Fully organic ITO replacement through acid doping of double-walled carbon nanotube thin film assemblies

Abstract

Transparent electrodes made from metal oxides, of which indium tin oxide (ITO) is most common, suffer from poor mechanical flexibility and electrochemical stability. Highly transparent and electrically conductive thin films based on double-walled carbon nanotubes (DWNTs) were assembled layer-by-layer (LbL) as a potential replacement for ITO. The alternate deposition of positively charged poly(diallyldimethyl ammonium chloride) [PDDA] and DWNTs, stabilized with negatively charged deoxycholate (DOC), exhibits linear film growth. A five bilayer (BL) assembly exhibits a sheet resistance of 309 Ω sq⁻¹, and visible light transmittance of 84%. In an effort to further reduce sheet resistance, these thin films were exposed to HNO₃ vapor. Sheet resistance of the 5 BL film was reduced to 104 Ω sq⁻¹ (4200 S cm⁻¹ conductivity, 22.9 nm thickness), with no change in transparency. Single-walled carbon nanotube assemblies show an even greater change when exposed to HNO₃ vapor, with a 5 BL assembly (%T > 85) decreasing from 1236 to 237 Ω sq⁻¹. The DWNT-based assemblies maintained their low sheet resistance after repeated bending and also showed excellent electrochemical stability relative to ITO. This work demonstrates the excellent optoelectronic performance, mechanical flexibility, and electrochemical stability of nanotube-based assemblies, which are potentially useful as transparent electrodes for a variety of flexible electronics.

---

Introduction

Transparent electrodes continue to be heavily studied because of increasing demand for optoelectronics, such as photodiodes for solar cells and electronic displays (e.g., touch screens and light-emitting diodes). These electrodes not only transport electrons, but in many cases must transmit visible light because they are applied directly onto the surface of a transparent substrate. Especially for display applications, it is important for these electrodes to exhibit high transparency (>85%) and low sheet resistance (<1000 Ω sq⁻¹). Several materials have achieved the combination of resistance and transparency necessary to meet the needs of most optoelectronic applications. Numerous reviews describe the advantages and disadvantages of these transparent electrodes.^1–3 Metal-based thin films (e.g., silver, gold, or copper) have largely been too opaque,^4,5 while intrinsically conductive polymers (e.g., PEDOT) suffer from photo-oxidative degradation.^6,7 The best combination of high transparency and low sheet resistance is currently found in metal oxides (e.g., ATO, ITO, ZnO:Al, etc.).^8–13

Beginning with cadmium oxide in 1951, doped, wide-bandgap transition metal oxides (primarily In₂O₃, SnO₂ or ZnO) have been widely investigated due to their high electrical conductivity and optical transmittance.¹ These studies ultimately led to the development of doped In₂O₃:Sn (also known as indium tin oxide or ITO), which has excellent electrical and optical properties.¹⁴ These films are hard, dense, strongly adherent to glass, and chemically inert, relative to metal thin films. Most of these oxides, such as ZnO:Al,⁸SnO₂:Sb,⁹ZnO_x,¹⁰ and ITO,^11–13 also exhibit very low resistivity (<5 × 10⁻⁴ Ω cm). ITO remains the most used and widely studied oxide, along with SnO₂:F and ZnO:Al, for transparent electrodes, including flat panel displays.^15,16 Additionally, ITO layers can be directly coated on flexible, transparent plastic substrates.¹⁷

Despite having low sheet resistance and high transparency, ITO thin films are hindered by their brittleness.^18–20 The difference in thermal expansion coefficient and elastic properties of ITO films and polymer substrates results in large mechanical stresses.¹⁹ Flexing an ITO coated polyethylene terephthalate (PET) film causes stress cracks that reduce its electrical conductivity. Several studies have examined the effect of deflection on the resistivity of ITO on a PET or polycarbonate (PC) substrate, focusing on the relationship between cracking and electrical properties.^21–23 These studies show that cracks are initiated at a strain of 1.28%, when films are stretched, and at a curvature of 10 cm, when films are bent. These limitations have led to extensive exploration of alternative transparent electrodes (i.e., ITO replacement).

Carbon nanotube (CNT)-based thin films are a more recently studied alternative to existing transparent conductive layers. Single-walled carbon nanotubes (SWNTs), consisting of one layer of hexagonal graphite lattice rolled to form a seamless cylinder with a radius up to a few nanometres,²⁴ are especially promising. SWNT thin films have been generated by several methods in recent years, such as vacuum filtration,²⁵ transfer printing,²⁶spin coating,²⁷ direct CVD growth,²⁸ air-spraying,²⁹ and rod coating.³⁰ These films show good optoelectronic performance, with values in the middle of the range of commercial ITO-coated PET, which typically has 50–200 Ω sq⁻¹ sheet resistance and ∼83% transmission at 550 nm. Although transfer printing is one of the best methods for producing CNT thin films, there has been difficulty with scale-up, breakage of the film during transfer, and comparatively brittle final films. Other deposition methods also tend to exhibit problems with regard to film quality, electrical performance, and processing complexity.³¹ These challenges have created an opportunity for the use of layer-by-layer (LbL) assembly to produce these thin films.

