Study of the electrochemical and optical properties of fullerene and methano[60]fullerenediphosphonate derivatives in solution and as self-assembled structures

Abstract

Four methanofullerenes with phosphonate groups attached to a C₆₀ core were synthesized to probe their electrochemical and optical properties both in solution and as self-assembled monolayer structures (SAMs). As the methano[60]fullerenediphosphonate is a water soluble derivative of the fullerene it has high potential as an imaging molecule in biological applications and in optoelectronics. For the processing of fullerene based SAMs different electrode substrates (ITO, Au and Si) with specific anchoring groups (zirconium, cysteamine and amino-silane) were used. The formation of the C₆₀-SAMs to the surfaces were investigated by atomic force microscopy (AFM), infrared spectroscopy, contact angle measurements and cyclic voltammetry. Using cyclic voltammetry it was shown that the reduction potentials of substituted methanofullerenes, both in solution and as SAMs, were slightly higher as compared with formal potentials of the redox reactions of C₆₀. The AFM results show that the fullerene molecules produce surface features with an apparent height of ∼2 nm. The self-assembly strategy aims towards fabrication of electronic devices with improved interfacial contact, a prerequisite in order to obtain enhanced electron transfer between acceptor–donor materials.

---

1 Introduction

Since the discovery of fullerenes (C₆₀)¹ and mainly after the development of methods for production on a macroscopic scale² these carbon clusters have led to tremendous scientific interest. Electrochemical methods have played a significant role in investigation of the properties of these clusters as these methods provide direct information on the thermodynamics and kinetics of the electron transfer processes. The high electron affinity of C₆₀ is nowadays well established from cyclic voltammetric studies of dissolved molecules. The ability of fullerene to reversibly accept up to six electrons (electrochemically detectable), generating stable multianions, is the most attractive property for developing electronic devices. Of particular interest is the C₆₀ moieties that serve as excellent electron acceptors in molecularly designed donor–acceptor systems. Most of the fullerene-based applications require the immobilization of the carbon cage onto solid surfaces as thin films. Understanding the chemistry, characteristics and to structurally control these interface systems will be the key factor for the future realization of many nanoscale applications. The abundance of studies on attaching fullerenes to different substrates have focused on applying different techniques including spin coating, solution evaporation, the Langmuir–Blodgett (LB) technique, interfacial hydrogen bonding, electrochemical deposition and self-assembled monolayers (SAMs) to reproducibly build well-defined, stable interfaces. The self-assembly approach is by far the most powerful methodology for organizing ordered and well-defined interfaces on a nanometer scale containing fullerenes on solid surfaces using simple solution-based procedures.^3–5 Different immobilization techniques have been developed towards fullerene self-assembly onto surfaces. Either a fullerene can be modified with functionalities that allow for self-assembly,^6–8 or a surface may be chemically modified with a reagent that undergoes a bond forming reaction with the fullerene species as amines that are often used in the case of silicon substrates.⁹ In the field of fullerene research much attention has therefore been devoted to the synthesis and investigation of various properties of functionally substituted fullerene derivatives. The development of new fullerene based materials is not only important for covalently attaching fullerenes to a solid support, but also for tailoring the fullerene to exhibit a desired property as the nature, geometry, structure and number of addends will all influence on the sphere's electronic or optical response and on its solubility. Functionalization of C₆₀ with hydrophilic diphosphonate groups, especially in the case of bis- and higher adducts, gives the possibility to synthesize water soluble fullerene derivatives.¹⁰ Introducing electron-donating substituents into the fullerene molecule makes the affinity to electron decrease and the reduction potential shift in the negative direction. The same occurs when increasing the number of such substituents.^11,12

The successful formation of fullerene-modified SAMs has been reported using a variety of microscopic methods including transmission electron microscopy (TEM),¹³ atomic force microscopy (AFM)³ and scanning tunneling microscopy (STM).⁵ Also spectroscopic techniques such as attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR), UV-Vis spectroscopy and X-ray photoelectron spectroscopy (XPS) or electroanalytical techniques such as cyclic voltammetry (CV)^4,14,15 has been used. However, only a limited number of studies attempted to combine microscopic and spectroscopic methods.

Herein we report the electrochemical characterization of fullerene and four methanofullerene derivatives (structures 1, 2, 3 and 4 in Scheme 1) with phosphonate groups attached to a C₆₀ core, in solutions and as self-assembled structures. The information on the reduction potentials is of interest from the viewpoint of practical applications being based on the electron-accepting properties of the materials. The fullerene functionalizations were designed aiming for water solubility as well as self-assembly properties of the compounds. Derivatized fullerenes and metallofullerene-based materials with their small size and big surface area offer a promise as new diagnostic tools in biomedicine.¹⁶ For the self-assembly various substrates, such as indium-tin-oxide (ITO), gold and Si has been used. We focus on different chemistries to attach fullerene molecules to the substrates using amino-terminated SAMs for Si and gold. It has been reported that the fullerene molecules react readily with the amine group and can thus be used as a reliable probe for surface reactivity.^4,5 In the case of metal oxides phosphonic groups form SAM without the need of surface anchoring groups. In order to improve the electrochemical stability of SAMs on ITO we used the metal phosphonate approach utilizing zirconium. The zirconium phosphonate chemistry has been used in order to form mechanically and chemically stable and well-ordered LB films.^17,18 The functionalized surfaces were characterized by cyclic voltammetry, FTIR spectroscopy and contact angle measurements, whereas the surface topography was examined by AFM and scanning electron microscopy (SEM).

Scheme 1 Chemical structure of fullerene derivatives (1–4).

