Patterned polyaniline & carbon nanotube–polyaniline composites on silicon

Abstract

Atomic force anodisation lithography has been used to create patterned silicon oxide nanostructures on a p-type silicon (100) substrate. Modification of these oxide nanostructures either by acid treatment or self assembly of 3-aminopropyltriethoxysilane yields a chemically heterogeneous surface containing hydrophilic and hydrophobic regions allowing the site-selective deposition of thin films of conductive polyaniline. Carbon nanotubes, with high carboxylic acid functionality, were immobilised on the patterned hydrophilic regions using condensation reactions. Due to the hydrophobic nature of the nanotube sidewalls polyaniline is found to adsorb to the carbon nanotube structure and has allowed the creation of patterned carbon nanotube–polyaniline composites. Such composite materials may prove to be useful in the field of molecular electronics. These two approaches allow the construction of what are basically positive and negative resists using polymers.

---

Introduction

Since their discovery π-conjugated organic polymers, otherwise known as conducting polymers, have generated enormous interest due to their unique electronic properties. These properties provide great potential application in the semiconductor industry such as the development of cheap, flexible, all plastic, microelectronic devices.^1–3 Such devices will utilise the metal–-insulator transition induced by doping and un-doping of the polymer, controllable optoelectronic properties through molecular design and the many attractive mechanical properties and processing advantages available.⁴ Among the most commonly studied conjugated polymer is polyaniline due to its environmental stability and the ease with which changes in its conductivity can be obtained through proton doping, humidity, or redox reactions.⁵ Several devices have already been fabricated. For example, the un-doped, semi-conductive form of polyaniline known as emeraldine base has been used in the fabrication of light-emitting diodes and transistors^1,5 whilst the doped, conductive, emeraldine salt form is being investigated as potential interconnects, wires and conductive channels.¹

However, for many of these applications, in particular for the use as active elements in circuits, the lateral control of formation or deposition on the nano-scale of the conductive polymers is essential.^1,6 The ability of these materials to be simply and conveniently processed into specific structures is one of the major limitations for device fabrication and remains a challenge in the field of nanotechnology.^7,8 Several approaches to this problem have already been presented and typically rely upon surface templates or molecular matrices of predetermined structure.^1,7 Chi et al. have used a combination of micro-contact printing with a lift-off process to create a high density of polypyrole wires.⁴ Yang et al. have used a photomask and ultra-violet radiation to create a patterned polymer surface, to which polyaniline would preferentially deposit on the un-irradiated regions.⁹ Similarly, Whitesides et al. have selectively deposited polypyrole and polyaniline to patterned regions of octadecyltrichlorosilane on glass by a polydimethylsiloxane stamp.¹ Common to each of these techniques is that they rely upon patterned surfaces with different physical and chemical properties in different locations to control the nucleation and deposition of polymer. One property that is proving to be particularly promising in the fabrication of patterned conductive polymers is the surface free energy, which is controlled by the terminal group of a self-assembled monolayer.¹⁰ It has been demonstrated that for a hydrophobic surface, that the rate of deposition of conducting polymer is high but the adhesion is weak while for a hydrophilic surface, the rate of deposition is low but the adhesion is strong.^1,9–11

Conducting polymer composites, in particular with carbon nanotubes, have also received a great deal of attention^12–15 as a result of the exceptional electronic and mechanical properties of nanotubes.^16,17 Fabrication of hybrid materials in this way provides a possible avenue to enhance the characteristics of materials as the composite material often exhibits characteristics different to the individual components.¹² In the case of carbon nanotube–polyaniline composites it is expected that the electrical and mechanical properties may be able to be significantly improved compared to pure polymer films.^13,15,18 An order of magnitude increase in electrical conductivity of a carbon nanotube–polyaniline composite compared to pure polyaniline has already been demonstrated by Maser et al.¹⁴ Furthermore, a suggested mechanical improvement of carbon nanotube–polymer composites is that the nanotubes in the polymer could act as a nanometric heat sink, preventing device break down.^13,15

As with pure polymer films, the ability to create controlled surface architectures of composite carbon nanotube–polymer films remains a challenge. A large body of work exists where carbon nanotube–polymer films have been spin-coated, cast or electrochemically deposited onto a surface.^19–21 However, work directed towards the development of patterned composite surface architectures remains in the initial stages. Blanchet et al.²² have printed a 500 µm wide carbon nanotube–polyaniline structure via laser ablation of a thin composite layer coated on a metal support. Similarly Chen et al.²³ have used a laser-assisted photo thermal imprinting method to pattern micro-features into a carbon nanofibre reinforced polyethylene composite. Unfortunately, whilst these techniques are capable of producing arrays of carbon nanotube–polymer composites on designed substrates in a controllable manner, they require high energy processes and are only capable of producing macroscale features.

