Optical and electrochemical properties of tunable host-guest complexes linked to plasmonic interfaces

This article describes the use of localized surface plasmon resonance (LSPR) interface to detect complexation/decomplexation steps of a controllable host–guest system at the solid–liquid interface. The LSPR interfaces consist of a sandwiched structure comprising a tin-doped indium oxide (ITO) substrate, gold nanostructures (Au NSs) and a thin ITO ﬁlm overcoating. ‘‘Click’’ chemistry was used to covalently link an alkyne-functionalized p -electron deﬁcient tetracationic cyclophane cyclobis(paraquat-p -phenylene) (CBPQT 4+ ) unit to an azide-terminated LSPR interface. The modiﬁed interfaces were characterized using X-ray photoelectron spectroscopy (XPS), cyclic voltammetry and UV-vis transmission spectroscopy. Tetrathiafulvalene (TTF) was used as a model guest molecule to demonstrate the possibility to follow the complexation/decomplexation events by monitoring the change in the LSPR signal. The results demonstrate that redox controlled host–guest complexation at the surface can be monitored effectively using LSPR.


Introduction
2][3][4] One of the most widely studied host systems is the p-electron deficient tetracationic cyclophane cyclobis(paraquat-p-phenylene) (CBPQT 4+ ), which readily forms redox-controllable host-guest systems with p-electron rich molecules such as tetrathiafulvalene (TTF) and 1,5-dioxynaphthalene derivatives. 5Although these systems have been extensively studied in solution phase using a range of analytical techniques, for device orientated applications, it is desirable that these systems are immobilized onto a surface. 6An early report on linking interlocked supramolecular systems of this type onto a gold surface was based on complexes formed between CBPQT 4+ and bis-thiol functionalised thread featuring a 1,4dialkyloxy phenyl unit. 7,8More recently, more elaborate systems have been developed using silane 9,10 or gold-sulfur based selfassembled monolayers (SAMS) [11][12][13][14][15] and electropolymerization techniques [16][17][18][19] to deposit CBPQT 4+ host-guest systems onto surfaces. 6,16,20,21An important consideration in designing surfaces functionalized with supramolecules is that the behavior of the anchored systems remains the same after surface attachment.
''Click'' chemistry, a generic term describing a range of chemical transformations characterized by high efficiency, selectivity and tolerance to a variety of solvents and functional groups, has proved to be a useful method to link supramolecular systems to solid substrates. 22The concept was introduced by Sharpless et al. 23 and is usually based on the copper(I) catalyzed triazole formation through the classic Huisgen 1,3-dipolar cycloaddition between azides and alkynes.The appealing characteristics of the azide-alkyne click reactions have led to its rapid a Institut de Recherche Interdisciplinaire (IRI, USR-3078), Universit ed e Lille1, Parc de la Haute Borne, 50 avenue de Halley, BP 70478, 59658 Villeneuve d'Ascq, France.E-mail: sabine.szunerits@iri.univ-lille1.fr;Fax: +33 3 62 53 17 01; Tel: +33 3 62 53  17  4À and CBPQT 4+ as well as cyclophane-modified ITO; UV/Vis spectra of alkynyl-functionalized CBPQT 4+ , TTF and the TTF-CBPQT 4+ guest-host complex.See DOI: 10.1039/c0jm03293j utilisation in a range of chemistry applications including organic, medicinal, polymer and materials chemistry.][26] However, to the best of our knowledge, this approach has not until now been used to surface-link CBPQT 4+ units to ITO based surfaces.
Here, we report on the use of ''click'' chemistry to anchor alkyne-functionalized CBPQT 4+ to plasmonic interfaces.We have subsequently exploited the presence of the CBPQT 4+ unit to specifically and reversibly modulate, under redox control, the plasmonic response using host-guest interactions with a TTF unit.8][29] The interface used in this study is based on a multilayer architecture (Fig. 1) developed in our group. 30It consists of gold nanostructures (Au NSs) deposited onto an indium tin oxide (ITO) substrate and coated with an ITO overlayer of different thicknesses.Indeed, one of the initial problems for accurate LSPR sensing was the poor adhesion of metallic nanostructures to the underlying surface, leading to morphological changes upon exposure to solvents and analytes. 31,324][35][36][37][38][39][40][41][42] In addition, it widens the surface functionalization schemes for the coupling of organic and/or biological molecules. 33,34,36,39) were cleaned with acetone and isopropanol, rinsed with Milli-Q water and dried under a nitrogen stream.A 2 nm thin gold film was deposited by thermal evaporation (0.1 A ˚sÀ1 ) at a base pressure of 10 À7 to 10 À8 Torr using a MECA 2000-1 system (Plassys) with about 10% accuracy.Post-deposition annealing of the gold film was performed under nitrogen stream at 500 C for 1 min using a rapid thermal annealer (Jipelec Jet First 100).