2 Experimental

2.1 Materials

Zirconium oxychloride (ZrOCl₂·8H₂O, Sigma-Aldrich >99.99%), 3-aminopropyl triethoxysilane (APTES, Fluka, >98%), bromotrimethylsilane (Alfa Aesar, >97%) and cysteamine (Sigma-Aldrich, >98%) were used as received. Tetrabutylammonium hexafluorophosphate (TBAPF₆, Sigma-Aldrich > 99%) and tetrabutylammonium tetrafluoroborate (TBABF₄, Sigma-Aldrich 99%) were dried at 120 °C for 2 h prior to use. Dimethylformamide (DMF, VWR International) and 1,2-dichlorobenzene (Acros Organics, 99%) were dried over basic alumina (Aldrich). Fullerene (C₆₀, SES Corporation, 99.5%) was used without any purification. Fullerene derivatives (1–4) were synthesized similar to the procedure described elsewhere¹⁰ and are explained in Supplementary Information. Bis(diisopropoxyphosphoryl)-bromomethane and bis(diethoxyphosphoryl)-bromomethane were prepared as previously described.¹⁹

2.2 Methods

The Au and Pt electrode (1 mm diameter, Cypress Systems Inc.) was polished mechanically with DP-Paste (Struers A/s) of different sizes (15, 7, 3 and 0.25 μm) before each experiment and thoroughly rinsed. For the Au electrode the procedure was completed by cycling the electrode between −0.2 and +1.15 V vs. Ag/AgCl at 0.1 V s⁻¹ potential scan rate, in 0.5 M H₂SO₄ solution, until a repeatable cyclic voltammogram, typical of a clean Au electrode, was obtained.

Si(100) wafers (Okmetic, Finland) 12 mm × 12 mm were cleaned in a fresh piranha solution (concentrated H₂SO₄–30% H₂O₂ in proportion 3 : 1), rinsed thoroughly with quartz distilled water and dried. Double quartz distilled water was used in all aqueous experiments. (Warning: piranha solution is extremely corrosive and must be treated with extreme caution. It should not be stored in tightly closed vessels.) The piranha cleaned silicon substrates were silanized with (3-aminopropyl)triethoxysilane in toluene solution (1% v/v) for 4 min at 60 °C.²⁰ After silanization the wafers were rinsed thoroughly with toluene solution. The silanized silicon wafers were thereafter placed in a 1 mM toluene solution of C₆₀ for 24 h (preparation steps shown in Scheme S1, ESI†). The used fullerene solution was filtered through a 0.45 μm PTFE filter. Then the surface was rinsed thoroughly with toluene to remove physisorbed C₆₀ and thereafter dried before characterization. The thickness of the natural oxide layer on cleaned Si wafers was measured by ellipsometry to be around 2 nm and showed a contact angle of 34°. In the text these Si oxide covered substrates are referred to and marked as Si. After silanization a thickness of 3.5–4 nm was obtained by ellipsometry and the contact angle was 75°.

On some of the piranha cleaned Si wafers a thin gold layer (ca. 150 nm) was evaporated using the Edwards E306A coating system. Before use these substrates were additionally cleaned with O₂ plasma using the Harrick PDC-3XG plasma cleaner. The plasma cleaned Au surface was immediately soaked in a 10 mM toluene solution of cysteamine for 1 h. After that the surface was rinsed with toluene, ethanol and water for 30 min each and dried. The cysteamine covered Au surface was then immediately placed in a 1 mM toluene solution of C₆₀ for 24 h (preparation steps shown in Scheme S2, ESI†). Then the surface was rinsed with toluene and dried.

ITO electrodes (Delta Technologies) were cleaned for 10 min in successive ultrasonic baths in acetone, ethanol and water. Some of the cleaned ITO electrodes were further treated in the solution mixture of NH₃–H₂O₂ (3 : 1) for one hour, rinsed thoroughly with water and dried. Before use these substrates were additionally cleaned with O₂ plasma using a Harrick PDC-3XG plasma cleaner. Some of the OH-treated ITO wafers were further placed in a 10 mM solution of ZrOCl₂·8H₂O for one hour and thoroughly rinsed and dried. Treated wafers were then immersed in a toluene solution of fullerene derivatives 2 and 4 (1 mM) for 24 h (preparation steps shown in Scheme S3, ESI†). The substrates were taken from the solution and rinsed with toluene to remove excess molecules from the surface and finally dried. For the ITO electrodes also another OH-treatment was used. In this case the cleaned wafers were placed in an H₂O₂–NH₄OH–H₂O (1 : 1 : 5) solution and heated to 80 °C. After 30 min the wafers were rinsed with water.

2.3 Equipment and measurements

2.3.1 Characterization of fullerene and fullerene derivatives in solution. All electrochemical experiments were performed in a conventional three-electrode one-compartment cell with an Autolab (PGSTAT101) potentiostat. The electrochemical cell consisted of an Au-working electrode, a Pt-wire as counter electrode and an Ag/AgCl wire was used as quasi-reference electrode. All potential values given are referred to this reference electrode, calibrated versus the ferrocene/ferrocinium couple giving E_redox of 490 mV for the solvent mixture containing 0.1 M TBABF₄. The electrochemical reduction of fullerene and fullerene derivatives (1–4) were carried out in 1,2-dichlorobenzene–DMF mixture (3 : 1) containing 0.1 M TBABF₄ as electrolyte salt in the potential range between 0 and −2.0 V using 50 mV s⁻¹ as scan rate. The concentration of fullerene and fullerene derivatives was 1 mM. Prior to recording cyclic voltammograms the solutions were deaerated with nitrogen for 15 min and a slight nitrogen overpressure was maintained during the reaction.

The UV-Vis spectra were recorded through a 1 cm path length quartz cuvette using a Hewlett Packard 8453 spectrophotometer in the range 300–800 nm. Measurements were performed at different concentrations of the compounds in the same solvent mixture used in the electrochemical measurements and in toluene.

The luminescence measurements were performed with a Varian Cary Eclipse spectrophotometer (Varian Scientific Instruments, Mulgrave, Australia).

