Influence of ozone on N-doped multi-walled carbon nanotubes

Abstract

Nitrogen-doped multi-walled carbon nanotubes (N-MWCNT) were selectively grown on oxidised silicon substrate by means of catalytic chemical vapour deposition with decomposition of acetonitrile in the presence of ferrocene (FeCp₂) at 900°C. The synthesised N-MWCNT film was initially treated with continuous ozone flow for various time periods and characterised by means of scanning electron microscopy and Raman spectroscopy. The electrochemical response of ozone-treated N-MWCNT film towards ferrocyanide/ferricyanide, [Fe(CN)₆]^3−/4− redox couple was investigated by means of cyclic voltammetry (CV) and electrochemical impedance spectroscopy. The results were compared with those obtained for untreated N-MWCNT film. The findings reveal the occurrence of structural defects on surface of ozone-treated N-MWCNT film due to oxidation of nanotubes. Furthermore, the handling of N-MWCNT film with ozone increases the barrier for interfacial electron transfer and slows down the kinetics of redox reaction occurring on this particular electrode.

Keywords:

• chemical vapour deposition
• electrochemical impedance spectroscopy
• multi-walled carbon nanotubes
• ozone
• Raman spectroscopy

1. Introduction

The carbon nanotubes (CNTs) are allotropes of carbon with cylindrical nanostructure that were recently discovered and attracted special attention due to their unique electronic and mechanical properties as well as to their resistance to acid/basic media 1–3. CNT can be synthesised by means of several techniques based on either chemical 4 or plasma 5 processes and applied as materials for the development of biological and chemical gas sensors 6,7, supercapacitors 8, hydrogen storages 9–11, solar cells 12, drug deliveries 13 and electron field emitters 14. Furthermore, from CNT various carbon-based nano-electronic and magneto-electronic devices (e.g. one-dimensional GaN nanorods) can be produced 15–18. The modification of CNT leads to essential changes in their electronic structure as well as their functional properties and leads to the creation of novel materials with incredible interesting properties 19,20. One of the simplest approaches in modification of CNT involves their oxidation with various oxidising agents and the functionalisation of the active oxygen-containing groups that are formed on their tube walls. The oxidation of CNT has gained a lot of attention for many scientists the last years since it enhances the chemical reactivity of the graphitic network. Specifically, it was found that through a treatment of pristine CNT with an oxidising agent, oxygen-containing groups are formed on graphitic surface, which facilitate the exfoliation of CNT bundles and increase their solubility in polar media 21,22. It is well known that diluted HCl solutions remove the residues of the metal catalyst from CNT, but they do not affect considerably their structure 23. However, strong oxidising agents, such as HNO₃, cause defects on the structure of CNT leading, consequently, to the formation of oxygen-containing groups, such as hydroxyl, carbonyl and carboxyl groups, on their tube walls via covalent functionalisation 24. The oxidised CNT can be later modified at their functional groups through various methods that allow organic molecules to be directly attached to the nanotube sidewalls. Depending on application for which the modified CNT are intended, methods such as cyclo-addition, electrophilic, nucleophilic or radical addition, as well as various polymerisation techniques, can be used for the functionalisation of CNT 25. The gas treatment for the oxidation of CNT is a simple and effective method that does not require the use of solvents and was already adopted in numerous works dealing with CNT chemistry 26. The handling with ozone was preferred from many scientists since it is one of the strongest known oxidants and although ozone itself is toxic, ozone treatment leaves no toxic residues that have to be treated or disposed 27. The reaction between ozone and CNT was already investigated theoretically 28 and experimentally 29. A main feature of ozonolysis is the preferential oxidation of CNT by ozone attack and the release of CO₂. For multi-walled carbon nanotubes (MWCNT), the oxidation tends to start near the tips permitting the cutting of the tips and the opening of the tubes. The oxidation proceeds layer by layer, resulting in thinner tubes 30.

In this work, nitrogen-doped MWCNT (N-MWCNT) were grown on oxidised silicon substrate by using the chemical vapour deposition (CVD) technique, treated with continuous ozone flow for different time periods (60, 120 and 180 s), and characterised by means of scanning electron microscopy (SEM) and Raman spectroscopy. Furthermore, the electrochemical response of untreated and ozone-treated N-MWCNT film towards [Fe(CN)₆]^3−/4− was probed by means of CV and EIS techniques. The results are discussed in terms of structural changes of N-MWCNT occurring upon oxidation.