This work attempts to address this issue, providing a novel approach for the fabrication of patterned polyaniline and carbon nanotube–polyaniline composites on the nanoscale on a silicon substrate utilising a low energy, solution chemistry approach. A scanning probe method called atomic force anodisation lithography is used to create regions of silicon oxide in an insulating self-assembled monolayer of hydrophobic, –CH₃ terminated, hexadecyltrichlorosilane (HTS) on silicon substrates. The silicon oxide regions are then treated with both piranha solution and 3-aminopropyltriethoxysilane (APS) to yield hydrophilic, –OH or –NH₂ terminated regions. As has been seen previously in literature^1,9 polyaniline is then able to be selectively adsorbed to the hydrophobic regions to create controllable patterns. Instead of directly depositing polyaniline, carbon nanotubes can also be chemically attached to the patterned regions as has been shown in our previous work.^24–27 Single-walled carbon nanotubes, with high carboxylic acid functionality can be immobilised to the –OH and –NH₂ regions using condensation reactions to create ester and amide linkages, respectively. As the sidewalls of carbon nanotubes are hydrophobic this has the effect of removing the patterned hydrophilic/hydrophobic nature of the silicon surface as the hydrophilic region is now covered in nanotubes. However, upon deposition of polyaniline the strength of polymer adhesion to the carbon nanotube is found to be greater than to the hexadecyltrichlorosilane layer and the polymer on the hydrophobic layer can be preferentially removed with adhesive tape. Upon removal of surrounding polyaniline on the hexadecyltrichlorosilane layer an area selective carbon nanotube–polyaniline composite is created.

Materials & methods

Atomic force microscopy

Atomic force microscope images were taken in air with a multi-mode head and Nanoscope IV controller, Digital Instruments, Veeco, Santa Barbara, operating in tapping mode. Commercially available silicon cantilevers (FESP-ESP series, Veeco probes, Santa Barbara) with fundamental resonance frequency between 70 and 85 KHz were used. Topographic (height) and amplitude images were obtained simultaneously at a scan rate of 1 Hz with the parameters set point, amplitude, scan size and feedback control optimised for each sample. Anodisation lithography was conducted in tapping mode with platinum–iridium-coated silicon cantilevers (SCM-PIT series, Veeco probes, Santa Barbara) with fundamental resonance frequency between 60 and 75 KHz. Precise movement of the cantilever was controlled using scripts written in Microsoft Visual C++ v6.0 (Microsoft Corporation) utilising the Nanolithography software v5.1.2 (Veeco, Santa Barbara). An applied cantilever voltage of −11 V with a tip velocity of 1 µm s⁻¹ was used. Prior to patterning the amplitude set point was reduced to 70% of its imaging value to reduce the tip–substrate distance. Atmospheric humidity levels ranging from 30 to 45% (as obtained from the Australian Bureau of Meteorology) were used for patterning. All images presented are background subtracted data (using the flatten feature in the DI or WSxM²⁸ software).

Fabrication of patterned hydrophobic/hydrophilic silicon surfaces

The process to create patterned hydrophobic/hydrophilic silicon surfaces is shown schematically in the first five steps of Fig. 1 and has been described in detail previously.^24,27 In brief, highly boron doped p-type silicon (100) (Virginia Semiconductor, Inc. USA) was cut into 1 cm × 1 cm sized wafers and cleaned with a 1 : 3 (v/v) mixture of 30% H₂O₂ (Chem-Supply, South Australia) and 98% H₂SO₄ (Labscan Asia Co. Ltd, Thailand), at 80 °C for 15 minutes to yield a hydroxylated silicon surface. After cleaning, the silicon wafers were rinsed with HPLC grade water produced by A MilliQ Plus system (Millipore), hereafter referred to as MilliQ water. Before self-assembled monolayer formation, the silicon substrate was thoroughly blown dry under a stream of ultra-pure nitrogen. The hydroxyl-terminated silicon wafers were then immersed into a 0.5% solution of 90% hexadecyltrichlorosilane (Fluka Production GmbH) in 99% hexadecane (Sigma-Aldrich) for 40 minutes. After removal from solution the silicon wafers were rinsed in chloroform (Ajax-Finechem) and allowed to dry. Atomic force anodisation lithography was then used to selectively remove the hexadecyltrichlorosilane layer, creating regions of silicon oxide on the surface. The now patterned silicon substrate was then placed into a 1 : 3 (v/v) mixture of 30% H₂O₂ and 98% H₂SO₄ for 3 minutes at 80 °C to produce silanol groups on the oxide. The substrate was then washed with MilliQ water and thoroughly dried with nitrogen. This allowed a silicon surface with patterned –CH₃ and –OH surface groups to be created. These surface hydroxyl groups also provided the point of attachment for 3-aminopropyltriethoxysilane and hence enable the creation of patterned –NH₂ regions. The substrate was then placed into a 0.5% solution of 99% 3-aminopropyltriethoxysilane (Sigma-Aldrich) in 99% hexadecane (Sigma-Aldrich) for 40 minutes. The substrate was then removed from solution, washed in chloroform and allowed to dry.