The fullerene and fullerene derivatives (2 mg) were grinded with spectral grade KBr (200 mg) and pellets were made using standard procedure in hydraulic pellet press (GEC machines Ltd.). These pellets were subjected to FTIR spectroscopic analysis recorded on dry-air-purged Nexus 870 FTIR spectrometer (Nicolet) equipped with a DTGS detector. The IR spectra were acquired in the transmission mode in the range 4000–500 cm⁻¹. For each spectrum 128 scans at 4 cm⁻¹ spectral resolution were collected.

The ¹H and ¹³C NMR spectra were recorded with a Bruker AVANCE 600 spectrometer equipped with a direct X-detection probe (BBO-5mm-Zgrad). The NMR spectra were measured from a CDCl₃ solution at a temperature of 298 K (uncalibrated) with tetramethylsilane (TMS) as an internal reference (δ_TMS = 0.00 ppm for both nuclei). The ¹H NMR spectrum was acquired with a simple pulse-acquire sequence (Bruker pulse program ‘zg30’ using a 30° flip angle and 4.4 s pulse repetition time) and the ¹³C NMR spectrum was acquired with broadband ¹H decoupling (Bruker pulse program ‘zgpg30’ using a 30° flip angle, 8.9 s pulse repetition time, and waltz16 decoupling sequence).

2.3.2 Layer characterization. For the electrochemical characterization of fullerene modified surfaces the electrochemical setup described above was used, applying the substrate with SAM as working electrode.

External reflection infrared measurements were done from SAMs on Si substrates using Nicolet Nexus 870 FTIR spectrometer equipped with a Harrick Seagull variable angle reflection accessory and liquid nitrogen cooled mercury-cadmium-telluride (MCT-A) detector. Spectra were measured at 65, 70, 75 and 80° angles of incidence relative to the surface normal and for each spectrum 1024 scans at spectral resolution of 4 cm⁻¹ were recorded. A piranha cleaned wafer was used as reference spectra.

The surface morphology was investigated by means of Leo (Zeiss) 1530 Gemini FEG scanning electron microscopy (SEM) and atomic force microscopy (AFM) on a diCaliber AFM microscope from VEECO Instruments Inc. Analysis was carried out in tapping mode. A NSG-11 tip from NT-MDT with a 11.5 N m⁻¹ force constant was used. The raw AFM data obtained were processed by flattening and plane fitting with the WSxM 5.0 data visualization and analysis software.

The film buildup was followed on a SE 400 ellipsometer (Sentech Instruments GmbH) operating at 632.8 nm using the reflection angle of 70°. The refractive index of 1.70 was used for SiO. At least three different sampling points were made in order to get the average thickness value.

Water contact angles were measured on a CAM 200 contact angle measurement system (KSV Instruments).

3 Results and discussion

3.1 Electrochemical reduction of phosphonium substituted fullerenes (1–4)

Due to earlier studies it is well known that fullerenes and its derivatives have the ability to readily accept electrons in both electrochemical reactions and upon interaction with organic electron donors.^21,22 In this work the electrochemical behavior of four different fullerene derivatives has been investigated by cyclic voltammetry in dichlorobenzene–DMF solution mixture (3 : 1) and compared to that of fullerene, as shown in Fig. 1. The fullerene derivatives have been synthesized in the group and are marked in Scheme 1 with 1–4. In the potential range between 0 V and −1.8 V three reduction peaks and correspondingly, three oxidation peaks are observed in the CV of C₆₀ and the methanofullerenes 1–2 using a Pt electrode in the aforementioned solution mixture containing 0.1 M TBABF₄. The cathodic and anodic potentials vs. Ag/AgCl for the current peaks of the compounds are summarized in Table 1. It is well known that addition of electron withdrawing substituents to the fullerene core leads to loss in conjugation and a LUMO that is higher in energy, and thus to a decreased electron affinity for derivatized fullerenes. As can be seen from Table 1 and Fig. 1 the potentials, especially for the first two reduction peaks of derivatives, are shifted to negative values as compared to formal potentials of the corresponding redox-transitions of C₆₀. These displacements are in the range 90–380 mV for derivatives 2 and 4 whereas it is 190–520 mV for derivatives 1 and 3 (monomethanofullerene 190–320 mV versus tetraphosphonobismethanofullerene 340–520 mV). The shift is dependent on the amount of substituents showing the biggest displacement for bismethanofullerenes. The third reduction peak is shifted approximately 100 mV for all studied fullerene derivatives. In Fig. 1 it can be seen that with increasing degree of functionalization of the fullerene cage reductions become more difficult and irreversible. This can be explained in terms of a stepwise loss of conjugation, which shifts the LUMO to higher energy and so decreasing the acceptor property. Several cases have been reported in which derivatization, regardless of nature or number of addends, decreases the electron affinity of the C₆₀ sphere. Typically, cathodically shifted waves, with shifts ranging from 30 to 350 mV per adduct with respect to those of pure C₆₀, have been observed by CV.^23–25 The stability of anionic intermediates that form during reduction of fullerene and their derivatives is defined by the anion charge, the nature of fullerene, environment and substituents and the number of substituents. In general, the anion stability drops with increasing charge. Upon going from lower to higher fullerenes and with decreasing number of substituents, such species turn more stable. Reduction of some of these functionalized derivatives has resulted in the release of the attached fragments and recovery of pure C₆₀. As an example; C₆₀H₂ converts to C₆₀ but not until after the third reduction wave is reached. Furthermore it is solvent and temperature dependent.²³ Not just removal of the addend but also electropolymerization can occur as a result of electroreduction.^26,27 For methanofullerenes retrocyclopropagation have been studied and found to take place at potentials of the second reduction peak.^27,28 After this process the third and fourth reduction peak and also the potentials for the first, second and third oxidation peak coincide with the corresponding peak potentials for the fullerene. Detailed studies by cyclic voltammetry has been made and revealed understanding of the mechanism of methano fragment elimination as well as enabled determination of rate constants of this process. In our work the potentials of all three oxidation peaks are shifted compared to those of unsubstituted fullerene with 20–360 mV for derivatives 2 and 4 compared to 70–220 mV for derivatives 1 and 3. This indicates weakening of the acceptor properties of the fullerene fragment and moreover stability of the fullerene derivatives. However, for methanofullerenes it has been reported that bulk electrolysis, when performed after the second reduction wave, lowering of the scan rate and change of electrolyte solution will all have a big influence on the stability.^27,28

Fig. 1 CV for fullerene and fullerene derivatives 1–4 obtained at Pt-electrode in the 1,2-dichlorobenzene–DMF mixture (3 : 1) containing 0.1 M TBABF₄. Scan rate 50 mV s⁻¹.

