Hydrophilic Micro- and Macroelectrodes with Antibiofouling Properties for Biomedical Applications

Abstract

Implantable neural electrodes are generally used to record the electrical activity of neurons and to stimulate neurons in the nervous system. Biofouling triggered by inflammatory responses can dramatically affect the performance of neural electrodes, resulting in decreased signal sensitivity and consistency over time. Thus, long-term clinical applications require electrically conducting electrode materials with reduced dimensions, high flexibility, and antibiofouling properties that can reduce the degree of inflammatory reactions and increase the lifetime of neural electrodes. Carbon nanotubes (CNTs) are well known to form flexible assemblies such as CNT fibers. Herein, we report the covalent functionalization of predefined CNT fiber and film surfaces with hydrophilic, antibiofouling phosphorylcholine (PC) molecules. The electrochemical and spectroscopic characteristics, impedance properties, hydrophilicity, and in vitro antifouling nature of the functionalized CNT surfaces were evaluated. The hydrophilicity of the functionalized CNT films was demonstrated by a decrease in the static contact angle from 134.4° ± 3.9° before to 15.7° ± 1.5° after one and fully wetting after three functionalization cycles, respectively. In addition, the extent of protein absorption on the functionalized CNT films was significantly lower than that on the nonfunctionalized CNT film. Surprisingly, the faradic charge-transfer properties and impedance of the CNT assemblies were preserved after functionalization with PC molecules. These functionalized CNT assemblies are promising for the development of low-impedance neural electrodes with higher hydrophilicity and protein-fouling resistance to inhibit inflammatory responses.

Graphical Abstract

1. INTRODUCTION

Implantable medical electrodes and neuroprosthetic devices are used to record reliable and accurate information as well as stimulate target tissues.¹ Hence, consistent performance is critical for long-term applications. Although a large variety of materials are available for developing sensitive and consistent implantable electrodes, most electrodes are prone to failure over time during in vivo applications.² Current commercial devices have a short life time of a few months to several years. The surgical implantation of medical devices results in tissue injury, which elicits a chain reaction of chemical and biological events that lead to acute and chronic inflammatory responses.^3,4 Such inflammatory responses cause plasma protein and inflammatory cells such as microglia, macrophages, and astrocytes to bind to the medical device, thus creating a physical, chemical, and electrical barrier between the target tissue and the electrode surface. The complete encapsulation of the electronic devices leads to device failure because signal communication is interrupted.^5,6 Among numerous materials that have been employed as microelectrodes (MEs) in neuroprosthetic applications, traditional metals (titanium and tungsten), alloys (stainless-steel, nickel-chromium, and titanium-iridium), and silicon ME arrays cause considerable tissue damage, resulting from the mechanical mismatch between soft tissues and rigid metals, and induce faster biofouling, leading to performance inconsistency during long-term applications.^1,6,7 The hydrophobic nature of most MEs facilitates the initial adsorption of plasma proteins on the electrode surface, which promotes the subsequent attachment of inflammatory cells in the damaged tissue enviroment.^3,8,9 Two main strategies have been developed to minimize inflammatory responses and to reduce initial plasma protein and inflammatory cell attachment: (1) the use of soft and flexible conductive materials to reduce tissue damage during insertion and (2) the grafting of hydrophilic and antibiofouling coatings on the surface of the sensing area to minimize protein and inflammatory cell attachment.^9,10

More recently, carbon-based nanomaterials such as carbon nanotubes (CNTs) and graphene have shown impressive performances in terms of electrochemical characteristics and reduced inflammatory responses. Owing to their unique nanoscale structural and electronic features, CNTs are suitable for various biological applications such as implantable biosensors and neuroprosthetic devices, which require direct communication between single cells and man-made devices. CNTs are hollow, cylindrical nanostructures composed of rolled graphene sheets made of sp² hybridized carbon atoms.¹¹ Single-walled CNTs (SW-CNTs), which comprise a single graphene layer, have diameters in the range of 0.5–3 nm, whereas multiwalled CNTs (MW-CNTs) comprise an array of concentric cylinders, typically with diameters greater than 3 nm.¹² Because of their unique arrangement of sp² hybridized carbon atoms, CNTs show excellent electrical and thermal properties, a high degree of chemical resistance (inertness), and superior mechanical properties (elastic modulus high as 1 TPa and tensile strength up to 300 GPa).^12,13 The inherent features of CNTs should allow the development of advanced bioelectronic devices that are capable of efficient signal communication in highly corrosive environments with minimal tissue damage and long-term durability and flexibility. As CNTs also show good biocompatibility and safe charge-transfer properties, they are promising as electrodes for the human-electronic interface development.^7,14

Despite the above-mentioned properties, the full potential of CNTs has not been exploited to overcome the disadvantages of the typical materials employed in biomedical applications. CNTs have commonly been used as coatings on metals to improve the electronic properties, impedance, and nanoscale surface features of MEs.¹⁵ However, CNTs can also be used independently to develop implantable electrodes and interfaces. Recently, CNTs in the form of aligned films, vertically aligned ME arrays, and fibers have been employed to fabricate bioelectronic devices.^7,16–18 The Alvarez lab recently reported highly densified CNT fiber cross sections consisting of millions of exposed open-ended CNTs, which are effective as an electrode–electrolyte interface for the real-time detection of picomolar concentrations of dopamine in cell cultures.¹⁹ This novel interface was shown to have higher sensitivity for neurotransmitters and a lower impedance (23.6 MΩ μm²) than other metal- and CNT-based electrodes.^19,20 In addition, several research groups have demonstrated that CNT fibers can be used as neural electrodes.^7,21,22 Lu et al. reported that soft CNT fiber MEs precisely recorded high-quality single-unit neural signals from rats for up to 4–5 months with greatly reduced inflammatory responses.²² Furthermore, the charge-transfer characteristics of the reported CNT fiber MEs could be improved by surface modification. Vitale et al. used the electrical stimulation of an in vivo model to demonstrate the stability of CNT fiber electrodes, with no degradation of recording quality observed over a 4 week period. This study also revealed that the electrochemical stability and chemical stability of the CNT fiber electrodes were superior to those of polymer-coated metal electrodes. In addition, the CNT fiber electrodes were shown to reduce inflammatory responses as compared to Pt electrodes.⁷ These studies indicate that the use of flexible CNT fibers as implantable sensors can reduce tissue damage and biofouling, and the improved electrochemical properties can be obtained by modifying their surface chemistry. CNTs can be further improved by functionalization using antibiofouling molecules.