Experimental section
2.2.2.Deposition of ITO overlayers.ITO overcoatings were deposited on the ITO/Au NSs interface using r.f.sputtering (Plassys MP 450S) at 8.10 À8 mbar (turbomolecular rotary pump system). 30,43,44The deposition chamber contains an In 2 O 3 -SnO 2 (In 2 O 3 90% w/w); SnO 2 10% w/w 99.999% purity) ceramic sputtering target (75 mm in diameter).The deposition temperature is measured with a thermocouple set behind the sample holder.ITO deposition is carried out at a r.f.power of 13.56 MHz under oxygen/argon atmosphere using the following parameters: r.f.power ¼ 38 W, total pressure ¼ 0.012 mbar, O 2 / Ar ratio ¼ 0.051, deposition rate ¼ 0.6 nm min À1 , and substrate temperature ¼ 25 C.The gold nanorod is characterized by the height h and the width l.The lattice parameter ''a'' is defined as the distance between two nearest neighboring gold nanorods.The input source is placed in the substrate and the detector in air or solvent.times), ethanol (5 min, 2 times) and finally with water, and dried under a nitrogen stream. 303.3.''Clicking'' alkyne-functionalized cyclophane (1) to azide-terminated surface.The azide-terminated ITO/Au NSs/ITO surface was immersed in 10 mL of a solution of acetonitrile with alkyne-functionalized cyclophane (2 mM), CuI (2 mM) and DBU (0.1 M) and kept for 48 h at 70 C.The resulting surface was washed with acetonitrile, ethanol and water, and dried under a stream of nitrogen.30 2.4.Instrumentation 2.4.1.X-Ray photoelectron spectroscopy.X-Ray photoelectron spectroscopy (XPS) measurements were performed with an ESCALAB 220 XL spectrometer from vacuum Generators.A monochromatic Al Ka X-ray source (1486.6 eV) was operated in the CAE (constant analyzer energy) mode (CAE ¼ 100 eV for survey spectra and CAE ¼ 40 eV for high-resolution spectra), using the electromagnetic lens mode.No flood gun source was needed due to the conducting character of the substrates.The angle between the incident X-rays and the analyzer is 58 .The detection angle of the photoelectrons is 90 , as referenced to the sample surface.

Contact angle measurements.
Water contact angles were measured using deionized water.We used a remotecomputer controlled goniometer system (DIGIDROP by GBX, France) for measuring the contact angles.The accuracy is AE2 .All measurements were made in ambient atmosphere at room temperature.
2.4.3.Electrochemical measurements.Cyclic voltammetry (CV) was performed with an Autolab potentiostat 30 (Eco-Chemie, Utrecht, The Netherlands).A platinum mesh and a silver wire were used as a counter and a reference electrode, respectively.The scan rate was 50 mV s À1 and the active surface area was 0.04 cm 2 .
2.4.4.UV/vis spectrometry.Absorption spectra were recorded using a PerkinElmer Lambda UV/Vis 950 spectrophotometer in polystyrene cuvettes with an optical path of 10 mm.The wavelength range was 400-800 nm.
2.4.5.Simulation method.Calculations are performed using a Finite Difference Time Domain (FDTD) method, which solves the Maxwell's equations by discretizing both time and space and by replacing derivatives with finite differences. 46,47Our calculation is performed in a two-dimensional (2-D) box (along x and y axes, see Fig. 1) with a propagation along the y axis.Perfect Matching Layer (PML) conditions are applied at the boundaries y of the box, in order to avoid reflections of outgoing waves. 48long the x direction, the unit cell is repeated periodically and the structure is supposed to be infinite along the z direction.Space is discretized in both x and y directions using a mesh interval equal to Dx ¼ Dy ¼ 1 nm.The equations of motion are solved with a time integration step Dt ¼ Dx/(4c), where c is the velocity of the light in vacuum and the number of time steps equals to 2, 20 which is the necessary tested time for a good convergence of the numerical calculation.
The incoming pulse, having TM polarization, is generated at the bottom part of the unit cell by a current source parallel to the x axis and having a planar profile along the x direction.The current is generated during a short period of time in such a way as to excite the electromagnetic waves in the frequency domain of interest.The transmitted signal, probed at the end of the upper part of the unit cell, is recorded as a function of time and finally Fourier transformed to obtain the transmission coefficient versus frequency.