Table 1 Characteristics of reduction steps for fullerene and derivatives 1–4 versus ferrocene [V]^a

	1. Red.	2. Red.	3. Red.	1. Ox.	2. Ox.	3. Ox.
a Electrochemical measurements were performed using a Pt-working electrode, an Ag/AgCl pseudoreference electrode and a Pt counter electrode. TBABF4 (0.1 M) was used as supporting electrolyte in solvent mixture (1,2-dichlorobenzene–DMF), concentration of analyte 1 mM and the scan rate was 50 mV s^−1.
C60	−0.44	−0.89	−1.40	−0.37	−0.81	−1.32
Deri 1	−0.63	−1.21	—	−0.56	−1.03	—
Deri 2	−0.57	−0.98	−1.51	−0.39	−0.82	−1.36
Deri 3	−0.96	−1.23	−1.47	−0.5	−0.88	−1.38
Deri 4	−0.81	−1.27	—	−0.57	−1.17	—

---

The peak currents for the reduction are characteristic for the compound nature, especially for the second reduction where higher currents than one-electron transfer processes would involve are seen. The structure specific influence on the height of the current peaks has been studied thoroughly for phosphorylated methanofullerenes.²⁸ Also in this work the number of electrons transferred at potentials of the second reduction step is higher compared to that of the fullerene, showing the biggest divergency for derivative 3.

3.2 Electronic absorption spectra of fullerene derivatives

Room temperature electronic absorption spectra recorded between 380 and 750 nm for the four C₆₀ derivatives and of fullerene in the solvent mixture 1,2-dichlorobenzene–DMF (3 : 1) are shown in Fig. 2a. Changes in the extension of the conjugated π-chromophore of the fullerene have a pronounced effect on the spectral shape and position of the absorption band. When comparing the spectra of C₆₀ with the ones obtained for derivatives 1–4, C₆₀ itself gives a purple solution thus exhibiting a broad absorption between 435 and 665 nm, with maxima at 500, 540, 570, 600 and 625 nm (seen more clearly in ESI Fig. S1,† fullerene UV-Vis spectra recorded at different concentrations). Also a sharp band at 406 nm can be observed in the visible spectral region for C₆₀ and for [6,5]-ring open methanofulleroids. These spectral features are absent for the fullerene derivatives. In contrast the fullerene derivatives show in the visible region other characteristic spectral features. UV-Vis spectra recorded in the solvent mixture and in toluene for fullerene derivatives 1 and 2 are shown in Fig. 2a and b, respectively. For the methanofullerene 1 and 2 the following three distinct features can be mentioned, a sharp band at 427 nm, a broad band peaking at 486 nm and a group of weak structures in the 650–700 nm region whose strongest feature is at 694 nm. These are characteristic for all C₆₀ mono-adducts.²⁹ The broad absorption band in the Vis range seen in the spectrum of C₆₀ is generally less structured and hypsochromically shifted in the spectra of mono-adducts, maximum observed around 500 nm. The sharp band at 427 nm and the wide absorption between 450 and 650 nm, with the maximum around 500 nm gives a reliable criterion of the formation of cycloadducts at the closed [6,6]-bond.³⁰ These three new bands, not present in C₆₀, are very similar to those reported for several other methanofullerenes.^29,31

Fig. 2 Absorption spectra between 380 and 750 nm of C₆₀ and fullerene derivatives 1–4 in (a) 1,2-dichlorobenzene–DMF (3 : 1) solvent mixture (10⁻⁴ M) and (b) in toluene (10⁻³ M).

The spectrum of bis-adducts 3 and 4, which gives orange-brown solutions, resembles in its general shape the spectrum of a mono-adduct, although the broad Vis absorption band displays a shift from 500 nm to 470 nm (UV-Vis spectra recorded at different concentrations of bis-adduct 4 shown in ESI, Fig. S2.† In Fig. S2† no shift is observed for the position of the peak maxima with concentration showing the absence of aggregation). In a characteristic manner, the diagnostic mono-adduct peak around 427 nm is no longer observed in the spectrum for the bis-adducts. The observed shift of the broad end absorption in the UV-Vis spectra could be nicely correlated with the first reduction potential obtained in the electrochemical measurements. The higher the energy of the optical end absorption the higher the reduction potential will be.

3.3 Photoluminescence measurements

The photoluminescence (PL) spectra of a water soluble polythiophene (PT) based polymer synthesized in the group³² (structure shown in Scheme S4, ESI†) and of the polymer blended with fullerene derivative 4 obtained in water solutions are shown in Fig. 3. PL spectra can give information about the photoexcited electron transfer and recombination processes and has been used to study the fate of electron–hole pairs successfully. The polymer shows an absorbance band maximum at 378 nm in the UV-Vis spectra and therefore it was used as the excitation wavelength for PL studies. From the PL spectra it can be seen that adding the fullerene derivative to PT, resulted in a significant quenching which indicates that charge transfer takes place between the two component materials. This indicates that the fullerene acceptor property has not been too much decreased by the derivatisation. More detailed measurements are currently under study using different fullerene derivatives as also PL measurements for thin film structures.