Alternatively, zwitterionic polymers with phosphorylcholine (PC), carboxybetaine (CB), and sulfobetaine (SB) as functional groups have been used to modify the surfaces of bioimplants. This approach provides superior antifouling properties, with low serum protein adsorption (<0.3 ng cm⁻²).^23–26 Antibiofouling zwitterionic polymer coatings have been shown to reduce inflammatory cell attachment in brain implants. The Cui lab reported that silicon probes coated with zwitterionic poly(sulfobetaine methacrylate) (PSB) and polydopamine (PDA) reduced protein adsorption by 89% and reduced fibroblast adhesion by 86% compared to bare silicon probes.⁹ In mouse brain studies, the PSB/PDA-modified silicon probes showed a significant decrease in reactive astrocytes within 70 μm of the electrode–tissue interface as compared to an unmodified silicon implant. In another study, the Cui lab used a photoiniferter approach to directly graft PSB on silicon probes which significantly reduced microglia endfeet spreading on the probe surface, as revealed by in vivo two-photon imaging.²⁶ These findings indicate that zwitterionic coatings can effectively reduce acute inflammatory responses. In addition, Zhang et al. reported that zwitterionic hydrogels prepared from CB strongly inhibit fibrotic capsule formation in a mouse model for at least for 3 months.²⁷ However, the modification of electrodes with zwitterionic polymers tends to increase impedance. Therefore, with such coatings, a trade-off between impedance and fouling resistance must be considered.^28,29 Gui et al. reported an alternative approach for introducing zwitterionic PC molecules onto electrode surfaces without building up a large charge-transfer barrier at the electrode–electrolyte interface.³⁰ Several groups have modified electrode surfaces with zwitterionic phenylphosphorylcholine (PPC) molecules to achieve antifouling properties using a similar approach. Zwitterionic PPC has been successfully grafted on gold, glassy carbon (GC), indium tin oxide (ITO), and reduced graphene oxide nanosheets by electrografting aryl diazonium salts.^30–35

This report demonstrates the covalent functionalization of two types of CNT assemblies (macro and microassemblies) with extremely hydrophilic and antibiofouling PPC molecules. The microassembly was fabricated by encapsulating a CNT fiber (60 μm diameter) in a polymer, which then was sliced into 100–150 μm sections, with the CNT fiber cross section (CNT-Fcs) serving as a ME. The macroassembly was fabricated by drawing a CNT array into a woven film on a flat substrate (CNT film electrode). Electrochemical functionalization of the CNT-Fcs ME and CNT film electrode with PPC molecules was accomplished using aryl diazonium chemistry, as shown in Figure 1. The spectroscopic and electrochemical characteristics, impedance, hydrophilicity, and protein-fouling resistance of the CNT assemblies are reported. Surface functionalization of the thin CNT films was confirmed using spectroscopic methods, and the functionalization parameters were subsequently applied to the CNT-Fcs MEs.

Figure 1.

Electrochemical functionalization of CNT-Fcs ME and CNT film electrodes with PPC molecules. (A) Illustration of a CNT fiber cross section displaying the open-ended nature of the CNTs and their functional groups. (B) Illustration of side-wall functionalization of the CNTs in the CNT film electrode.

2. EXPERIMENTAL SECTION

2.1. Reagents and Materials.

4-Aminophenylphosphorylcholine (C₁₁H₁₉N₂O₄P, ≥98%, Cayman Chemical Company, Ann Arbor, MI) was used for electrochemical functionalization of CNT assemblies. Sodium nitrite (NaNO₂, ≥97%) was obtained from Chem-Impex International Inc. (Wood Dale, IL). Fuming hydrochloric acid (HCl, 37%), potassium ferricyanide(III) (K₃[Fe(CN₆)], 99%), potassium hexacyanoferrate(II) trihydrate (K₄[Fe(CN)₆].3H₂O, 98.5%), potassium chloride (KCl, 99%), acetone (99.7%), albumin–fluorescein isothiocyanate conjugate (FITC-albumin, (catalog # A9771), and phosphate-buffered saline (PBS) were purchased from Sigma-Aldrich (St. Louis, MO). Hexaammineruthenium(III)chloride ([Ru(NH₃)₆]-Cl₃, 98%) was purchased from Acros Organics (Waltham, MA). Alexa Fluor 294 goat antirabbit IgG (catalog # A-11012) was purchased from ThermoFisher Scientific (Waltham, MA). To prepare the CNT-Fcs ME, densified CNT fibers were embedded in an embedding polymer kit (EMBed-812, Electron Microscopy Sciences). The polymer kit consists of EMBed-812 resin, monomers dodecenyl succinic anhydride (DDSA) and methyl-5-norbornene-2,3-dicarboxylic anhydride (NMA), and the curing agent n-benzyldimethylamine (BDMA). Silver paint was purchased from Electron Microscopy Sciences (Hatfield, PA). Milli-Q water (18.2 MΩ cm) was used to prepare all the reagent solutions. Vertically aligned CNTs (VA-CNTs) were synthesized using a chemical vapor deposition (CVD) method with ethylene (Wright Brothers, Cincinnati, OH) and Fe/Co (Goodfellow, Coraopolis, PA) as the carbon source and catalyst, respectively.

2.2. CNT Fiber Synthesis and Densification.

Spinnable VA-CNT arrays were employed for the synthesis of CNT fibers. The CVD process for synthesizing VA-CNTs has been reported by our team elsewhere.³⁶ Dry spinning was initiated by drawing a CNT film from the edge and continuously spinning into a fiber using a homemade twisting setup. As the freshly spun fiber was not densely packed, it was placed in acetone for 72 h at room temperature to achieve a higher CNT density within the fiber. After the densification process, the diameter of the CNT fiber was 60 μm, and those densified CNT fibers were used to fabricate CNT-Fcs MEs.

2.3. Fabrication of CNT-Fcs MEs.

The fabrication process of the CNT-Fcs ME consisted of three steps: encapsulation of the densified CNT fibers, sectioning of the encapsulated CNT fibers, and CNT-Fcs ME assembly. The fabrication of the CNT-Fcs MEs was previously reported by our group.^19,37 First, the EMBed-812 resin mixture was prepared according to the manufacturer’s instructions by mixing EMBed-812 (48%) with the monomers, DDSA (36%) and NMA (18%). The curing agent BDMA (3%) was added to the mixture. Then, three 60 μm densified CNT fibers were placed vertically in a 2 mL cylindrical mold, into which the embedding polymer mixture was added. After curing at 80–85 °C for 6 h, the polymer-embedded CNT fibers were removed from the mold. To prepare cross sections of embedded CNT fibers, the cylindrical cured polymer was sliced into 100–150 μm cross sections using a microtome. The prepared polymer films consisted of three fibers embedded in the polymer matrix with exposed fiber cross sections on both sides. Silver paint was used to make an electrical connection with one side of the fiber cross section. The silver paint was applied where the CNT fibers were exposed in the cross section and a Cu wire was attached. The electrical connection was then sealed with an epoxy resin, leaving the opposite side of the cross section with exposed CNT fibers for functionalization with PPC molecules. Schematic representation of the fabrication process for CNT-Fcs MEs is shown in Figure S1A.