Fig. 1 Creation of patterned hydrophobic/hydrophilic silicon to allow the selective deposition of polyaniline.

Surface polymerisation of polyaniline (PANI) films

The preparation of polyaniline films has been extensively reported in the literature.^1,9,29,30 Prior to use, aniline (Unilab) was purified by distillation under vacuum and each solution was freshly made. Firstly the patterned substrate was immersed in an aqueous solution containing 1 mL of purified aniline in 50 mL of 1M HCl (Ajax-Finechem). A second solution containing 0.575 g of 99% ammonium persulfate (Analar) in 25 mL of 1M HCl was added to initiate polymerisation. The solutions were briefly stirred to ensure complete mixing and then allowed to rest and polymerisation took place at 18–25 °C. After 50 minutes the substrate was removed and placed into an aqueous solution containing 2 mL of aniline in 1M HCl for 30 minutes to reduce the oxidation state of the deposited polyaniline from the pernigraniline to the emeraldine.^1,9 Finally the substrate was placed into a 1M HCl solution for 5 minutes to ensure complete protonation of the polyaniline. The sample was removed from the solution and thoroughly dried under nitrogen. As shown schematically in the final step of Fig. 1, this results in the creation of a patterned polyaniline film.

Assembly of carbon nanotube–polyaniline composites

As we have previously demonstrated^24,26,27,31 single-walled carbon nanotubes can be immobilised on patterned silicon substrates both directly, using an ester linkage, and with the use of amine-terminated 3-aminopropyltriethoxysilane as a molecular anchor to create amide linkages and these processes are shown schematically in Fig. 2. Single-walled carbon nanotubes were functionalised with carboxylic acid groups using a 3 : 1 v/v solution of 98% H₂SO₄ and 70% HNO₃ under ultrasonication. For an 8 hour acid treatment carbon nanotubes are expected to have a log normal distribution of lengths with an average length of 380 nm and a maximum length of 850 nm.^26,32 For attachment using 3-aminoproplytriethoxysilane a patterned –CH₃–NH₂ substrate was prepared and then placed into a 5 mL 99.9% N,N-dimethyl formamide (DMF) (Southern Cross Scientific Pty. Ltd) solution containing both 2.5 mg of 99.0% N,N′-dicyclohexylcarbodiimide (DCC) (Fluka Production GmbH) and 1 mg of functionalised carbon nanotubes. Finally, the silicon wafer was rinsed in acetone and allowed to dry. To directly attach carbon nanotubes to silicon, a patterned silicon wafer was immersed in a 1 : 1 : 5 solution of 30% NH₄OH (Sigma-Aldrich), 30% H₂O₂ (Chem-Supply, South Australia) and MilliQ water for 15 minutes at 80 °C. The silicon substrate was then rinsed with MilliQ water and thoroughly blown dry under a stream of ultra-pure nitrogen prior to immersion into a 1 : 1 : 5 solution of 36% HCl (Ajax-Finechem), 30% H₂O₂ (Chem-Supply, South Australia) and MilliQ water for 15 minutes at 80 °C. The silicon wafer was then once again rinsed with MilliQ water and dried with nitrogen prior to exposure to carbon nanotube solution. The now hydroxyl-terminated silicon wafer²⁵ was incubated at 80 °C in a 5 mL 99.9% dimethyl sulfoxide (DMSO) (Sigma-Aldrich) solution containing both 2.5 mg of 99.0% N,N′-dicyclohexylcarbodiimide (DCC) (Fluka Production GmbH) and 1 mg of functionalised carbon nanotubes. Finally the silicon wafer was rinsed in acetone and allowed to dry. Prior to use, both of the carbon nanotube solutions were ultrasonicated for 5 hours to evenly disperse the carbon nanotubes within solution. The attachment of carbon nanotubes typically took 8 hours.

Fig. 2 Patterned assembly of single-walled carbon nanotubes on silicon allows the creation of patterned carbon nanotube–polyaniline composites.

The patterned carbon nanotube substrates were then immersed in the polyaniline polymerisation solution for a period of 50 minutes and the same procedure followed as previously described. After deposition of polyaniline a piece of adhesive tape (Highland Invisible Tape, 3M, USA) was firmly pressed to the silicon surface to ensure homogenous contact before being slowly peeled off, hence creating the patterned carbon nanotube–polyaniline composite upon removal of polyaniline from surrounding hydrophobic regions. Most polyaniline in the hexadecyltrichlorosilane regions would come off after a single adhesive tape peel but as the repeated application of adhesive tape showed no degradation of the patterned carbon nanotube–polyaniline structure this was repeated several times to ensure complete removal.