Theoretical considerations
To use the optimal LSPR interface in terms of refractive index sensitivity, theoretical calculations were performed on ITO/Au NSs interfaces coated with differently thick ITO overlayers.The Lorentz and Drude model was used to investigate the optical behavior of ITO/Au NSs/ITO substrates based on the interface configuration in Fig. 1. 35 It consists of a layer of gold nanostructures of diameter l, height h and interparticle distance a deposited on ITO (refractive index n 1 ¼ 2.00), coated with a dielectric of the same refractive index n 2 ¼ 2.00.This interface is furthermore covered by a non-absorbing medium of refractive index n 3 such as air, water, 1,3-propanediol and carbon disulfide.For the calculations, the following geometrical parameters were chosen: l ¼ 25 nm, h ¼ 15 nm, a ¼ 70 nm.This corresponds to an average metal coating of 37%.Fig. 2A shows the theoretical change in l max as a function of the thickness of the ITO coating layer when immersed in solvents of increasing refractive index.At short distances from the nanoparticles surface, the LSPR shift follows a steep slope with a change in l max by about 60 nm.As the distance from the nanoparticles increases, the curve bends over and shows a subsequent blue shift, followed by another sharp red-shift.The periodicity of the oscillation can be calculated by the equation: The periodicity of the oscillation is independent of the solvent used (i.e., independent of the refractive index n 3 ).However, the amplitude of the oscillation is strongly affected by the value of the refractive index n 3 of the solvent.
Fig. 2B shows in the form of a bare diagram the evolution of the sensitivity S in nm RIU À1 (change of nanometre per refractive index unit) as a function of the ITO layer thickness.The sensitivity decreases with the deposition of ITO overlayers.For d ITO # 20 nm, the calculated sensitivities are comparable to uncoated ITO/Au NSs interfaces.Negative sensitivities are observed for interfaces coated with 40-80 nm and 180-240 nm.We observed that the sensitivity becomes zero for d ITO z 90, 165 and 250 nm.The sensitivities of the ITO/Au NSs/ITO interface architectures were compared with results on glass/Au NSs/glass and glass/Au NSs/ITO interfaces as reported previously. 49Generally, it can be observed that the sensitivities on ITO/Au NSs are lower than that obtained on glass/Au NSs for thin ITO overcoatings (i.e., d ITO # 20 nm).However, for d ITO > 30 nm, the sensitivity is more pronounced on ITO/Au NSs than on glass/Au NSs.One can also notice that large negative sensitivities are observed with ITO overcoatings as thick as 240 nm.This is rather different to the other cases and shows the complex interplay between the refractive indexes of the overcoating.However, generally the sensitivities obtained are of the same magnitude.In the following, two different ITO overlayer thicknesses were used: 20 and 240 nm.