2. Experimental

2.1. Reagents and solutions

Potassium hexacyanoferrate(III), K₃Fe(CN)₆ (∼99.0%), potassium hexacyanoferrate(II) trihydrate, K₄Fe(CN)₆ · 3H₂O (>98.5%) and potassium chloride, KCl (>99.0%) were purchased from Sigma-Aldrich and used as received without any further purification. The measured solutions consisted of K₃Fe(CN)₆/K₄Fe(CN)₆ mixture (1.0 × 10⁻³mol L⁻¹) were prepared by dissolving the appropriate amounts of the electro-active compounds in aqueous solution of KCl (0.1 mol L⁻¹). The solutions were prepared immediately prior the electrochemistry experiments by using double-distilled water having specific conductivity 0.50 µS cm⁻¹.

2.2. Apparatus and procedures

The scanning electron micrographs of N-MWCNT were obtained on FEI/Philips (model XL30 ESEM) computer-controlled SEM with an accelerating voltage of 20 kV and magnification factors in the range of ×500–×6000. The cyclic voltammograms (CVs) were recorded using a computer-controlled system Zahner (model IM6/6EX). The effect of the uncompensated resistance was reduced by applying the positive feedback technique. The electrochemistry measurements were carried out using a three-electrode cell configuration. The working electrode was either treated or ozone-untreated N-MWCNT (geometric area 1.0 cm²) film, the counter electrode was Pt plate (geometric area 2.0 cm²) and the reference electrode was Ag/AgCl (sat. KCl). For all the electrochemistry experiments, a three-compartment electrochemical cell specially designed to minimise the distances between the electrodes with a total solution volume of 20 mL, was used. Before each measurement process, the solution was purged with high purity argon in order to eliminate interference from oxygen. The CVs were recorded in the potential range from −0.2 to +0.8 V (versus Ag/AgCl) with scan rates ranging from 0.05 to 0.10 V s⁻¹. All measurements were carried out at room temperature (20°C). The electrochemical impedance spectra (EIS) were recorded with a computer-controlled system Zahner (model IM6/6EX) by applying very small amplitude of the alternate voltage (10 mV) in the frequency range from 0.1 Hz to 100 kHz at the room temperature (20°C). All electrochemistry experiments were performed on either untreated or ozone-treated N-MWCNT (geometric area 1.0 cm²) film against the reference electrode Ag/AgCl (sat. KCl), while Pt plate (geometric area 2.0 cm²) was served as counter electrode. The EIS spectra were recorded at the half-wave potential of the investigated redox couple [Fe(CN)₆]^3−/4− (E _1/2 ≈ +0.21 versus Ag/AgCl). The recorded EIS spectra were analysed by means of the equivalent electrical circuit (R _s + (C _dl/(R _ct + Z _w))) 31,32 (software Thales, version 4.15), which includes a resistance in series with a parallel combination of a capacitive path and charge- and mass-transfer paths. The elements of the electrical circuit are clarified as follows: the resistors R _s and R _ct represent the electrolyte and the charge transfer resistances, respectively, the capacitor C _dl represents the double-layer capacitance and Z _W is the Warburg impedance element 33. The mean and maximum modified impedance simulation errors were estimated in all cases to be less than 0.4% and 7%, respectively. The Raman scattering measurements were carried out using a Jasco RFT-6000 FT-Raman attachment coupled with an FT/IR-6300 spectrometer at an excitation wavelength of 1064 nm (excitation energy of 1.17 eV). The Raman spectra were recorded directly from either untreated or ozone-treated N-MWCNT film with a laser power of 50–200 mW and spectral resolution of 4 cm⁻¹ in the backscattering geometry.

2.3. Fabrication and treatment of N-MWCNT

Vertically aligned MWCNT having diameter in the range of 30–150 nm were selectively synthesised on oxidised porous silicon wafer with geometrical area of 1.0 cm² (Siegert Consulting e.K.) by means of CVD technique using acetonitrile (ACN) as carbon source and FeCp₂ as catalyst. The used silicon wafer having a thickness of about 525 µm was fully covered with a SiO₂-coating of about 1.7 µm thickness. The ACN solution including FeCp₂ (1% w/w) was introduced to the furnace at 900°C through a syringe with a flow rate of 0.2 mL min⁻¹. The synthesis process was performed using argon as a carrier gas. The average thickness of the synthesised N-MWCNT film was estimated to be about ∼27 µm. Elemental composition analysis of synthesised MWCNT performed by means of X-ray photoelectron spectroscopy reveal the existence of significant amount of incorporated nitrogen in nanotubes (7.0%). The MWCNT-based working electrode used for the electrochemistry experiments was prepared as follows: the produced N-MWCNT film was initially connected to platinum wire by using silver conducting coating and afterwards, the silver layer was fully covered with varnish protective coating. The scheme of CVD apparatus and additional details concerning the synthesis and characterisation of N-MWCNT were already reported in previous articles 34–36. Ozone was generated by an electric discharge in flow of pure oxygen (>99.99%) using an ozone-generator (Model 502, Fischer technology, Bonn, Germany) set at a flow rate of 50 L h⁻¹. The ozone treatment time lasted from 60 to 180 s.