Ultraviolet-visible spectroscopy (UV-Vis)

UV-Vis spectra were obtained from a Cary 50 Scan UV-Visible spectrophotometer with an incorporated xenon flash lamp, Varian, Inc. (USA). Dual beam mode was utilised with a scan rate of 60 nm min⁻¹ and resolution of 4 nm at a temperature of 20 °C. Polyaniline was deposited onto a pathology grade glass microscope slide (thickness 1.0–1.2 mm) by immersion in the polymerisation solution for a period of 50 minutes. By then placing the glass slide into a 5% ammonium hydroxide (Sigma-Aldrich) solution for 10 minutes the emeraldine salt was de-protonated to the non-conducting emeraldine base.²⁹ Hence spectra could be obtained for both the emeraldine salt and emeraldine base forms of polyaniline.

4-Point probe

4-Point probe measurements were conducted using a Keithley 2425 100 W source metre, Keithley Instruments, Inc., USA operating in 4 wire sense mode. Spring-loaded gold probes arranged in a square array with a spacing of 1 cm and were then gently applied to the polyaniline film. A current of 100 µA was then fed through two adjacent probes and voltage generated across the other two was measured.

Results & discussion

Effect of surface functionality on polyaniline deposition

To investigate the effect that different terminal surface groups had on the rate of polyaniline deposition on the silicon substrate, surfaces covered with either a monolayer of hexadecyltrichlorosilane (CH₃ terminus), 3-aminopropyltriethoxysilane (NH₂ terminus) or hydroxyl groups (OH terminus) were placed into the polymerisation solution for periods of 5, 10, 15, 20, 30 and 50 minutes. The time quoted is exposure time to the two polymerisation solutions. The deposition of polyaniline was monitored with atomic force microscopy by applying bearing analysis in the DI software to obtain the percentage of a 25 µm² area covered by polyaniline. Fig. 3 shows the surface coverage for the different terminal groups with increased polymerisation time. It can be seen that for –CH₃ terminated hexadecyltrichlorosilane, that polymerisation begins almost immediately and after 30 minutes the polyaniline surface coverage approaches 100%. For the –NH₂ terminated 3-aminopropyltriethoxysilane layer and the –OH terminated silicon substrates, initiation of polymerisation onto the surface lagged behind that of the –CH₃ layer by approximately 5 minutes and only reached a maximum surface coverage of 60 and 34%, respectively. Yang et al.⁹ and Whitesides et al.¹ have performed similar experiments where the adsorption of polyaniline to hydrophilic and hydrophobic polymer and glass surfaces was investigated using UV-Vis spectroscopy and scanning electron microscopy, respectively. They found a similar time lag of 5–6 minutes but it was also observed that polyaniline would eventually polymerise to form a complete layer on hydrophilic surfaces. The tendency for aniline to polymerise onto hydrophobic surfaces is attributed to the “like dissolves like” principle³³ where the aniline monomer, which is hydrophobic, is preferentially adsorbed from the aqueous solution onto a hydrophobic surface, thus favouring more rapid polymerisation. Using this principle and examining the static water contact angles of the –CH₃, –NH₂ and –OH terminated substrates, which were found to be 115°, 50° and 0°, respectively, it can be seen why the surface coverage of polyaniline is highest for –CH₃ and lowest for –OH terminated silicon.

Fig. 3 Polyaniline surface coverage on –CH₃, –NH₂ and –OH terminated silicon with increased polymerisation times.

Electronic structure of polyaniline films

Two common techniques used to characterise polyaniline films are UV-Vis spectroscopy and 4-point probe measurements. UV-Vis spectroscopy is able to provide information about the π conjugation along the polymer backbone and hence conduction electron mobility,¹ allowing the conducting and non-conducting forms of polyaniline to be easily distinguished. 4-point probe measurements are then able to provide accurate information on the sheet resistance of the deposited films. After removal of the substrate from the polymerisation solution, it was coated in a thin transparent green layer, which is characteristic of the protonated, conductive form of polyaniline, emeraldine salt, which upon immersion in 5% ammonium hydroxide could be converted into blue, non-conducting, emeraldine base.²⁹ The UV-Vis spectra of both adsorbed emeraldine salt and emeraldine base is shown in Fig. 4. The spectra for the un-protonated emeraldine base shows two absorption peaks at 340 and 602 nm, which are due to the π – π* transition of the benzenoid rings and the exciton absorption of the quinoid rings, respectively.^30,34,35 In the case of the protonated emeraldine salt, three absorption peaks at 360, 430 and 796 nm are distinguishable with the absorption at 602 nm observed for the emeraldine base disappearing, indicating a protonation of the imine sites.^30,34–36 The peak at 340 nm for the emeraldine base is red shifted⁹ to 360 nm and the new peaks at 430 and 796 nm are related to a doping level and the formation of a polaron band transition.^5,34 Furthermore the peak at 430 nm is a result of a polaron–bipolaron transition.^9,34 These observations are in agreement with the literature^5,7,9,34,35 and confirm the formation of conductive polyaniline.