Optical characterization of the ITO/Au NSs/ITO interfaces
Fig. 3 shows the experimentally obtained UV/vis spectra of ITO/ Au NSs and ITO/Au NSs/ITO interfaces in air.The experimentally obtained l max correlates well with theoretically obtained values (Fig. 2).An additional optical feature next to the main plasmonic peak is seen between 400 and 550 nm, which originates most likely from classical Fabry-Perot cavities due to uncompensated thin film interferences.Indeed, the position of these bands does not change when immersed into different solvents.This behavior is different from that of the main plasmonic peak and the detected maximum wavelength depends on the incident angle of light.The main difference in the LSPR band of the ITO/Au NSs interfaces coated with 20 or 240 nm thick ITO layers is the peak full width at half maximum (fwhm).For an uncoated ITO/Au NSs interface, the fwhm is $80 AE 2 nm, while its value increased to $90 AE 4 nm for d ITO ¼ 20 and $134 AE 4nm for d ITO ¼ 240 nm.

Linking alkyne-functionalized cylcophane to azideterminated ITO surface
Fig. 4 illustrates the surface functionalization strategy employed.A sufficient amount of surface hydroxyl groups was generated on the ITO/Au NSs/ITO LSPR platform by UV/ozone treatment for 10 min. 50Amine termination was obtained by silanization of the surface hydroxyl terminal groups with APTMS.Chemical coupling of 4-azidobenzoic acid to the amine-terminated ITO surface resulted in the formation of an azide termination.Finally, the 1,3-dipolar cycloaddition between the azide-terminated surface and 1-alkynyl-cyclophane in the presence of Cu(I) was used to incorporate cyclophane moieties on the ITO/Au NSs/ITO substrate.
Water contact angle measurements and X-ray photoelectron spectroscopy (XPS) were used to characterize the modified interfaces.The initial ITO/Au NSs/ITO surface exhibits a water contact angle q ¼ 30 AE 2 .This value decreased to q ¼ 10 AE 2 after photochemical oxidation.Silanization with APTMS resulted in a water contact angle q ¼ 40 AE 2 .Incorporation of the azide function led to an increase of the water contact angle to q ¼ 63 AE 2 .After clicking cyclophane groups to the azide terminal groups, the contact angle dropped to q ¼ 55 AE 2 .
X-Ray photoelectron spectroscopy was used in parallel to evaluate the changes in the surface chemical composition during surface derivatization.Fig. 5 displays the XPS survey spectra of ITO/Au NSs/ITO surfaces before and after silanization with APTMS, coupling of 4-azidobenzoic acid, and clicking cyclophane moieties to the surface.The initial interface shows peaks due to indium at $17 eV (In 4d), 150 eV (In 4s), 444 eV (In 3d), 665 (In 3p), 703 eV (In 3s), a small contribution due to doping with Sn at 493 eV (Sn 3d) and a peak at 532 eV due to O 1s.An additional band at 285 eV due to C 1s from surface contamination is also observed (Fig. 5a).After silanization of the terminal hydroxyl groups with APTMS, a new peak at $400 eV due to N 1s (-NH 2 ) appeared in the XPS survey spectrum (Fig. 5b).From the high resolution XPS of the N 1s, the peak can be deconvoluted into two bands at 399.2 eV due to free amine (-NH 2 ) and  a smaller contribution at 400.6 due to protonated amine groups (-NH 3 + ) (Fig. 6a). 51Covalent coupling of the terminal amine groups with azidobenzoic acid increased the overall nitrogen, carbon and oxygen contents (Fig. 5c).The broad N 1s signal in the high resolution scan was fitted and deconvoluted into three peaks: 400.4 eV (N]N]N), 401.2 eV(-HN-C]O) and 402.6 eV (N]N]N) with a ratio of 2.7 : 1.3 : 1 close to the expected ratio of 2 : 1 : 1 (Fig. 6b).
Upon clicking of 1-alkynyl cyclophane to the azide-terminated interface, the high-resolution N 1s shows a broad signal centered at 401.7 eV, which was fitted and deconvoluted into four peaks: 398.8 eV (N-N]N), 400.3 eV (N]N) and 400.8 eV (-HN-C]O) and 401.7 eV (]N + -) with a ratio of 1 : 1.5 : 1.7 : 4 (Fig. 6c).This is consistent with the formation of surfaceconfined triazole groups and incorporation of the cyclophane ring. 52