3. Results and discussion

Representative SEM micrographs taken for untreated and ozone-treated N-MWCNT film are shown in Figure 1. In order to identify better the morphology of nanotubes, different level of magnifications was used form taking the SEM images. As was previously reported in literature, the nanotubes doped with nitrogen are bamboo-like shaped and perfectly ordered 37. The SEM images confirm that the nanotubes of untreated N-MWCNT are well ordered and separated with each other leading to well-organised film surface with insignificant faults (Figure 1a). The ozone-treated N-MWCNT film consists of closely packed and highly entangled CNT bundles. The surface of this particular film contains ‘holes’ having diameters in the range 5–40 µm (Figures 1b, c). In order to verify the micro-physical analysis from the SEM micrographs, a quantification of the defects was made by using Raman spectroscopy. Representative Raman spectra of untreated and ozone-treated N-MWCNT are presented in Figure 2. The Raman shifts along with the Raman bands intensity ratios are listed in Table 1. As it can be seen in Figure 2, the Raman spectra of N-MWCNT film consist of two main bands, namely the so-called D-band (∼1290 cm⁻¹) and the G-band (∼1597 cm⁻¹) 38, and their frequency is unaffected by ozone treatment. The G-band corresponds to the high frequency E_2g first-order mode and can be attributed to the opposing movement of two-neighbouring carbon atoms in graphene sheet 39, while the D-band results from defects in curved graphene sheets, as well as from ends and finite size crystalline domains of CNTs 40. It is well known that the ratio of relative intensities of G- and D-bands (I _G/I _D) is a measure of the degree of structural order. For MWCNT with different surface modifications, the change in (I _G/I _D) intensity ratio can be attributed to the change in surface functionalities or local defects resulting from bond breakage between C–C bonds in nanotubes 41. In this work, the ozone treated N-MWCNT films were found to exhibit smaller value of intensity ratio (I _G/I _D) compared to the untreated film, indicating that the structural defects prevail in ozone handled films 42. The findings suggest that a certain amount of structural disorder in ozone-treated N-MWCNT was generated by the continuous decoration of functional groups on the sidewalls of the nanotubes. The covalent functional groups that were introduced to the sidewalls of the nanotubes are assumed to cause difficulties to the tubes to retain their perfect cylindrical shape 43. The destruction of the tubular morphology of nanotubes with the ozone treatment can probably explain the ‘holes’ with various geometric shapes and sizes that were observed on surfaces of the ozone-treated N-MWCNT films (Figure 1). The (I _G/I _D) ratio progressively decreases with the ozonolysis time indicating an enhancement of the degree of defects with ozonolysis (Figure 3). It is interesting that after a treatment for about 180 s insignificant variation of the (I _G/I _D) ratio occurs. Typical CVs obtained with the scan rate of 0.05 V s⁻¹ showing the effect of ozonolysis time on electrochemical performance of N-MWCNT film are displayed in Figure 4. The CVs were recorded in the presence of [Fe(CN)₆]^3−/4− (1.0 × 10⁻³mol L⁻¹) redox couple in KCl solution (0.1 mol L⁻¹). The extracted CV parameters are presented in Table 2. A couple of well-defined symmetric redox peaks with equal cathodic () and anodic () peak currents (/ ≈ 1) were observed in CVs recorded on either untreated or ozone-treated N-MWCNT films. The findings exhibit that the amounts of oxidised and reduced forms of electro-active compound are equal and not consumed in coupled chemical reaction, confirming, that the charge-transfer process occurring on studied N-MWCNT films can be considered as Nernstian. Besides, the determined value of charge transfer coefficient of a = 0.5, which is a typical value for reversible reactions, supports further the last conclusion 44,45. The half-wave potential (E _1/2) of [Fe(CN)₆]^3−/4− redox system was found to be independent of ozonolysis time. Nevertheless, an influence of ozone handling time of N-MWCNT film on anodic and cathodic peak potential separation ΔE _p (ΔE _p = E _p ^ox− E _p ^red) was observed, namely, an increase of ΔE _p with the ozone-handling time of N-MWCNT film occurs. Considering that ΔE _p is related to heterogeneous charge transfer rate constant (k _s) of redox reaction, the findings indicate an important falling of kinetics of [Fe(CN)₆]^3−/4− redox couple with ozonolysis. Specifically, the k _s values of [Fe(CN)₆]^3−/4−, determined by means of electrochemical absolute rate relation 46, reveal that an ozone treatment of N-MWCNT film for only 60 s leads to a decrease of more than 80% in kinetics of the redox reaction occurring on this particular electrode (Table 2). It is also remarkable that further ozonolysis slightly affects the kinetics of redox reaction. The specific surface area of the untreated N-MWCNT film determined as 1.71 cm² increases sharply to 4.18 cm² after an ozonolysis treatment for about 60 s, while further ozone handling causes a trivial improvement of the surface area, namely an increase of about 4% was observed for an additional ozone treatment for about 60 s (Table 2). In general, it was observed that the current density on ozone-treated N-MWCNT film is about 7% greater of that of the untreated film. The results suggest that the treatment of N-MWCNT film with ozone increases its specific area upon cutting off (opening) the nanotube tips by the oxidation process 47. Complementary information concerning the electrochemical response of either the untreated or ozone-treated N-MWCNT films towards [Fe(CN)₆]^3−/4− was obtained from EIS measurements. Representative EIS spectra illustrating the effect of ozonolysis on impedance performance of N-MWCNT film are shown in Figure 5. In the EIS spectra the complex impedance is presented as a sum of the real (Z _re) and imaginary (Z _imag) components that originate mainly from the resistance and capacitance of the cell, respectively. The EIS spectra were analysed by means of the equivalent electrical circuit (R _s + (C _dl/(R _ct + Z _w))). The determined EIS parameters are included in Table 2. As it can be seen in Figure 5, the EIS spectra include a semicircle portion that corresponds to the electron-transfer-limited process, and a linear part that is attributed to diffusion-limited electron transfer. The semicircle diameter, obtained by extrapolation of the semicircle on Z _re-axis, corresponds to the electron transfer resistance (R _ct), a parameter that is attributed to electrochemical reaction occurring on electrode's surface. As it can be observed in EIS spectra displayed in Figure 5, the handling with ozone leads to an increase of the semicircle diameter, indicating that the electro transfer process occurring on ozone-treated N-MWCNT, is quite limited. It was observed that the R _ct values of the [Fe(CN)₆]^3−/4− redox system on N-MWCNT film tend to increase with the ozonolysis time. Specifically, the findings are in absolute accordance with the CV results and demonstrate that the barrier for interfacial electron transfer increases with the ozone handling time.