Fig. 4 UV-Vis spectra of polyaniline films deposited for 50 minutes in the a) non-conducting and b) conducting forms. (spectra offset for clarity).

Upon performing 4-point probe measurements on a polyaniline film deposited on a hexadecyltrichlorosilane-coated silicon substrate a sheet resistance for the polyaniline could be determined using eqn (1) where R_s is the sheet resistance (Ω cm⁻¹), I is the current fed in through any two adjacent probes (A) and V the voltage measured across the other two (V).³⁷ It was then possible to determine the conductivity, using eqn (2), where σ is the conductivity (S cm⁻¹) and d is the film thickness (cm).³⁷

A series of conductivity measurements were taken at different locations on the substrate, where the supplied current to adjacent probes was maintained at 100 mA and using the measured thickness of the deposited polyaniline film as obtained from AFM (73 nm), the conductivities could be calculated and are summarised in Table 1. The average conductivity of the deposited polyaniline film was found to be 12.69 S cm⁻¹. This compares well to the literature^{9,30,35,38,39} where the conductivity of deposited polyaniline films have been found to range from 0.3 to 21 S cm⁻¹. Whitesides et al.¹ have reported a difference in the conductivity of polyaniline films dependant upon the surface functionality to which it is deposited. In this work due to the incomplete formation of polyaniline films on the –NH₂ and –OH terminated substrates it is difficult to comment on whether the surface functionalities might be effecting the conductivities observed. However, Yang et al.⁹ have provided evidence that surface functionality has no effect on the conductivity of polyaniline films. It is therefore believed that the polyaniline deposited in the case of the carbon nanotube–polyaniline composite material will exhibit a similar conductivity of ∼13 S cm⁻¹. The effect on conductivity of placing carbon nanotubes within this polyaniline film is at present unknown, however, based on the work by Maser et al.¹⁴ a conductivity of up to ∼130 S cm⁻¹ could be expected.

Table 1 4-Point probe conductivity measurement of polyaniline deposited on a silicon substrate with hexadecyltrichlorosilane monolayer

Supplied Current, I/A	Measured voltage, V/V	Sheet resistance, Rs/Ωcm^−1	Conductivity, σ/S cm^−1
1 × 10^−4	0.170	0.1078	8.89
	0.098	0.0622	15.43
	0.110	0.0698	13.75
		Average	12.7 ± 3.4

---

Patterned polyaniline

Fig. 5 shows an AFM image of a patterned silicon substrate from anodisation lithography before (a) and after (b) the deposition of polyaniline for 50 minutes. Prior to polyaniline deposition a silicon oxide pattern, approximately 2–5 nm in height, surrounded by a monolayer of hexadecyltrichlorosilane can be seen. This silicon oxide pattern is then treated with 3-aminopropyltriethoxysilane or piranha solution to yield the patterned hydrophobic/hydrophilic surface. After deposition, polyaniline is seen to selectively adsorb to the surrounding hydrophobic hexadecyltrichlorosilane layer. Unlike in the case where a substrate with a monolayer of –NH₂ or –OH terminal groups was placed into polyaniline solution, no adsorption in the nanoscale hydrophilic patterned regions is observable, even when left in the polymerisation solution for 3 hours or upon successive depositions. This is contrary to the literature where polyaniline has been found to deposit onto large macroscale hydrophilic surfaces.^1,9 This difference is likely due to the decreased size of the patterned area in this work to yield a nanoscale heterogeneous surface. The very small hydrophilic regions may mean preferential adsorption of monomer to, initially, the uncoated hexadecyltrichlorosilane layer or later further polymerisation with regions of the surface already polymer coated, instead of to the hydrophilic –NH₂ and –OH terminated region. After some time, the ability of the monomer to get to the bottom of a reasonably deep canyon in the midst of material that the monomer would easily react with would be limited. The fact that no adsorption is observed after many hours may also indicates that the nanoscale features are simply not large enough for effective polymerisation and even if growth starts in the hydrophilic areas, the small polymer chains formed do not provide effective adhesion and the chain migrates to the growing polymer layer on the hydrophobic regions.

Fig. 5 AFM image of a patterned silicon substrate by anodisation lithography (a) before and (b) after deposition of polyaniline.