Cyclic voltammetry using Fe(CN) 6
4À as a solution-based redox couple was first performed to evaluate the conductive properties of the modified ITO/Au NSs/ITO interfaces.The ITO/Au NSs/ ITO hybrid interfaces show different behaviors, depending on the ITO overlayer thickness (Fig. 7A).Increasing the ITO overlayer thickness led to an increase of the detected current and electron transfer rate for the oxidation of Fe(CN) 6

4À
. Indeed, the 20 nm thick ITO overlayer shows an enlarged DE p ¼ 200 AE 15 mV.This is rather typical for a rough electrode material due to the presence of uncompensated resistance within the ITO overlayer. 53In the case of a 240 nm thick ITO overlayer, the i-E curve   is comparable to that of a naked massive bare ITO, where a peak to peak separation DE p ¼ 180 AE 15 mV was determined with an apparent electron transfer rate constant k app z 0.003 cm s À1 (see Fig. S1a, ESI †) Similar behavior was observed when the CBPQT 4+ -modified LPSR interfaces were immersed into acetonitrile/TBAPF 6 (0.1 M) (Fig. 7B).In the case of a 240 nm thick ITO, the CV of the surface-bound CBPQT 4+ mirrors that obtained for CBPQT 4+ on ITO in solution (see ESI, Fig. S1b †), indicating that its redox properties are not significantly altered by its immobilization onto ITO. 21Indeed, in the case of a 240 nm thick ITO overlayer, the CV gives rise to two-electron quasi-reversible reduction waves at E 01 ¼À 0.32 V and E 02 ¼À 0.76 V (Ag/AgCl) corresponding to the two step reduction (CBPQT 4+ / CBPQT 2+ c / CBPQT 0 )of the cyclophane unit. 21,16In the case of a 20 nm thick ITO overlayer, the CV also exhibits two redox waves.However, the redox waves are shifted by about 90 mV to more cathodic potentials and the second redox wave is irreversible.Indeed, we have shown previously that a minimal thickness d (ITO) z 60 nm is required to provide a good electrical contact. 30The irreversible redox behaviour observed for the reduction of CBPQT 2+ c / CBPQT 0 may be due to the high resistivity of the ITO film.
The CBPQT 4+ surface coverage G can be estimated by integrating the cathodic peak area, for the first oxidation wave, according to G ¼ Q a /nFA, where F is the Faraday constant, n the number or electrons exchanged (n ¼ 2) and A the surface area (A ¼ 0.04 cm 2 ).A surface coverage G ¼ (1.64 AE 0.9)Â10 13 molecules cm À2 was obtained.This is lower than the surface coverage obtained for ferrocene molecules grafted onto ITO using a similar click chemistry approach (G ¼ (5.16 AE 0.9)Â10 14 molecules cm À2 ). 30This might be due to the bulky nature of CBPQT 4+ and the likely associated Coulombic repulsion between the cationic CBPQT 4+ units attached to the surface.
Finally, UV/vis spectra of the CBPQT 4+ -modified ITO/Au NSs/ITO interfaces were recorded.Fig. 8 shows the absorption spectra of ITO/Au NSs/ITO interfaces coated with 20 and 240 nm ITO layers before and after grafting CBPQT 4+ units.The plasmonic signal changes when CBPQT 4+ is present on the interface showing in both cases a blue shift together with a decrease in the absorption intensity.The origin of the blue shift is currently not clear.However, it has been shown previously by us, that when ITO/Au NSs interfaces are coated with overlayers such as SiO x having a refractive index of 1.48 (similar to that of an organic layer), a blue shift was observed. 42The 20 nm ITO overlayer shows better plasmonic signals and was used in the following experiments to demonstrate the formation of hostguest complexes with TTF.