2.4. CNT Film Electrode Preparation.

CNT films were prepared by drawing layers of CNTs from a VA-CNT array onto a polyethylene terephthalate (PET) substrate. Four layers of CNT films were drawn in a crosshatch pattern, and then the prepared film was densified by placing acetone (~1 mL) on the film. After densification, the film was connected to a copper wire using silver paint and the electrical connection was sealed with the epoxy resin. Schematic representation of the fabrication process for CNT film electrodes is shown in Figure S1B.

2.5. Electrochemical Measurements and Functionalization.

A PalmSens3 electrochemical workstation (Houten, The Netherlands) was used for the electrochemical characterization. A PalmSens4 electrochemical workstation (Houten, The Netherlands) was used for electrochemical impedance spectroscopy (EIS) measurements. All the electrochemical experiments were performed using a three-electrode system with CNT-Fcs ME or CNT film electrode as the working electrode, a Pt wire as the counter electrode, and an Ag/AgCl electrode as the reference electrode, unless otherwise specified. All potentials are reported versus the Ag/AgCl reference electrode at room temperature.

Surface functionalization of both CNT assemblies was conducted as follows. The CNT electrode was washed with Milli-Q water before functionalization. A 5 mM 4-aminophenylphosphorylcholine solution in 0.25 M HCl and a 5 mM NaNO₂ aqueous solution were freshly prepared. After mixing equimolar 4-aminophenylphosphorylcholine and NaNO₂, the solution was degassed with N₂ for 30 min before electrochemical functionalization. The CNT electrode was functionalized by cycling the potential between 0.6 and −1.0 V at a scan rate of 50 mV s⁻¹ in the degassed PPC diazonium salt solution. After the functionalization, the electrode was washed with Milli-Q water and air-dried.

2.6. Electrochemical Evaluation of Functionalized CNT Assemblies.

2.6.1. Cyclic Voltammetry (CV).

Two different redox probes, Ru(NH₃)₆³⁺ and Fe(CN)₆^3−/4−, were used to evaluate surface functionalization. CV was performed with both redox probes before and after the surface functionalization with 0.1 M KCl as the supporting electrolyte. To analyze the behavior of Ru(NH₃)₆³⁺, the potential was cycled between +0.2 and −0.45 V at a scan rate of 50 mV s⁻¹ in 0.1 M KCl aqueous solution containing 2 mM Ru(NH₃)₆Cl₃. To analyze the behavior of the Fe(CN)₆^3−/4− couple, the potential was cycled between +0.8 and − 0.5 V at a scan rate 100 mV s⁻¹ in 0.1 M KCl aqueous solution containing 4 mM K₃[Fe(CN)₆]/K₂[Fe(CN)₆] (1:1).

2.6.2. EIS Measurements.

EIS characterization of the functionalized CNT surfaces in 0.01 M PBS (pH 7.4) was performed in the frequency range of 1 MHz to 1 Hz with an amplitude of 10 mV. For each electrode, the impedance was measured before and after surface functionalization. A CNT-Fcs ME prepared by embedding only one fiber instead of three fibers was employed to minimize complications resulting from surface damage during microtoming.

2.7. X-Ray Photoelectron Spectroscopy (XPS) Analysis.

The surface chemistry of the functionalized CNT-Fcs MEs and CNT films was analyzed using XPS. For CNT-Fcs MEs, XPS measurements of the 60 μm fiber cross section were performed using a PHI VersaProbe II XPS scanning microprobe (Physical Electronics, Chanhassen, MN, USA) with an Al Kα X-ray microfocused monochromator and a multichannel detector. High-resolution XPS spectra were collected at a takeoff angle of 40 ° with a spot size of 50 μm diameter. High-resolution scans were performed for C 1s, O 1s, N 1s, and P 2p with a step size of 0.20 eV, a dwell time of 50 ms/step, and 15 scans for each element. The XPS spectra acquired for fiber cross sections were analyzed using the MultiPak analysis software, and atomic percentages were obtained for each species.

XPS measurements on the CNT films were performed using a KAlpha X-ray photoelectron spectrometer system (Thermo Scientific, Waltham, MA, USA) with an Al Kα X-ray microfocused monochromator and a multichannel detector. High-resolution XPS spectra were collected at a takeoff angle of 90° with a spot size of 400 μm. High-resolution C 1s, O 1s, N 1s, and P 2p scans were performed with a step size of 0.1 eV and a dwell time of 50 ms. The number of scans was 10 for C and O, whereas the number of scans was increased to 20 for N and P. The XPS spectra acquired for CNT films were analyzed using Avantage surface chemical analysis software, and atomic percentages were obtained for each species.

Nonfunctionalized CNT-Fcs ME and CNT films were analyzed as controls. Origin Pro 8.5 was used to deconvolute the C 1s, O 1s, N 1s, and P 2p spectra. For each type of CNT surface, at least two samples were analyzed.

2.8. Water Contact Angle Analysis.

The hydrophilicity of the functionalized CNT films was assessed, and nonfunctionalized CNT films were used as a control. The functionalized CNT films were thoroughly washed with Milli-Q water before water contact angle measurements. The contact angle measurements were performed using a Rame-´Hart 200-F1 standard goniometer (Succasunna, NJ, USA) with a manual tilting base. Images were captured using a high-resolution camera and analyzed with DROPimage Standard software. The goniometer tilting base was set to 0°, and the sample stage was manually adjusted to ensure that each sample surface was leveled. A 10 μL droplet of deionized water was placed on the sample surface using a microsyringe, and the contact angle was recorded. For each sample, static contact angles were measured for three individual water droplets placed at different locations across the surface, and the average is reported. Measurements were performed approximately 10 s and 2 min after placing the droplet on the CNT film. As there was no significant difference between the recorded contact angles, the reported static contact angles correspond to the measurements taken immediately after placing the droplet on the sample surface.