Figure 1. SEM micrographs taken for untreated (a) and ozone-treated N-MWCNT films (b, c, d). The SEM images were obtained with an accelerating voltage of 20 kV and magnification factors of 1600× (a), 500 × (b), 2000× (c) and 6000× (d).

Figure 2. Raman spectra recorded for either untreated (a) or ozone-treated N-MWCNT film for 60 s (b) and 180 s (c). The symbols D and G represent the Raman D- and G-bands, respectively.

Figure 3. Variation of the intensity ratio of G- and D-bands (I _G/I _D) with the ozonolysis time of N-MWCNT film.

Figure 4. Representative CVs displaying the effect of ozonolysis time on electrochemical performance of N-MWCNT film. The CVs were recorded at v = 0.05 V s⁻¹ in the presence of [Fe(CN)₆]^3−/4− (1.0 × 10⁻³mol L⁻¹) in KCl (0.1 mol L⁻¹) on either untreated (dotted line) or ozone-treated N-MWCNT film for 60 s (dashed line), 120 s (dotted-dashed line) and 180 s (solid line).

Figure 5. Representative EIS spectra illustrating the effect of ozonolysis time on impedance behaviour of N-MWCNT film. The EIS spectra were recorded in the presence of [Fe(CN)₆]^3−/4− (1.0 × 10⁻³mol L⁻¹) in KCl (0.1 mol L⁻¹) either on untreated (filled square) or ozone-treated N-MWCNT film for 60 s (filled circle), 120 s (filled upward triangle) and 180 s (filled downward triangle). For better readability, the solution ohmic resistance (∼10 Ω) was subtracted from the real part of the impedance data. The EIS spectra were recorded at the half-wave potential of [Fe(CN)₆]^3−/4− (E _1/2 ≈ +0.21 V versus Ag/AgCl).