Formation of guest-host complexes
TTF has proven to be a versatile redox-active system for the construction of switchable host-guest and interlocked systems. 135][56][57] Here, we have exploited the faculty of TTF to form redox tunable complexes with CBPQT 4+ as a means to specifically and reversibly modify a surface via complexation with surface-confined CBPQT 4+ units.Fig. 9B shows the UV/vis spectra of the CBPQT 4+ modified ITO/Au NSs/ITO (20 nm) interface before and after immersion of the interface in a TTF solution for 20 min.The addition of the TTF results in the appearance of a characteristic absorption band consistent with the formation of a TTF-CBPQT 4+ host-guest complex (see Fig. S2 in the ESI † for the UV/vis spectra of CBPQT 4+ and TTF-CBPQT 4+ host-guest complex recorded in acetonitrile solution).In particular, upon the addition of TTF, a broad absorption band centered around l ¼ 875 nm is observed, which is a diagnostic of the formation of a TTF-CBPQT 4+ pseudorotaxane-like system. 54In addition, a significant red shift of z22 nm in the LSPR peak position is observed.This observation correlates with the increase in the refractive index of the surrounding area. 56To further prove the ability of the new interface containing surfacelinked CBPQT 4+ to bind TTF, the CV of the CBPQT 4+ modified ITO/Au NSs/ITO (20 nm) upon the addition of TTF was recorded (Fig. S3 in the ESI †).The CV indicates that the first reduction wave of the CBPQT 4+ unit is shifted by about À20 mV upon addition of TTF, presumably due to donor-acceptor interactions resulting from complex formation.Similar small shifts were reported by Stoddart et al. 5 We next turned our attention to whether the surface-bound complexes could be disrupted by oxidizing the TTF unit with Fe(ClO 4 ) 3 . 58,59Chemical oxidation of TTF was carried out by immersion of the TTF-complexed interface into Fe(ClO 4 ) 3 (1 mM) in acetonitrile for 20 min.The recorded UV/Vis spectrum shows a blue shift of the LSPR peak, due to the decrease in the refractive index of the surrounding area. 56Indeed, the initial LSPR position is obtained, thereby suggesting the dethreading of the TTF unit from the CBPQT 4+ unit, presumably due to electrostatic repulsion of the CBPQT 4+ ring and the oxidized TTF unit.In addition, no signal due to the host-guest complex at l ¼ 856 nm is observed, which is a good indication that TTF unit is no longer bound to the cyclophane and that the plasmonic properties of the ITO/Au NSs/ITO interface can be modified by complexation of the CBPQT 4+ species.Indeed, formation and disruption of the TTF-CBPQT 4+ complex using Fe(ClO 4 ) 3 can be performed several times without degradation of the LSPR signal and significant change in the LSPR shift (Fig. 9C).

Conclusion
The use of click chemistry to covalently link alkyne-functionalized p-electron deficient CBPQT 4+ to ITO/Au NSs/ITO interfaces has been demonstrated.The presence of surface-linked CBPQT 4+ was shown by XPS, CV and UV/vis measurements.The interfaces were used to demonstrate the possibility of recording the formation and dethreading of guest-host complexes using the LSPR.The capture and release of guest molecules can potentially be used to program LSPR interfaces to selectively capture and release specific species (e.g.TTF functionalized biomolecules) from a complex mixture.The applications of such ''tunable'' plasmonic interfaces are thus broad.In addition, besides LSPR other sensing techniques such as Quartz microbalance (QCM) transducers 60 will make possible such type of investigations enlarging the scope of the developed surface chemistry.

Fig. 1
Fig. 1 Schematic representation of the structure studied in this work.The gold nanorod is characterized by the height h and the width l.The lattice parameter ''a'' is defined as the distance between two nearest neighboring gold nanorods.The input source is placed in the substrate and the detector in air or solvent.

Fig. 2 (
Fig. 2 (A) Theoretically calculated evolution of l max of an ITO/Au NSs LSPR interface coated with increasingly thick ITO overlayers: air (filled black circles) water (filled grey circles), 1,3-propanediol (open black squares), and CS 2 (open grey squares) using a ¼ 70 nm, l ¼ 25 nm, h ¼ 15 nm.(B) Change of refractive index sensitivity for different LSPR interfaces as a function of the top layer thickness: ITO/Au NSs/ITO (black), glass/Au NSs/glass (grey) and glass/Au NSs/ITO (bright grey).