It would be interesting to probe the critical minimum dimensions that would allow growth in the channels. However, the geometry of the AFM probe limits the maximum lateral dimensions of the water meniscus critical in anodisation lithography and this limits the lateral size of features which can be drawn using this technique to the order of 300 to 400 nm. The feature in Fig. 5 has lateral dimensions near the maximum that can be drawn with anodisation lithography. Larger features with dimensions between 500 nm and a few microns could be formed using other approaches and this would elucidate the critical dimensions required for polymer growth.

Upon taking a cross-section of the patterned area after polyaniline deposition it can be seen that the polymer film has a thickness of 73 nm, which is in agreement with the literature value of 75 nm⁹ where the same approach to polymerisation was used to create macroscale regions of polyaniline. It should also be noted that the slight ‘bump’ at the base of the cross-section is from the silicon oxide formed during surface patterning. The dimensions of the polyaniline lines created in this work are 300 nm in width, spaced approximately 4 µm apart, which is about to 10 times narrower than the smallest features produced in previous work^1,9 where features of between 2–82 µm were fabricated. Furthermore, atomic force anodisation lithography has been shown to produce features with lateral dimensions below 100 nm,⁴⁰ hence this approach may provide a means to produce high resolution polyaniline wires by ‘tuning’ the line width and spacing of the patterned oxide, which can be used to define polyaniline wires and this would have many applications in electronic device manufacture.

Patterned carbon nanotube–polyaniline composites

Fig. 6 shows AFM images of each of the surface treatment steps involved in the fabrication of the patterned carbon nanotube–polyaniline composite material beginning with a patterned silicon substrate (Fig. 6a). Previously we have reported the attachment of carbon nanotubes to silicon both directly and with the use of 3-aminopropyltriethoxysilane.^24–27 After attachment (Fig. 6b) it can be seen that the carbon nanotubes are vertically aligned forming a distribution of heights and that there is no adsorption to the –CH₃ terminated silane layer. A cross-section of the immobilised nanotube region shows the underlying silicon oxide stripe, which is approximately 5 nm in height, with a vertically aligned carbon nanotube attached. The length of attached carbon nanotubes is significantly less than would be expected with most having lengths between 8 and 20 nm with a maximum height of 57 nm.²⁴ It has been shown in the literature by Liu et al.^41,42 that adsorption kinetics facilitate the assembly of shorter carbon nanotubes over their slower moving longer counterparts. Whilst the observed length for the carbon nanotube is quite short, this is in agreement with the literature^24,26,41,42 where predominately short carbon nanotubes are found to chemically attach to form vertically aligned arrays. When polyaniline is deposited onto this patterned array of carbon nanotubes it is found to cover the entire surface (Fig. 6c). This can be explained by considering the structure of the immobilised carbon nanotubes. Due to the hydrophobic nature of the carbon nanotube sidewalls it is possible for the aniline monomer to adsorb to, and initiate polymerisation on, the nanotube.³³ Essentially the deposition of carbon nanotubes has had the effect of removing some of the heterogeneous hydrophilic/hydrophobic nature of the patterned substrate. Surprisingly upon removal of the polyaniline from the substrate, polyaniline was found to remain on the patterned region containing carbon nanotubes (Fig. 6d). This can be explained by further consideration of the carbon nanotube structure. Due to the concentrated acid treatment, the end caps and sidewalls of the carbon nanotube are decorated with hydrophilic, oxygen-containing functional groups such as carboxylic acids.^24–26,43 It is the presence of hydrophilic regions in the form of both carboxylic acid groups on the carbon nanotube along with the underlying –NH₂ or –OH pattern to which the nanotubes are attached that explains why adhesive tape can be used to selectively remove the polyaniline from the hexadecyltrichlorosilane layer to leave the patterned composite. It has been shown in the literature^1,9,11 that the adhesion strength of polyaniline is predominately determined by the surface free energy. High energy, hydrophilic surfaces have been seen to promote adhesion while low energy hydrophobic surfaces remain passive.¹¹ This can be examined quantitatively for our system using the Young–Dupre equation:⁹

ΔG^a = −γ(1 + cosθ) (3)

where ΔG^a is the adhesion free energy (N m⁻¹), θ is the static water contact angle on the surface and γ is the water surface tension (N m⁻¹), which is assumed to be 0.0720 N m⁻¹.¹¹Table 2 summaries the static water contact angle for unpatterned silicon with carbon nanotubes attached directly and with a 3-aminopropyltriethoxysilane linkage along with surfaces containing purely –NH₂, –COOH,⁴⁴ –OH and –CH₃ functionalities. The adhesion free energy of each of these surfaces has been calculated and it can be seen that there is a significant decrease in the adhesion free energy for each surface compared to the –CH₃ terminated surface from which the polyaniline can easily be removed. The carbon nanotube should be thought of as a hydrophobic surface to which the aniline monomer can adsorb. However the nanotubes also contain many small hydrophilic regions mainly near the nanotube ends to facilitate adhesion. Indeed contact angle measurements on an unpatterned nanotube surface show that it is quite hydrophilic (see Table 2). The growth of the polyaniline is likely to initially be on top of the –CH₃ terminated surface with growth up the largely hydrophobic walls of the nanotubes. After long enough growth the entire nanotube structure is covered including the ends which contain a significant hydrophilic character due to the presence of the carboxylic acid groups. It is this region of the nanotubes that provide the stronger adhesion to the polyaniline allowing differential removal of the polymer from the hydrophobic substrate without removal from the nanotubes.