Fig. 3 PL spectra for (a) thiophene and (b) thiophene–fullerene derivative 4 in water solutions. Excitation wavelength of 380 nm was used.

3.4 FT-IR spectra of fullerene derivatives

The FT-IR spectra of C₆₀ and fullerene derivatives 3 and 4 as KBr pellets in the wavenumber range 3700–470 and 1470–470 cm⁻¹ are shown in Fig. 4a and b, respectively. The characteristic absorptions of the C₆₀ skeleton appear at 1430, 1183, 576 and 528 cm⁻¹.³³ The most prominent of the four absorptions seen in the spectrum of C₆₀, at 528 cm⁻¹ is maintained and appears as a strong band in the narrow range between 524 and 531 cm⁻¹. This vibration in C₆₀ has been assigned to a pure radial breathing mode. This breathing mode of the C-sphere seems to be quite insensitive to the degree, pattern and nature of derivatization. Fullerene derivative 4 shows a broad OH stretching band from O=P–OH at around 2700–2000 cm⁻¹. The P–OH stretch shows maxima at 2725–2525, 2350–2080 and 1740–1600 cm⁻¹ from which the one at 1740–1600 cm⁻¹ is not present due to the two OH groups attached to phosphorus. For derivative 3 a sharp and strong peak is observed at 1262 and at 1106, 1005 and 606 cm⁻¹, corresponding to the vibrational frequencies of P=O and P–O, respectively. For derivative 4 the peak from the P=O vibration is much broader and the peak maxima observed at lower wavenumbers due to the presence of the OH groups. For derivative 3 the P–O–C stretch, observed at 1088–920 cm⁻¹, overlap with the vibrational frequencies of P–O. Presence of the ethyl groups for derivative 3 can be confirmed due to the strong peaks observed at 2980, 2933 and 2873 cm⁻¹, from the C–H stretch of the alkyl group.

Fig. 4 FT-IR spectra (KBr) of C₆₀ and fullerene derivatives 3 and 4 in the wavenumber range (a) 3700–470 cm⁻¹ and (b) 1470–470 cm⁻¹.

3.5 Electrochemical studies on SAMs

The electrochemical studies of the SAMs on a amine-terminated Si surface were performed in a 0.1 M solution of TBABF₄ in 1,2-dichlorobenzene–DMF (3 : 1). It is well known that the amine-terminated thiols act as bifunctional building blocks, where the sulfur atoms bind to the gold surface while the amino groups may be employed for the attachment of other functional groups to the SAMs,^34,35 in this work to C₆₀. Gold substrates and alkane-thiol SAMs are another preferred chemical system used as C₆₀ support material.³⁶ Schemes S1 and S2 in ESI† show the preparation steps for C₆₀ SAMs on Si and gold used in this work. Fig. 5 shows a typical cyclic voltammogram of the SAM of C₆₀ on cysteamine-Au/Si (curve b) compared to the response obtained only for support material (curve a). Two electrochemically reversible waves are observed, and the potential for the first and second one-electron reductions are −0.61 and −1.19 V, respectively. For these SAMs the currents diminish somewhat during repeated cycling and when using silanized Si substrates no current response from C₆₀ could be obtained. Coupling of C₆₀ to this support material could anyhow be confirmed by infrared spectroscopy and AFM, and will be discussed later on in the text, indicating that the SAMs of C₆₀ on silane/Si are formed but not stable upon potential cycling.

Fig. 5 Cyclic voltammograms of (a) Si/Au-cysteamine and (b) the SAM of C₆₀ on the Si/Au-cysteamine electrode in 1,2-dichlorobenzene–DMF (3 : 1) containing 0.1 M of TBABF₄ using 50 mV s⁻¹ as scan rate in the potential range 0 to −2.0 V.

Next the self-assembly of fullerene derivatives 2 and 4 on ITO and modified ITO substrates were investigated. Phosphonic fullerene compounds have been found to form stable SAMs on ITO and SnO₂.³⁷ The voltammetric response from the SAM of 2 and 4 on ITO plates (curves c and d) compared to that obtained for C₆₀ and derivative 2 in solution (curves a and b) are shown in Fig. 6a. The cyclic voltammetry of derivative 2 monolayer modified ITO is different from that of derivative 2 in solution on a bare ITO electrode. In Fig. 6a we see three reversible reduction–reoxidation waves of solution with derivative 2 on bare ITO. For ITO modified with a fullerene derivative monolayer, however, the first redox step appears to be reversible and both the first and especially the second redox potentials are shifted positive in potential compared to that obtained for the solution. The first fullerene centered reduction peak for the SAM of 2 is 70 mV less negative (−0.55 versus −0.48 V) than that obtained for derivative 2 in solutions. The origin could lie in fast electron transfer between the ITO substrate and the fullerene derivative directly attached onto the electrode surface. As can be seen from Fig. 6a the potentials, especially for the first two reduction peaks of derivatives, are shifted to negative values as compared with formal potentials of the corresponding redox-transitions of C₆₀. This was observed also when using Pt as working electrode and has already been discussed in Section 3.1. In order to obtain uniformly distributed –OH groups on the ITO surface the ITO glass plates were hydrolyzed using two different methods (as explained in the Experimental section) and immersed directly into solutions of derivative 2 or 4 in order to form SAMs. These structures were characterized by CV and the response obtained for derivative 4 on hydrolyzed ITO plates in a 0.1 M solution of TBABF₄ in 1,2-dichlorobenzene–DMF (3 : 1) is shown in Fig. 6b (curve e) together with the responses from Fig. 6a but now in a narrower potential range from 0 to −1 V. Compared to the unmodified ITO plates both derivatives showed upon continuous cycling higher and stable current response, especially for the first redox wave with the peak maximum located at −0.47 V. The stability indicates that there is specific bonding between the C₆₀ derivatives and the OH groups on the hydrolyzed ITO surface. There are reports on using aqueous ZrOCl₂ for functionalization of silicon and gold surfaces for the attachment of phosphonic anchoring agents.³⁸ In an previous study by our group on phosphoquinone SAMs, both native and zirconated ITO were used in order to understand the interactions between the electrode surface and the adsorbed molecules.¹⁷ In this work we used this concept for modification of the ITO plate in order to get a anchoring group for the phosphonate groups on fullerene derivatives 2 and 4. The CV curve recorded for derivate 2 self-assembled onto the Zr functionalized ITO surface is shown in Fig. 6b (curve f). Compared to the other SAMs the peak potentials for the first redox response stay the same but the current response and stability is greatly improved. The reoxidation of the SAMs are shifted in potential compared to the case of fullerene solutions and it is more complicated than a one-electron oxidation step. The shift and merging of multiple oxidation peaks into almost a single broad anodic wave indicates that it is hard to reoxidize the films after reduction.