2.9. Analysis of Antibiofouling Performance of Functionalized CNT Surfaces by Fluorescence Microscopy.

The antiprotein fouling performance of nonfunctionalized and functionalized CNT films was assessed by quantifying nonspecific adsorption of fluorescently tagged proteins. Albumin-FITC and Alexa Fluor 294 goat antirabbit IgG proteins were used. The CNT films were soaked in 2 mg mL⁻¹ FITC-albumin and IgG solution for 1 and 4 h at room temperature under dark conditions. After the incubation, the CNT films were washed thrice with PBS to remove weakly attached proteins and then air-dried. Fluorescence images were captured using a Leica DMI4000 microscope (Buffalo Grove, IL) with a 3 s exposure time at 10× magnification. A total of 10–15 images were taken across two samples prepared per condition (nonfunctionalized CNT films and functionalized CNT films using one, three, and five CV cycles) at the 1- and 4-h time point for FITC-albumin and IgG. The total fluorescence intensities per unit area of the CNT film were measured using CellProfiler software (Cambridge, MA) and normalized to the nonfunctionalized condition for each time point.

2.10. Morphology Analysis.

2.10.1. Scanning Electron Microscopy (SEM).

High-resolution imaging of CNT-Fcs ME and CNT film electrode surfaces was performed using an FEI Apreo scanning electron microscope (Thermo Scientific, Waltham, MA, USA) at acceleration voltages of 5–15 kV and a current of 0.2 nA.

2.10.2. Transmission Electron Microscopy (TEM).

The TEM analysis was conducted using an FEI Talos F200X instrument (North Billerica, MA, USA) using an accelerating voltage of 200 kV. For TEM analysis, the CNT film electrodes were prepared in a similar manner described in Section 2.4. CNT films were prepared by drawing layers of CNTs from the VA-CNT array onto a TEM Cu grid instead of a PET substrate, and two layers were drawn in a crosshatch pattern instead of four layers. TEM analysis was performed on nonfunctionalized CNT film electrodes and CNT film electrodes functionalized by three CV cycles.

3. RESULTS AND DISCUSSION

3.1. Functionalization of CNT Assemblies by Electrochemical Reduction of PPC Diazonium Salt.

In this study, aryl diazonium chemistry was used to functionalize the CNTs with PC functional groups. Aryl diazonium chemistry is commonly used to introduce various functional groups containing substituted aryl rings onto surfaces. Surface modification using diazonium chemistry provides high stability because covalent bonds are formed between the surface and the substituted aryl group.^38–41 Herein, electrochemistry was used to trigger reaction between the diazonium salt and the CNTs. The electrochemical generation of reactive aryl radicals in the vicinity of CNTs leads to subsequent attachment of aryl moieties on to CNTs. In addition to aryl diazonium salts, electrochemical methods have also been successfully used to graft other types of monomers onto CNTs by generating reactive radicals. Zhang et al. reported the electrochemical functionalization of SW-CNTs with an acrylate monomer, N-succinimidyl acrylate.⁴² The authors also introduced an ionic liquid-supported three-dimensional network-based SW-CNT assembly to homogeneously functionalize SW-CNT-based electrodes to minimize localized irregular grafting. In various studies, the electrochemical grafting of aryl diazonium salts has been successfully applied to introduce important functional groups onto CNTs.^43–47 Electrochemical grafting of diazonium salts provides control over the surface modification process, resulting in minimal damage to the CNT structure, unlike to wet chemical methods.⁴⁵ 4-Aminophenylphosphorylcholine was used as a precursor to generate diazonium salt of PPC. Subsequently, the CV technique was used to generate the aryl radical of PPC, which then reacted with the CNT surface. Characteristic voltammograms were obtained for the PPC diazonium salt reduction on the CNT assemblies, as shown in Figure 2.

Figure 2.

Consecutive cyclic voltammograms for the electrochemical reduction of PPC diazonium salt on CNT assemblies: (A) CNT-Fcs ME displaying up to the 20th cycle and (B) CNT film surface displaying up to the 5th cycle in 0.25 M HCl aqueous solution containing 5 mM PPC and 5 mM NaNO₂ at 50 mV s⁻¹.

The CV curves in Figure 2A confirm the successful reduction of the PPC diazonium salt on the CNT-Fcs ME surface. During the first cycle, an irreversible reduction peak was clearly observed at −0.84 V, and the intensity of this peak gradually decreased over 20 cycles. An irreversible reduction peak during diazonium salt grafting can be attributed to the cleavage and loss of N₂.⁴¹ It is proposed that the PPC radical generated by diazonium salt reduction can attack the exposed open ends of the CNTs in the CNT-Fcs ME. The decrease in the intensity of the reduction peak observed with each subsequent cycle confirmed the binding of the PPC radical to the CNT ME, thus decreasing the reaction between the electrode surface and the diazonium salt. Consequently, electron transfer between the diazonium salt and the electrode surface becomes difficult as potential cycling continued. The reaction mechanism for electrochemically induced aryl diazonium chemistry on the CNT-Fcs ME is shown in Figure S2. A similar trend has been reported for gold, GC, and ITO surfaces.^30,31,33 Nevertheless, the presence of a reduction peak during the 20th cycle suggests that the surface was not entirely blocked by PPC molecules. Thus, this technique allows the controlled functionalization of a predetermined microsized area of CNT fibers, thereby advancing the development of implantable CNT fiber electrodes with antibiofouling properties. CNT films were also functionalized and compared with the CNT-Fcs ME characteristics.

The CNT film electrode was also functionalized using this approach, as shown in Figure 2B. In a similar manner to the CNT-Fcs ME, the PPC-functionalized CNT films showed a characteristic reduction peak at approximately −0.70 V, which decreased in intensity with potential cycling. In the film assembly, the PPC radicals could attack the open ends and sidewalls of CNTs, resulting in PPC radicals covalently attached to the CNT film surface. The difference in the reduction potentials of the PPC diazonium salt on the CNT-Fcs ME and the CNT film electrode was attributed to variations in the surface functional groups between sp² carbon on the CNT sidewalls and exposed open-ended CNTs of the two types of assemblies. In addition, the shift may be due to the intrinsic hydrophilic nature of the cross section on the CNT-Fcs ME.

