase guest. The inclusion process in the mixed SAM depended on the bulk concentration of [Co(biptpy)2]2+/3+ guest in solution, and this dependence was fitted to the Langmuir isotherm. Experimental data indicated that the association constant was larger at the monolayer than for corresponding systems in solution. Although alkyl thiol induced orientational changes have been noted previously in a small number of studies of mixed SAMs, this is the first time they have been noted in a SAM containing a host molecule and suggests a significant degree of control may be exerted over the monolayer architecture and its inclusion ability. This may have important implications for potential applications of host based SAMS in molecular sensing devices or molecular wire construction. LA070212A