Fig. 6 In (a) cyclic voltammograms of C₆₀ (a) and derivative 2 (b) in solution and from SAMs of derivative 2 (c) and 4 (d) on ITO in 1,2-dichlorobenzene–DMF (3 : 1) containing 0.1 M of TBABF₄ using 50 mV s⁻¹ as scan rate in the potential range 0 to −2.0 V. In (b) the potential range 0 to −1.0 V is shown upon functionalization of the ITO surface.

The CV curve recorded for derivate 2 self-assembled onto the Zr functionalized ITO surface and further self-assembled with the soluble polythiophene polymer (structure shown in Scheme S4, ESI†) is shown in Fig. 7. Compared to the CV curve for derivate 2 the first redox response show a shift in potential with the redox maximum located at −0.65 V (shift of 180 mV compared to curve f in Fig. 6b). The conducting polymer shows an oxidation response at 0.63 V.

Fig. 7 Cyclic voltammograms from SAMs of polythiophene-derivative 2 onto Zr functionalized ITO surface in 1,2-dichlorobenzene–DMF (3 : 1) containing 0.1 M of TBABF₄ using 50 mV s⁻¹ as scan rate in the potential range −1.0 to 1.0 V.

3.6 FTIR Spectroscopy of SAMs

Infrared reflectance measurements were used to characterize the SAM of C₆₀ on silanized Si substrates. The obtained spectra, using a cleaned Si wafer as reference, in the range from 650 to 1750 cm⁻¹ and from 2700 to 3600 cm⁻¹ is shown in Fig. 8. From the spectrum the two C–H stretching modes, asymmetrical and symmetrical at 2919 and 2848 cm⁻¹, respectively, reveal successful attachment of the (3-aminopropyl)triethoxysilane onto Si.³⁹ The single absorption at 1113 cm⁻¹ is due to the Si–O–Si asymmetric stretching, accompanied by the symmetric stretch at 818 cm⁻¹. The broad band at around 3600–3200 cm⁻¹ corresponds to the N–H stretching vibration due to secondary amine, formed by addition of hydrogen in primary amine to double bond in C₆₀.⁴⁰ In addition, the presence of several vibrational features ascribed to the C₆₀ skeleton ring vibration observed at 1541, 1418, 1390 and 1185 cm⁻¹ showed that C₆₀ was successfully grafted to the silanized Si surface.^4,41

Fig. 8 FT-IR spectra of C₆₀ attached to the silanized Si surface in the wavenumber range 3600–2750 cm⁻¹ and 1750–650 cm⁻¹.

3.7 Contact angles on SAMs

Depending on surface treatment of the ITO plate we should obtain a specific binding of the fullerene derivatives to the plates and in this way receive different contact angles of water. Contact angle measurements provide a sensitive probe of the nature of the terminal functional group of organic thin films. The contact angle measurements for fullerene SAMs on ITO are summarized in Table 2 and show increased hydrophilicity for SAMs of derivative 2 and 4 on hydrolyzed ITO plates. This indicates that the derivatives are bound to the surface with the hydrophilic phosphonate groups pointing out from the surface. Further functionalization of the hydrolyzed ITO plate with Zr will function as anchoring group for the phosphonate groups on fullerene derivatives. The hydrophobicity of the C₆₀ modified surfaces are confirmed by increase of the contact angle. Several groups have reported similar water contact values for C₆₀ terminated surfaces.^4,42

Table 2 Water contact angles for fullerene SAMs on ITO

Sample	Contact angle, deg
Washed ITO plate	70
ITO treated with OH	34
Fullerene derivative 2 on ITO (OH-treated)	16
Fullerene derivative 4 on ITO (OH-treated)	21
Fullerene derivative 2 on ITO (Zr-treated)	58
Fullerene derivative 4 on ITO (Zr-treated)	56

---

3.8 UV-Vis spectra of derivative 2 and 4 on ITO

Fig. 9 compares the UV-Vis spectra for derivative 2 in toluene (curve c) to the SAM of derivative 2 and 4 on hydrolyzed ITO plates (curve a and b) in the range 300–600 nm. The spectrum of the derivatives on ITO show one strong peak appearing at 320 nm and one broad peak at 400–600 nm, which resembles the characteristic absorption of the derivative in solution, indicating successful grafting of the derivatives onto the ITO surface. In comparison with that of the derivatives in solution, the peak at 320 nm shows a blue shift because of electron transport from C₆₀ derivatives attached to the ITO plate. Also the broad band at 520 nm, characteristic to fullerene derivatives, caused by the breakage of symmetry and electronic structure of C₆₀, indicates grafting onto the ITO plate.

Fig. 9 UV-Vis absorption spectra of SAMs of (a) derivative 2, (b) derivative 4 on ITO and (c) derivative 2 in toluene.