3.2. Influence of the Functionalization on the Microstructure of the CNT Assemblies.

The morphological features of CNT-Fcs MEs and CNT film electrodes were analyzed by SEM and TEM. Both CNT-Fcs MEs and CNT films were functionalized with PPC by cycling the potential between 0.6 V and −1.0 V for three cycles in the diazonium salt solution as described in Section 2.5 and compared with the nonfunctionalized CNT assemblies. Figure 3A,B shows the detailed morphological features of the nonfunctionalized CNT-Fcs ME at two different magnifications. The sectioning of the polymer-coated CNT fibers using a microtome causes opening of the individual CNTs. Therefore, CNT-Fcs ME consist of millions of individual CNTs with a high density of open-ended tubes (Figure 3B). After the covalent functionalization of CNT-Fcs MEs with PPC, no apparent morphological changes were observed as shown by the SEM images (Figure 3C,D). CNT-Fcs MEs could subject to damage by prolonged application of high potentials (potentials higher than +1.5 V). However, the electrochemical grafting of the PPC was performed in a safe potential range (0.6 V to −1.0 V) where it is expected that CNTs maintain their original structure. According to Figure 3C, there was no indication of possible ME structure damage due to electrochemical grafting of PPC. Also, there was no apparent formation of polymer coating on the surface according to SEM characterization. Generally, diazonium salt reactions could be highly reactive, forming polymer layers instead of monolayers. However, no significant difference was observed between individual CNTs between nonfunctionalized and functionalized surfaces (Figure 3B,D). According to these images, the morphological effect of functionalization could be in nanometer levels. These images suggest that the functionalization might be limited to monolayers or few layers which is not possible to detect by SEM. Figure 3E,F shows the nonfunctionalized and PPC-functionalized CNT films, respectively. Similarly, no structural differences were observed after the functionalization of CNT films. The effect of functionalization might be in nanometer levels suggesting that PPC grafting by electrochemical diazonium salt reduction might generate monolayers or few layers on the CNT film surface. To assess the impact of the PPC functionalization on individual CNTs, TEM characterization was performed on functionalized CNT film electrodes (Figure S3). TEM images of CNTs in nonfunctionalized electrodes are shown in Figure S3A and S3B while the functionalized ones are shown in Figure S3C and S3D. The differences between individual CNTs were not evident after the functionalization (Figure S3A and S3C). Higher magnification TEM images were also compared in Figure S3B and S3D, but a minimum change in thickness is observed. However, focusing of the functionalized CNT film at 245,000× magnification was more difficult than usual, suggesting that functional groups may be moving; thus the CNT walls were not clear.

Figure 3.

SEM characterization of CNT-Fcs MEs and CNT film electrodes. (A) Low- and (B) high-magnification SEM images of a nonfunctionalized CNT-Fcs ME. (C) Low- and (D) high-magnification SEM images of a CNT-Fcs ME functionalized by three CV cycles. (E) SEM image of a nonfunctionalized CNT film. (F) SEM image of a CNT film functionalized by three CV cycles.

3.3. Hydrophilicity Assessment.

The CNT films were functionalized with PPC by cycling the potential between 0.6 and −1.0 V for one, three, or five CV cycles in the diazonium salt solution as described in Section 2.5. The hydrophilicity of the functionalized CNT film electrodes was evaluated using contact angle measurements. However, such an analysis was not possible for the CNT-Fcs ME because of size limitations of the fiber cross section. Figure 4 shows the effect of functionalization on the wettability of the CNT films. The densified crosshatch-patterned CNT films were inherently hydrophobic with a static contact angle of 134.4° ± 3.9° (Figure 4A). It has been reported that the wettability of aligned CNT films can be altered through chemical modification with functional groups.⁴⁸ As shown in Figure 4B, the surface of the CNT film became hydrophilic after a single CV cycle in the diazonium salt solution, with the static contact angle decreasing drastically to 15.7° ± 1.5°. Complete wetting (contact angle = 0°) was observed for the CNT films functionalized using three and five CV cycles, indicating that the maximum hydrophilicity was achieved after three CV cycles (Figure S4 and Table S1). The wettability of the CNT surface was greatly increased after introducing PPC functional groups, and the observed hydrophilicity was attributed to electrostatically induced hydration by the high density of zwitterionic functional groups grafted during the electrochemical reduction of PPC. Zwitterionic polymers such as PSB, poly(carboxybetaine), and poly(2-methacryloyloxyethyl phosphorylcholine) (PMPC) have also been used to prepare highly hydrophilic surfaces.^24,49–51 Unfortunately, zwitterionic polymers are also known to introduce a huge electrical barrier for charge transfer, thus increasing the impedance on the electrode surface.

Figure 4.

Digital images and contact angle measurements of (A) nonfunctionalized CNT film displaying almost completely spherical water droplets (inset), and (B) functionalized CNT film after a single CV cycle showing wetting of the surface (the inset displays the difficulty in determining the contact angle due to almost complete wetting). In each inset, the black surface at the bottom is the goniometer stage and the light gray surface is the material of interest. In some images, the stage is tilted to level the surface of the material, thus compensating for the unevenness introduced by the glue used for electrode fabrication.

3.4. XPS Characterization of the PPC-Functionalized CNT Surfaces.

Both the CNT-Fcs ME and CNT film electrodes were analyzed by XPS. Figure 5 shows survey scans for the nonfunctionalized and PPC-functionalized CNT assemblies. The samples were functionalized using three CV cycles. For the nonfunctionalized CNT-Fcs ME, C 1s (284.9 eV) and O 1s (532.9 eV) peaks were observed without any N or P peaks (Figure 5A). In contrast, the presence of C 1s and O 1s and traces of N 1s and P 2p was found for the functionalized CNT-Fcs ME (Figure 5B), although the N and P signals were quite weak, which may be due to both the limited size of the cross sections and the attachment of a single PPC monolayer on the surface. As shown in Figure 5C, the survey spectrum of the nonfunctionalized CNT film contained major peaks for C 1s (284.9 eV) and O 1s (532.9 eV) with a trace amount of N 1s (399.8 eV). In contrast, the survey spectrum of the PPC-functionalized CNT film exhibited C 1s (285.0 eV), O 1s (532.1 eV), N 1s (402.3 eV), and P 2p (133.3 eV) peaks (Figure 5D). The N 1s and P 2p peaks in the XPS spectrum of the PPC-functionalized CNT film were more pronounced than those in the XPS spectrum of the PPC-functionalized CNT-Fcs MEs. Nevertheless, the observation of N and P peaks for both the functionalized micro- and macroassemblies is indicative of the presence of PC groups on the CNT surfaces.

Figure 5.