Table 2 Adhesion free energy for different surface functionalities

Surface group	Water contact angle/θ	Adhesion free energy, ΔG^a/N m^−1)
CH3	115	−0.042
APS attached nanotubes	75	−0.091
NH2	50	−0.118
Directly attached nanotubes	13	−0.142
COOH^44	10	−0.143
OH	0	−0.144

---

Fig. 6 AFM images of the various steps involved in the formation of carbon nanotube–polyaniline composites.

Examination of the carbon nanotube–polyaniline composite material shows that polyaniline now covers the entire patterned region and is coated around the vertically aligned carbon nanotubes. This is demonstrated by comparing a cross-section of the composite material to the cross-section before polymer deposition. The vertical height has remained almost the same, however the width has increased and the silicon oxide structure is no longer visible due to the deposition of polymer. To our knowledge this is the first example of a solution chemistry based approach to fabricate nanoscale carbon nanotube–polyaniline structures.

The two approaches to attaching polymer to this patterned substrate has in effect provided a simple approach using polymers to make both positive and negative resists of the nanoscale features created. Further deposition of a variety of materials to yield novel and ultimately important structures is certainly feasible.

Conclusion

Atomic force microscopy has been used to investigate the rate of polymerisation of polyaniline and the degree of surface coverage for various silane coated silicon substrates. Polyaniline, which was found to have a conductivity of 12.69 S cm⁻¹ preferentially adsorbed and formed higher quality layers on a hydrophobic, hexadecyltrichlorosilane surface compared to the hydrophilic 3-aminopropyltriethoxysilane and hydroxyl-terminated substrates. Atomic force anodisation lithography was used to create patterned hydrophilic regions on a hydrophobic silicon substrate yielding a spatially heterogeneous surface in terms of chemical functionality allowing the patterned deposition of polyaniline. Due to the high resolution achievable with atomic force anodisation lithography this method has allowed the fabrication of nanostructures up to 10 times smaller than the smallest features produced in previous work. This technique therefore may provide a straightforward alternative approach to fabricate electronic wires and contacts from conductive polymers. To these patterned hydrophilic regions single-walled carbon nanotubes were also immobilised. Due to the hydrophobic side walls and high carboxylic acid functionality, polyaniline was able to polymerise around the carbon nanotube forming a composite material. This composite material may prove to be of particular importance to the electronics industry. The polymer deposition approaches allow the construction of what are basically nanoscale positive and negative resists.