3.9 Surface morphology analysis

The assembled structures of C₆₀ and derivatives 2 and 4 on Si wafers were characterized by AFM. Fig. 10 shows the AFM images of Si (images A–C) and for Si with a gold film grown by vacuum deposition (images D–F) after different surface modification steps. The value inserted on the image gives the corresponding roughness value. At the starting point the cleaned Si wafer show a roughness (rms) of around 0.4 nm (4 Å) and upon evaporating it with gold a rms of 6 nm (A and D, respectively). Both surfaces show also crystalline domains ranging from 30–70 nm in height for Si–Au (0.4 μm in width) and 30 nm for Si (0.15 μm in width). After silanization by 3-aminopropyl triethoxysilane (APTES) of the Si wafer the surface became rougher with an rms of 2.5 nm (B). Similar behaviour is observed for the cysteamine modified Si–Au wafer showing an rms of ∼7 nm (E). Modification with C₆₀ molecules, further roughened the Si surface (rms 3.5–4 nm, C) but also the crystalline domains, that can be observed throughout all images, increase in height. Further on some even bigger grains can be found in the images. These grains, which are at maximum 0.7 μm (C) and 0.25 μm across (F) are several orders of molecular diameters of C₆₀. These grains could be due to C₆₀ aggregated clusters that are attached to the surface fullerenes through van der Waals attractions. Similar microcrystalline aggregated clusters have been reported for fullerene molecules adsorbed on silicon and mica surface.⁴³ The apparent height from the AFM images seems to be close to the 2–3 nm range generally detected by AFM for C₆₀ SAMs^14,44 why it is reasonable to consider that we have semimonolayer films on Si. It is important to take into account that it is very difficult to obtain atomic resolution AFM images of molecules that are smaller than 50 nm at room temperature, especially if they are adsorbed on top of a “soft” layer for reasons such as thermal motion and the broadening effect of the radius of the tip during AFM scanning.

Fig. 10 AFM images: (A) cleaned Si wafer, (B) APTES SAM on Si, (C) C₆₀ SAM on APTES Si, (D) Si wafer with Au, (E) cysteamine SAM on Si/Au and (F) C₆₀ SAM on cysteamine Si/Au. The value inserted on the image gives the rms roughness values.

The AFM images of assembled layers of derivative 4 on ITO that had been hydrolysed and treated with Zr are shown in ESI, Fig. S3A.† Treatment with Zr resulted in SAMs with more sample particles on the surface than if the ITO plate had been just hydrolyzed. As seen from Fig. S3A† the SAMs of derivative 4 showed an rms of 0.6 nm with particles at maximum 20 nm in height with a diameter of 0.4 μm. The surface image of derivative 4 on ITO obtained by SEM (Fig. S3B†) shows the typical surface structure of ITO with a coherent, but not fully covering SAM layer.

4 Conclusions

In this work four different methano[60]fullerenediphosphonate derivatives have been synthesized. The fullerene derivatives were firstly characterized in solutions by electrochemical and spectroscopic methods. With an increasing degree of functionalization of the fullerene cage the reductions became more difficult and irreversible. This shift could be nicely correlated with the obtained UV-Vis spectra that showed a shift of the absorption at 500 nm to higher energies upon increased functionalization. Furthermore we showed that fullerene based SAMs can be prepared on silicon, gold and ITO substrates and that the chemical modification of the substrate will play an important role for obtaining a stable electrochemical response. In case of ITO the modification of the surface with ZrOCl₂ facilitated anchoring of the methano[60]fullerenediphosphonate and also increased the stability of the redox response upon potential cycling. For the SAMs we obtained a shift of the redox potentials towards lower values in comparison to what was obtained for the fullerene derivatives in solution. The electrochemical results were combined with spectroscopic and microscopic methods in order to understand the chemistry of the attachments.

Acknowledgements

We gratefully acknowledge the financial support from Academy of Finland.