XPS survey spectra of (A) nonfunctionalized and (B) PPC-functionalized CNT-Fcs MEs. The inset in (A) shows the resolved area of the nitrogen peak for the nonfunctionalized (black) and functionalized (red) CNT-Fcs ME. The inset in (B) shows the resolved area of the phosphorus peak for the nonfunctionalized (black) and functionalized (red) CNT-Fcs ME and (C) nonfunctionalized and (D) PPC-functionalized CNT films. The highlighted bands in (C) and (D) show the expected N 1s, P 2s, and P 2p binding energies for the film assemblies.

Figure 6 displays the N 1s, P 2p, and O 1s core level spectra of both PPC-functionalized CNT assemblies. Two nitrogen peaks with different intensities were observed for each functionalized CNT assembly (Figure 6A,B). Deconvolution of the observed peaks at 402.5 and 400.0 eV for the functionalized CNT-Fcs ME (Figure 6A) and at 402.7 and 399.9 eV for the PPC-functionalized CNT film (Figure 6B) suggests that R–NR₃ and –N═N− are the corresponding functional groups. It is worth noting that corresponding N 1s spectra for the nonfunctionalized CNT-Fcs ME and CNT film electrodes are noisier (Figure 6A) or almost flat (Figure 6B) suggesting their absence. Nevertheless, binding energies of N observed for the functionalized CNT-Fcs ME are in agreement with the N peaks observed for the functionalized CNT film. The nitrogen peak at a binding energy of 402.7 eV is a characteristic feature for phosphorylcholine-functionalized surfaces, and it was attributed to quaternary nitrogen such as ammonium N in the PC group.^30,31 The peak at around 399.9 eV is often attributed to azo linkage,⁵² which could result from electrochemical reduction of diazonium salt promoting azo linkage formation due to multilayer grafting.³⁰ However, this peak was also observed on the nonfunctionalized CNT film (Figure 6B) which could also be attributed to inherited N impurities on the nonfunctionalized CNT surfaces. Deconvolution of the P2p core level spectra peaks in both PPC-functionalized CNT assemblies (Figure 6C,D) suggests that the peak at 133.5 eV for the functionalized CNT-Fcs ME and 133.6 eV for the functionalized CNT film corresponds to O–P–O. As expected, phosphorus was not present on nonfunctionalized CNT assemblies; therefore their XPS spectrum shows almost flat lines compared to the functionalized surfaces, Figure 6C,D. The peaks at 133.5 and 133.6 eV are similar to literature values that are attributed to phosphorus bound to the oxygen in the PC group.^30,31 The N/P atomic ratio calculated for PPC-functionalized CNT films gave an approximately 1:1 ratio, which represents the presence of charge-balanced zwitterionic PC groups on the surface. The atomic percentages for the functionalized CNT-Fcs ME were not calculated as the signal intensities were poor.

Figure 6.

Core level of N 1s, P 2p, and O 1s XPS spectra of functionalized and nonfunctionalized CNT electrodes. N 1s spectra of (A) CNT-Fcs ME and (B) CNT film electrode. P 2p spectra of (C) CNT-Fcs ME and (D) CNT film electrode. O 1s spectra of (E) CNT-Fcs ME and (F) CNT film electrode (deconvoluted spectrum of N 1s for the nonfunctionalized CNT film and deconvoluted spectra of O 1s for the nonfunctionalized CNT-Fcs ME and CNT film are not shown in Figure 6B,E,F respectively).

O 1s core level spectra of the CNT-Fcs MEs and CNT films are shown in Figure 6E,F respectively. The O 1s spectrum of the PPC-functionalized CNT-Fcs ME was fitted with three deconvoluted peaks at 533.1, 532.3, and 530.7 eV. It is worth noting that the nonfunctionalized surface of CNT-Fcs MEs also showed peaks at 533.1 and 532.3 eV (deconvoluted peaks for nonfunctionalized CNT-Fcs ME are not shown in Figure 6E). Because both of these peaks were present on the nonfunctionalized and functionalized surfaces, the peaks were attributed to oxygen-containing functional groups readily present on the bare fiber cross section. In a previous study, Alvarez lab reported that CNT-Fcs MEs contain significantly higher levels of oxygen (10–13%) on nonfunctionalized CNT-Fcs ME.¹⁹ The high percentage of O species on the nonfunctionalized CNT-Fcs ME is attributed to microtoming CNT fibers in the open air, which may promote reactions of the freshly exposed CNT open ends with oxygen or water in the environment. The peak at 533.4 eV was attributed to the hydroxyl oxygen in carboxyl and ester groups, and the peak at 532.3 eV may correspond to carbonyl oxygen in carboxyl and ester groups or hydroxyl groups.^53,54 The peak at 530.7 eV was only observed on PPC-functionalized CNT-Fcs MEs and it was attributed to oxygen in PC. Literature reports on gold and GC electrode functionalization also assigned the peaks at 530–531 eV to new O after PC functionalization, and they were attributed to O attached to phosphorus in the PC group.^30,31 O 1s spectral results (Figure 6F) of the CNT film assemblies were correlated with the results of CNT-Fcs MEs. The oxygen peak associated with the PC group was also observed on the functionalized CNT film as expected. The peak was centered at 530.9 eV and was more pronounced compared to the PPC-functionalized CNT-Fcs MEs. Two peaks at around 533.0 and 532.7 eV were also observed on the functionalized CNT films, and they were attributed to oxygen-containing functional groups (carboxyl, carbonyl, and hydroxyl) readily present on CNTs because they were also present in the nonfunctionalized CNT film. The presence of a nitrogen peak at around 402.7 eV, a phosphorus peak at 133.6 eV, and an oxygen peak at 530.7–530.9 eV is strong evidence of PPC molecule attachment onto the CNT surface in both assemblies.