Table 1. Raman D-bands, G-bands and intensity ratios of G- and D-bands (I _G/I _D) for the untreated and ozone-treated N-MWCNT for different time periods.

	Ozone treatment time (s)
	0	60	120	180
v (D) (cm^−1)	1289	1289	1288	1288
v (G) (cm^−1)	1597	1597	1598	1598
I G/I D	1.25	0.73	0.62	0.56

Table 2. Anodic (E _p ^ox) and cathodic (E _p ^red) peak potentials,^a peak potential separations (ΔE _p), anodic peak current densities (), cathodic and anodic peak current ratios (/), half-wave potentials (E _1/2),^a active surface areas (A), heterogeneous electron transfer rate constants (k _s),^b charge transfer coefficients (a),^c double layer capacitances (C _dl),^d charge transfer resistances (R _ct)^d and Warburg parameters (σ)^de for [Fe(CN)₆]^3−/4− (1.0 × 10⁻³mol L⁻¹) on untreated and ozone-treated N-MWCNT film for various time periods.

	Ozone treatment time (s)
	0	60	120	180
E p ^ox (V)	0.252	0.307	0.313	0.320
E p ^red (V)	0.164	0.114	0.108	0.094
ΔE p (V)	0.088	0.193	0.205	0.226
(10^−3 A cm^2)	0.150	0.160	0.160	0.160
/	1.04	1.01	1.05	1.07
E 1/2 (V)	0.208	0.210	0.210	0.207
A (cm^2)	1.71	4.18	4.36	4.44
k s (10^−3 cm s^−1)	5.56	0.87	0.75	0.60
A	0.50	0.50	0.50	0.50
C dl (10^−3 F)	0.09	0.09	0.03	0.03
R ct (Ω)	42.7	44.6	69.8	84.4
σ (Ω s^−1/2)	267	250	215	204
Notes: ^aAll potentials are reported versus Ag/AgCl (KCl sat.) reference electrode.
^bThe ks values were determined according to the relation: ψ = (Do /DR ) ^a ^/2 ks (nπFvDo/RT)^−1/2, where ψ is a kinetic parameter, a the charge transfer coefficient, D o, D R the diffusion coefficients of the oxidized and reduced species, respectively (D o ≈ D R) and n the number of the electrons involved in the reaction (n = 1) 46.
^cThe a values were determined graphically according to the following relation: Ep = E^o -(RT/anaF) ln(anaF/RTk s)-(RT/anaF)lnv 44,45.
^dThe EIS parameters were determined by using the electrical circuit model (R s + (C dl/(R ct + Z w))) (software Thales, version 4.15).
^eThe Warburg impedance ZW is connected with the Warburg parameter σ according to the relation: Z w = σ/(iω)^1/2, where ω is the angular frequency 33.

It is very interesting that analogous findings were obtained on ozone treated pristine MWCNT. Namely, a film consists of pristine MWCNT was produced with decomposition of benzene by means of CVD 34–36 and the effect of the ozone treatment on its electrochemical response towards [Fe(CN)₆]^3−/4− was investigated. The electrochemical behaviour of [Fe(CN)₆]^3−/4− on ozone treated pristine MWCNTs was comparable to that observed on N-MWCNTs. Specifically, an increase of peak potential separation of about 150% and decrease of heterogeneous rate constant of about 85% was observed on pristine MWCNTs after an ozone treatment for about 180 s. The findings reveal that the handling with a very powerful oxidising agent such as ozone causes nearly the same structure effects on both nanotube materials (pristine and N-MWCNT).

4. Conclusions

A novel N-MWCNT/oxidised silicon film was synthesised by means of CVD with decay of ACN in the presence of FeCp₂ at 900°C. The N-MWCNT film was treated with ozone and applied as working electrode for the electrochemical analysis of [Fe(CN)₆]^3−/4− redox couple. The findings reveal the occurrence of structural defects on surface of ozone-treated N-MWCNT film. The defective surface is probably resulted from the C–C breakage and the formation of carboxyl and hydroxide groups on sidewalls of the nanotubes. The findings demonstrate that the ozone treatment leads to significant increase of the film's active surface and enhances the current density. Moreover, the treatment of N-MWCNT film with ozone increases the barrier for interfacial electron transfer and slows down the kinetics of redox reaction occurring on this particular electrode. It is remarkable that an ozonolysis for about 60 s leads to decrease for about 80% of the kinetics of investigated redox system.

Acknowledgments

The authors would like to thank Mrs. D. Schneider and Mrs. Risch (TU Ilmenau).