Notes and references

1. H. W. Kroto, J. R. Heath, S. C. O'Brien, R. F. Curl and R. E. Smalley, Nature, 1985, 318, 162.
2. W. Kraetschmer, L. D. Lamb, K. Fostiropoulus and D. R. Huffmann, Nature, 1990, 347, 354.
3. V. V. Tsukruk, M. P. Everson, L. M. Lander and W. J. Brittain, Langmuir, 1996, 12, 3905.
4. R. R. Sahoo and A. Patnaik, J. Colloid Interface Sci., 2003, 268, 43.
5. A. Patnaik, K. K. Okudaira, S. Kera, H. Setoyama, K. Mase and N. Ueno, J. Chem. Phys., 2005, 122, 154703.
6. S. K. Hau, Y.-J. Cheng, H.-L. Yip, Y. Zhang, H. Ma and A. K.-Y. Jen, ACS Appl. Mater. Interfaces, 2010, 7, 1892.
7. G. Granicini, R. S. Santosh Kumar, A. Abrusci, H.-L. Yip, C.-Z. Li, A. K.-Y. Jen, G. Lanzani and H. J. Snaith, Adv. Funct. Mater., 2012, 22, 2160.
8. A. Abrusci, S. D. Stranks, P. Docampo, H.-L. Yip, A. K.-Y. Jen and H. J. Snaith, Nano Lett., 2013, 13, 3124.
9. K. Chen, W. B. Caldwell and C. A. Mirkin, J. Am. Chem. Soc., 1993, 115, 1193.
10. A. L. Mirakyan and L. J. Wilson, J. Chem. Soc., Perkin Trans. 2, 2002, 6, 1173.
11. C. Boudon, J. P. Gisselbrecht, M. Gross, L. Isaacs, H. Anderson and K.-D. Asmus, Helv. Chim. Acta, 1995, 78, 1334.
12. F. Cardullo, P. Seiler, L. Isaacs, J. F. Nierengarten, R. F. Haldimann, F. Diederich, T. Mordasini-Denti, W. Thiel, C. Boudon and J. P. Gisselbrecht, Helv. Chim. Acta, 1997, 80, 343.
13. H. Fujihara and H. Nakai, Langmuir, 2001, 17, 6393.
14. X. Zhang and A. V. Teplyakov, Langmuir, 2008, 24, 810.
15. F. Song, S. Zhang, D. Bonifazi, O. Enger, F. Diederich and L. Echegoyen, Langmuir, 2005, 21, 9246–9250.
16. A. Fuschbauer, O. A. Trishina, P. A. Troshin, R. Koeppe, R. N. Lyubovskaya and N. S. Sariciftci, Adv. Funct. Mater., 2008, 18, 2808.
17. A. Viinikanoja, J. Lukkari, T. Ääritalo, T. Laiho and J. Kankare, Langmuir, 2003, 19, 2768.
18. H. Byrd, J. K. Pike and D. R. Talham, Chem. Mater., 1993, 5, 709.
19. I. A. Nuretdinov, V. P. Gubskaya, N. I. Shishikina, G. M. Fazleeva, L. Sh. Berezhnaya, I. P. Karaseva, F. G. Sibgatullina and V. V. Zverev, Russ. Chem. Bull., 2002, 51, 337.
20. P. D. F. Siqueira, G. Wenz, P. Schunk and T. Schimmel, Langmuir, 1999, 15, 4520.
21. M. Eiermann, R. C. Haddon, B. Knight, Q. C. Li, M. Maggini, N. Martin, T. Ohno, M. Prato, T. Suzuki and F. Wudl, Angew. Chem., Int. Ed. Engl., 1995, 34, 1591.
22. P. A. Troshin, H. Hoppe, J. Renz, M. Egginger, J. Y. Mayorova, A. E. Goryachev, A. S. Peregudov, R. N. Lyubovskaya, G. Gobsch, N. S. Sariciftci and V. F. Razumov, Adv. Funct. Mater., 2009, 19, 779.
23. T. F. Guarr, M. S. Meier, V. K. Vance and M. Clayton, J. Am. Chem. Soc., 1993, 115, 9862.
24. S. A. Lerke, B. A. Parkinson, D. H. Evans and P. J. Fagan, J. Am. Chem. Soc., 1992, 114, 7807.
25. P. Boulas, F. D'Souza, C. C. Henderson, P. A. Cahill, M. T. Jones and K. M. Kadish, J. Phys. Chem., 1993, 97, 13435.
26. K. Winkler, D. A. Costa, A. L. Balch and W. R. Fawcett, J. Phys. Chem., 1995, 99, 17431.
27. M. Ocafrain, M. A. Herranz, L. Marx, C. Thilgen, F. Diedrich and L. Echegoyen, Chem.–Eur. J., 2003, 9, 4811.
28. V. V. Yanilkin, V. P. Gubskaya, V. I. Morozov, N. V. Nastapova, V. V. Zverev, E. A. Berdnikov and I. A. Nuretdinov, Russ. J. Electrochem., 2003, 39, 1147.
29. F. Cardullo, P. Seiler, L. Isaacs, J. F. Nierengarten, R. F. Haldimann, F. Diederich, T. Mordasini-Denti, W. Thiel, C. Boudon and J. P. Gisselbrecht, Helv. Chim. Acta, 1997, 80, 343.
30. R. V. Bensasson, E. Bienvenüe, C. Fabre, J.-M. Janot, E. J. Land, S. Leach, V. Leboulaire, A. Rassat, S. Roux and P. Seta, Chem.–Eur. J., 1998, 4, 270.
31. L. Isaacs, A. Wehrsig and F. Diederich, Helv. Chim. Acta, 1993, 76, 1231.
32. A. Viinikanoja, S. Areva, N. Kocharova, T. Aeaeritalo, M. Vuorinen, A. Savunen, J. Kankare and J. Lukkari, Langmuir, 2006, 22, 6078.
33. Z. C. Wu, D. A. Jelski and T. F. George, Chem. Phys. Lett., 1987, 137, 291.
34. A. Riklin and I. Willner, Anal. Chem., 1995, 67, 4118.
35. F. Arias, L. A. Godinez, S. R. Wilson, A. E. Kaifer and L. Achegoyen, J. Am. Chem. Soc., 1996, 118, 6086.
36. W. B. Caldwell, K. Chen and C. A. Mirkin, Langmuir, 1993, 9, 1945.
37. A. Sakamoto, Y. Matsuo, K. Matsuo and E. Nakamura, Chem.–Asian J., 2009, 4, 1208.
38. H. Lee, L. J. Kepley, H. G. Hong, S. Akhter and T. E. Mallouk, J. Phys. Chem., 1988, 92, 2597.
39. Y. Wu, B. Zhao, W. Xu, B. Li, Y. M. Jung and Y. Ozaki, Langmuir, 1999, 15, 4625.
40. T. Sainsbury and D. Fitzmaurice, Chem. Mater., 2004, 16, 3780.
41. J. Fabian, Phys. Rev. B: Condens. Matter, 1996, 53, 13864.
42. F. Song, S. Zhang, D. Bonifazi, O. Enger, F. Diederich and L. Echegoyen, Langmuir, 2005, 21, 9246.
43. R. R. Sahoo and A. Patnaik, Appl. Surf. Sci., 2005, 245, 26.
44. S. A. Backer, I. Suez, Z. M. Fresco, M. Rolandi and J. M. Frechet, Langmuir, 2007, 23, 2297.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis of fullerene derivatives (1–4), UV-Vis spectra for C₆₀ and derivative 4 at different concentrations, AFM and SEM images of SAMs of derivative 4, preparation schemes for SAMs of fullerene and fullerene derivatives 2 and 4 and a scheme showing the structure for the water soluble polythiophene used in PL measurements. See DOI: 10.1039/c3ra46740f

---