Analyzed and reported C 1s core level spectra for MWCNTs in the assemblies present some challenges due to the proximity of binding energies to different oxygen-containing carbons and the π–π* shake-up feature of CNTs with delocalized π-electron systems.⁵⁵ The main peak at 284.4–284.7 eV was observed in all the CNT assemblies, functionalized and nonfunctionalized Figure S5A–D and Table S2. The peak was attributed to sp² hybridized carbon originating from graphitic carbon on CNTs and aromatic rings in PC groups.^30,56 The peak at 284.9–285.3 eV (Figure S5C and S5D) was observed on both nonfunctionalized and functionalized CNT films, and it was attributed to sp³ hybridized carbons from defect sites in CNT sidewalls and open ends.⁵⁷ A peak was observed at 286.2–286.6 eV (C–O in ether or hydroxyl) in all CNT assemblies regardless of the functionalization state. However, this peak showed a significant increase in the PPC-functionalized film surface compared to the nonfunctionalized CNT film. In related observation, Gui et al. reported that the GC surface showed a significant increase in the carbon peak at 286.9 eV after functionalizing with PPC by aryl diazonium chemistry (from ~5 to ~20%).³⁰ Therefore, the observed C peak at 286.2 eV is most likely due to PPC on the functionalized CNT film and correlates with C–O and C–N in the PC molecular structure.^30,58,59 The peaks observed at around 287.4 eV (C═O in carbonyl) and 289.9 eV (C═O in carboxyl and ester) were attributed to surface functional groups readily present on the CNT surfaces.⁵⁷

3.5. Antibiofouling Behavior of the PPC-Functionalized CNT Surface.

To evaluate the in vitro antiprotein fouling behavior of the PPC-functionalized CNT surfaces, nonspecific adsorption of FITC-labeled albumin and IgG secondary antibody Alexa Fluor-594 was monitored using fluorescence microscopy. Albumin is the most abundant plasma protein (35–50 mg/mL), representing 50–60% of the total protein content in the plasma, while globulins constitute the remaining 40% (10–20%, immunoglobulin, IgG).^60,61 The implantation of invasive electrodes can rupture blood vessels in the brain, releasing plasma proteins such as albumin, which can then adsorb onto the electrode surface.^5,62 This nonspecific protein fouling can further mediate the attachment of inflammatory cells, such as microglia and astrocytes, onto the electrode surface and trigger a cascade of host immune responses.^3,8,9 Thus, in vitro protein adsorption assays can provide information about the antibiofouling properties of a surface and its ability to inhibit the initial response to foreign bodies. A protein adsorption study was only investigated on CNT film surfaces due to limitations associated with visualizing ME cross sections with fluorescence. Figure 7 shows the fluorescence quantification of FITC-albumin (Figure 7A–D) and IgG adsorption (Figure 7E–H) on nonfunctionalized and PPC-functionalized (1–5 cycles) CNT films. The functionalized CNT films were prepared using one, three, and five CV cycles and incubated with FITC-albumin and IgG proteins for 1 and 4 h at room temperature under dark conditions. Fluorescence intensity measurements revealed that protein adsorption was significantly lower on the CNT films functionalized using three and five CV cycles than on the nonfunctionalized CNT film after 1 h (Figure 7I; p < 0.0001 for three and five CV cycles) and 4 h (Figure 7J; p < 0.05 for three CV cycles, p < 0.0001 for five CV cycles) of incubation with FITC-albumin. IgG adsorption was similarly significantly lower on CNT films functionalized using three and five CV cycles and incubated with IgG for 1 h (p < 0.01 for three CV cycles; p < 0.001 for five CV cycles) and 4 h (p < 0.0001 for three and five CV cycles). Because artificial surfaces upon implantation in the body face a rapid adsorption of plasma proteins on the order of minutes, the 1 and 4-h timepoints should demonstrate the initial phase of biofouling in vitro.^63,64 Overall, the amount of protein adsorbed on the CNT film decreased as the number of CV cycles increased. The decrease in IgG fluorescence at the 4-h time point compared to the 1-h time point could be due to decreasing antibody stability over time at room temperature.⁶⁵ The mean fluorescence intensities are normalized to the nonfunctionalized sample to control for these variations. This improvement in the in vitro antiprotein fouling performance was attributed to the increase in PPC grafting as the number of CV cycles increased. These findings imply that electrochemical grafting can be used to control the degree of functionalization on the CNT surface. The electrografting of PPC by diazonium salt reduction allows strong covalent bond formation between the CNTs and PPC, providing an antibiofouling surface for long-term applications. The coatings produced by this approach are superior to antiprotein fouling polymer coatings developed by noncovalent functionalization. Therefore, the electrochemical characteristics of the CNT-Fcs ME and the CNT film electrode were evaluated to determine the impact of functionalization on their faradic charge-transfer properties and impedance characteristics (see Section 3.6).

Figure 7.

Fluorescence microscopic analysis of FITC-albumin adsorption on CNT films. (A) Colorized fluorescent images of FITC-albumin adsorbed on a nonfunctionalized CNT film and CNT films functionalized using (B) one CV cycle, (C) three CV cycles, and (D) five CV cycles. (E) Colorized fluorescent images of IgG adsorbed on the nonfunctionalized CNT film and CNT functionalized using (F) one CV cycle, (G) three CV cycles, and (H) five CV cycles. (I) and (F) show the quantified mean fluorescence intensity (error bars represent the standard error of the mean) for 1-h incubation and 4-h incubation, respectively, with FITC-albumin and IgG. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; n.s., not significant.

3.6. Electrochemical Characterization of PPC-Functionalized CNT Surfaces.

CV and EIS were used to analyze the charge-transfer characteristics of the functionalized CNT assemblies prepared using three CV cycles. Zwitterionic polymers, including PMPC, exhibit poor electrical conductivity due to the insulating nature of the polymer chains. Such insulation can interfere with ionic charge-transfer at the electrode–electrolyte interface, resulting in high electrode impedance. In contrast, diazonium salt grafting has been shown to be a promising approach for developing low-impedance electrodes with antibiofouling properties.^30,33

CV is a valuable tool for investigating the functionalization of electrode surfaces. Redox probes such as Ru(NH₃)₆³⁺ and Fe(CN₆)^3– are frequently used as surface-insensitive and surface-sensitive redox probes to identify electrode surface modifications.⁶⁶ The Ru(NH₃)₆³⁺ redox reaction was observed on the functionalized and nonfunctionalized CNT surfaces to evaluate the effects of PPC grafting. Figure 8A,B shows the redox behavior of Ru(NH₃)₆³⁺ on the functionalized CNT-Fcs ME and CNT film electrode, respectively. PPC functionalization of the CNT surface had little effect on the reversible Ru(NH₃)₆³⁺ redox reaction, causing insignificant changes in peak currents or peak separation in either type of assembly. Gui et al. reported similar Ru(NH₃)₆³⁺ redox chemistry for PPC-modified GC electrodes, with Ru(NH₃)₆³⁺ oxidation and reduction peak currents slightly attenuated upon PPC diazonium salt grafting.³⁰ The Ru(NH₃)₆³⁺ redox behavior can be explained by its electron transfer mechanism. Typically, Ru(NH₃)₆³⁺ acts as an outer sphere redox probe on carbon-based electrodes and is thus considered as a surface-insensitive redox probe.⁶⁶ For surface-insensitive redox probes, changes in the surface chemistry and thin monolayers (<5 Å) on the electrode surface have minor effects on the electron transfer kinetics.^30,66 Therefore, the observed results suggest that electrochemical diazonium salt grafting may have introduced a thin layer (<5 Å) of PPC onto the CNT surfaces without building up a polymeric barrier that would interfere with the Ru(NH₃)₆³⁺ electron transfer kinetics. Although Fe(CN)₆^3−/4− usually acts as a surface-sensitive redox probe and has been widely used to identify changes in surface chemistry on metaland carbon-based electrode surfaces, nonideal electrochemical behavior can occur due to the electroactive sites on CNT surfaces being blocked by species produced by ferricyanide decomposition.^67,68 The effect of PPC functionalization on Fe(CN)₆^3−/4− redox chemistry was studied using CNT-Fcs MEs, as discussed in the Supporting Information (Figure S6).