References

1. Z. Huang, P.-C. Wang, A. G. MacDiarmid, Y. Xia and G. M. Whitesides, Langmuir, 1997, 13, 6480.
2. J. Z. Wang, Z. H. Zheng, H. W. Li, W. T. S. Huck and H. Sirringshaus, Syth. Met., 2004, 146, 287.
3. E. Menard, M. A. Meitl, Y. Sun, J.-U. Park, D. J.-L. Shir, Y.-S. Nam, S. Jeon and J. A. Rogers, Chem. Rev., 2007, 107, 1117.
4. B. Dong, N. Lu, M. Zelsmann, N. Kehagias, H. Fuchs, C. M. Sotomayor Torres and L. Chi, Adv. Funct. Mater., 2006, 16, 1937.
5. R. Sfez, L. De-Zhong, I. Turyan, D. Mandler and S. Yitzchaik, Langmuir, 2001, 17, 2556.
6. L. F. Rozsnyai and M. S. Wrighton, Langmuir, 1995, 11, 3913.
7. M. Mazur and P. Predeep, Polymer, 2005, 46, 1724.
8. P. T. Hurley, A. E. Ribbe and J. M. Buriak, J. Am. Chem. Soc., 2003, 125, 11334.
9. P. Yang, J. Xie and W. Yang, Macromol. Rapid Commun., 2006, 27, 418.
10. L. Niu, R. M. Latonen, C. Kvarnstrom and A. Ivaska, Electrochim. Acta, 2004, 49, 4455.
11. L. F. Rozsnyai and M. S. Wrighton, Am. Chem. Soc., 1996, 8, 309.
12. G. M. Do Nascimento, P. Corio, R. W. Novickis, M. L. A. Temperini and M. S. Dresselhaus, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 815.
13. P. C. Ramamurthy, A. M. Malshe, W. R. Harrell, R. V. Gregory, K. McGuire and A. M. Rao, Solid-state Electron., 2004, 48, 2019.
14. M. Cochet, W. K. Maser, A. M. Benito, M. A. Callejas, M. T. Martinez, J.-M. Benoit, J. Schreiber and O. Chauvet, Chem. Commun., 2001, 1450.
15. P. C. Ramamurthy, W. R. Harrell, R. V. Gregory, B. Sadanadan and A. Rao, M. Synth. Met., 2003, 137, 1497.
16. J. Liu, M. J. Casavant, M. Cox, D. A. Walters, P. Boul, W. Lu, A. J. Rimberg, K. A. Smith, D. T. Colbert and R. E. Smalley, Chem. Phys. Lett., 1999, 303, 125.
17. Y. Joon Jung, B. Wei, R. Vajtai, P. M. Ajayan, Y. Homma, K. Prabhakaran and T. Ogino, Nano Lett., 2003, 3, 561.
18. J. Deng, X. Ding, W. Zhang, Y. Peng, J. Wang, X. Long, P. Li and A. S. C. Chan, Eur. Polym. J., 2002, 38, 2497.
19. S. Yun and J. Kim, J. Phys. D: Appl. Phys., 2006, 39, 2580.
20. X. Ma, X. Zhang, Y. Li, G. Li, M. Wang, H. Chen and Y. Mi, Macromol. Mater. Eng., 2006, 291, 75.
21. J. Huang, X. Li, J. Xu and H. Li, Carbon, 2003, 41, 2731.
22. G. B. Blanchet, C. R. Fincher and F. Gao, Appl. Phys. Lett., 2003, 82, 1290.
23. Y. Lu, D. B. Shao and S. C. Chen, Appl. Phys. Lett., 2004, 85, 1604.
24. B. S. Flavel, J. Yu, J. G. Shapter and J. S. Quinton, Carbon, 2007, 45, 2551.
25. J. Yu, J. G. Shapter, J. S. Quinton, M. R. Johnston and D. A. Beattie, Phys. Chem. Chem. Phys., 2007, 9, 510.
26. J. Yu, D. Losic, M. Marshall, T. Bocking, J. J. Gooding and J. G. Shapter, Soft Matter, 2006, 2, 1081.
27. B. S. Flavel, J. Yu, J. G. Shapter and J. S. Quinton, J. Mater. Chem., 2007, 17, 4757.
28. I. Horcas, R. Fernandez, J. M. Gomez-Rodriguez, J. Colchero, J. Gomez-Herrero and A. M. Baro, Rev. Sci. Instrum., 2007, 78, 13705.
29. J. Stejskal and R. G. Gilbert, Pure Appl. Chem., 2002, 74, 857.
30. Z. F. Li and E. Ruckenstein, Macromol., 2002, 35, 9506.
31. J. Yu, J. G. Shapter, M. R. Johnston, J. S. Quinton and J. J. Gooding, Electrochim. Acta, 2007, 52, 6206.
32. M. Marshall, S. Popa-Nita and J. G. Shapter, Carbon, 2006, 44, 1137.
33. Z. Huang, P.-C. Wang, J. Feng, A. G. MacDiarmid, Y. Xia and G. M. Whitesides, Synth. Met., 1997, 85, 1375.
34. M. Seong Cho, S. Young Park, J. Yeon Hwang and H. Jin Choi, Mater. Sci. Eng. C, 2004, 24, 15.
35. E. Ruckenstein and Z. F. Li, Adv. Colloid Interface Sci., 2005, 113, 43.
36. N.-R. Chiou and A. J. Epstein, Adv. Mater., 2005, 17, 1679.
37. L. I. Maissel and R. Glang, Handbook of Thin Films Technology, McGraw-Hill, New York, NY, USA, 1970, ch. 13, p. 5.
38. Z. F. Li and E. Ruckenstein, Synth. Met., 2002, 129, 73.
39. K. S. Lee, G. B. Blanchet, F. Gao and Y.-L. Loo, Appl. Phys. Lett., 2005, 86 074102.
40. Y. Kim, I. Choi, S. Koo Kang, K. Choi and J. Yi, Microelectron. Eng., 2005, 81, 341.
41. Z. Liu, Z. Shen, T. Zhu, S. Hou and L. Ying, Langmuir, 2000, 16, 3569.
42. X. Nan, Z. Gu and Z. Liu, J. Colloid Interface Sci., 2002, 245, 311.
43. J. J. Gooding, R. Wibowo, J. Lie, W. Yang, D. Losic, S. Orbons, F. J. Mearns, J. G. Shapter and D. B. Hibbert, J. Am. Chem. Soc., 2003, 125, 9006.
44. D. K. Aswal, S. Lenfant, D. Guerin, J. V. Yakhmi and D. Vuillaume, Anal. Chim. Acta, 2006, 568, 84.

---