Figure 8.

Electrochemical characteristics of CNT assemblies before and after functionalization. Cyclic voltammograms of (A) CNT-Fcs ME and (B) CNT film electrodes in 0.1 M KCl aqueous solution containing for 2 mM Ru(NH₃)₆³⁺. Bode plots of (C) CNT-Fcs ME and (D) CNT film electrodes in PBS (pH 7.4). The insets show a comparison of the impedance at 1 kHz.

To further confirm the charge-transfer characteristics of the functionalized CNT assemblies, EIS measurements were performed. EIS can be used to investigate changes in charge-transfer resistance and electrode impedance following electrode surface modification. The physisorption or chemical bonding of insulating materials usually increases the resistance at the electrode–electrolyte interface. Figure 8C,D shows impedance data for the CNT assemblies in PBS (pH 7.4) before and after functionalization. The corresponding Nyquist plots are shown in Figure S7. After functionalization of the CNT-Fcs ME by electrochemical diazonium salt grafting, the impedance decreased. The observed results were reproducible and a Bode plot for a single CNT-Fcs ME is shown in Figure 8C. As EIS is an extremely sensitive technique, the experimental conditions were kept constant before and after functionalization to minimize the effect of external factors. As shown in the inset of Figure 8C, the average impedance of the functionalized CNT-Fcs MEs (0.21 ± 0.04 MΩ at 1 kHz) was considerably lower than that of the nonfunctionalized CNT-Fcs MEs (0.72 ± 0.07 MΩ at 1 kHz).

Gui et al. reported that the electrografting of PPC by diazonium salt reduction can be used to produce low-impedance antibiofouling GC surfaces. However, the charge-transfer resistance at the functionalized GC interface increases after surface functionalization.³⁰ The decrease of impedance after the functionalization of CNT assemblies with PPC is interesting and beneficial as a strategy to introduce antibiofouling properties to CNT electrodes in chronic applications without compromising the impedance. In this case, the decrease of impedance after the functionalization might be due to increased hydrophilicity of the CNTs by zwitterionic PPC molecules. PPC molecules cause electrostatically induced hydration; therefore grafting of PPC on CNTs facilitates the interaction of electrolytes with CNTs more and thus facilitates the charge transfer compared to nonfunctionalized CNT electrodes. Figure 8D shows the EIS characteristics of the CNT film electrode. Functionalization of the CNT film surface had a negligible effect on the impedance, with the average value decreasing from 134 ± 2 Ω to 127 ± 11 Ω after functionalization (inset, Figure 8D). Antifouling strategies currently used for implantable electrodes include surface modification with polyethylene glycol, hydrogels, zwitterionic polymers, naturally occurring proteins, and peptides. However, many of these strategies introduce a large resistive barrier at the electrode surface due to coverage of electroactive sites by large insulating molecules. In contrast, according to both the CV and EIS data, PPC grafting on CNT surfaces does not passivate the electroactive area of the electrodes (Table S3). The stability of the functionalized electrodes is also important in terms of consistence performance in long-term applications. The complete wetting (contact angle = 0°) of the functionalized CNT film electrodes (by three CV cycles) after 10 days soaking in PBS buffer implies the consistency of the hydrophilicity and the stability of the functionalization. In Figure S8, the stability of the functionalized CNT-Fcs MEs and CNT film electrodes was analyzed after applying 50 CV cycles in PBS buffer. The Bode plots of a CNT-Fcs ME functionalized by three CV cycles are shown in Figure S8A. The impedance of the functionalized CNT-Fcs ME shows consistency over the frequency range before and after the application of 50 CV cycles. A control experiment was performed with a nonfunctionalized CNT-Fcs ME and presented in Figure S8B. In both MEs, the impedance was consistent within the analyzed frequency range. Because EIS is an extremely surface-sensitive technique, the removal of attached molecules from the surface and surface damage/oxidation due to repetitive CV cycles can be detected. However, both nonfunctionalized and functionalized CNT-Fcs MEs showed no significant difference before and after application of 50 CV cycles. This implies that the covalent bonding of PPC and the nonfunctionalized CNT-Fcs MEs are stable after the application of 50 CV cycles. A similar behavior was observed for the functionalized and nonfunctionalized CNT films with three CV cycles, respectively (Figure S8C and S8D). The electrochemical grafting of PPC on to CNTs converts a hydrophobic surface into extremely hydrophilic surface, as demonstrated by contact angle measurements, through the attachment of a thin layer of PPC without building up a significant charge barrier. These findings will be useful for developing flexible electrodes with antibiofouling coatings for long-term applications.

4. CONCLUSIONS

In this work, novel antibiofouling and highly hydrophilic CNT films and fiber cross-sectional surfaces were prepared by covalent functionalization of CNT micro- and macroassemblies with small zwitterionic PPC molecules using an electrografting method. The CNT films became hydrophilic after functionalization with PCC resulting in a decreased static contact angle from 134.4°± 3.9° to 15.7°± 1.5° after a single cycle and 0° after three functionalization cycles. Furthermore, functionalization of the CNT film surfaces was found to greatly reduce protein fouling in vitro, as determined by fluorescence microscopy. CV experiments probing the redox reaction of Ru(NH₃)₆³⁺ revealed insignificant changes to either peak current and potential after functionalization. EIS data revealed a decrease in impedance after functionalization of the CNT surface. The observed electrochemical characteristics indicate that the sensing and charge-transfer capabilities of the CNTs were maintained after functionalization. In addition, these results suggest that the electrochemical grafting method is suitable for building up thin layers of zwitterionic molecules, which are more conductive and have lower impedance than large zwitterionic polymers. This modification approach allows the controlled functionalization of a microsized area of CNT fibers, thereby advancing the development of implantable CNT fiber MEs with antibiofouling properties.