Uncertainty remains about the driving force for LbL assembly, although it is widely accepted that the multilayer buildup depends on the electrostatic attraction between oppositely charged molecules and the entropy gain from small counterions entering the water.^32–35Scheme 1 shows the buildup of CNT-based thin films via the LbL process, by alternately dipping a substrate in positively and negatively charged solutions. The strong electrostatic attraction, between deposited material on the substrate and oppositely charged molecules in solution, results in charge reversal of the original surface due to the adsorption of new molecules (or particles). Each positive and negative pair deposited is referred to as a bilayer (BL). Repetition of adsorption cycles with polyanions and polycations (or other charged ingredients) leads to LbL growth of multifunctional films. A key advantage of the LbL method is that there are no restrictions with respect to size, topology and chemistry of the substrate. Assemblies have been grown on glass, semiconductors, and numerous polymer films (PET, polystyrene (PS), polydimethylsiloxane, etc.).^36–46 There are several reports of nanotube-based LbL assemblies,^45–50 but none have simultaneously achieved visible light transmission >85% (at 550 nm) and sheet resistance below 100 Ω sq⁻¹, which is necessary to compete with ITO.

Scheme 1 Schematic of the layer-by-layer process that involves alternately substrate to cationic PDDA and DOC-stabilized anionic CNT mixtures, along with rinsing with deionized water and drying with filtered air. These steps generate one bilayer (BL) and are repeated to deposit thicker films.

In an effort to achieve true ITO transparency and sheet resistance, double-walled carbon nanotubes (DWNTs) stabilized with deoxycholate (DOC) were assembled with poly(diallyldimethyl ammonium chloride) [PDDA]. Replacing SWNTs from previous studies with DWNTs provides lower sheet resistance due to greater metallic nature.^51,52 The optoelectronic behavior of these nanotube-based thin films was characterized before and after nitric, sulfuric, and hydrochloric acid treatment. DOC is known to be a good surfactant for dispersion of individual nanotubes (see Fig. S1 in ESI†),^53,54 with its negatively charged tail causing nanotubes to behave as negatively-charged particles, which facilitates LbL assembly.^45,48 Exposing nanotube films to strong acid has been shown to enhance conductivity,^48,49 but its use with DWNT-based assemblies has never been explored.

Although the mechanism of acid treatment for nanotube-based assemblies is still uncertain, the increase in electrical conductivity is believed to be a combination of acid anion doping^55,56 and removal of insulating materials (i.e., stabilizer and polymer).^29,57 A 5 BL DWNT assembly exhibits a significant reduction in sheet resistance, from 309 to 112 Ω sq⁻¹, after two minutes of exposure to HNO₃ vapor. This film is highly flexible, transparent, and electrochemically stable, making it a potential alternative to ITO. After 100 bending cycles, these films exhibit the same sheet resistance, but ITO-coated PET increases two orders of magnitude. In addition to mechanical stability, cyclic voltammetry was used to demonstrate the electrochemical stability of these transparent electrodes. This combination of optoelectronic performance, mechanical flexibility, and electrochemical stability make this a suitable candidate for a variety of applications requiring a transparent electrode.

Experimental section

Materials

Purified electric arc (EA) SWNTs (P2-SWNT, individual tube: average 0.5–3 μm length and 1.4 nm diameter, C ≥ 90 wt%), synthesized using Ni/Y catalysts, from Carbon Solutions, Inc. (Riverside, CA) were used as a base of comparison for HiPCO DWNTs (XB type, +1 μm length and ∼3.0 nm diameter, C ≥ 90 wt%), purchased from Continental Carbon Nanotechnologies, Inc. (Houston, TX), used in this study. PDDA, with a molecular weight of 100 000–200 000 g mol⁻¹, and DOC (C₂₄H₃₉NaO₄, ≥98%) were purchased from Sigma-Aldrich (Milwaukee, WI). Fuming sulfuric acid (oleum, H₂SO₄·(SO₃)_x, 20% free SO₃ basis), fuming nitric acid (HNO₃, 99.5%), and fuming hydrochloric acid (HCl, 37%) were also purchased from Sigma-Aldrich and used for acid treatment. All chemicals were used as received. PET (0.175 mm Melinex ST505 purchased from Tekra Corp., New Berlin, WI) and PS (0.125 mm ST311125 from Goodfellow Cambridge Ltd., Cambridge, UK) films were used as flexible substrates for CNT assemblies. Single side polished (1 0 0) silicon wafers (University Wafer, South Boston, MA) were used as substrates for thickness measurement. A 100 Ω sq⁻¹ ITO-coated PET sheet was purchased from Sigma-Aldrich and cut to size to measure the optical transmittance and the change in sheet resistance during bending cycles.

Layer-by-layer assembly

A cationic 0.25 wt% PDDA solution was prepared with 18.2 MΩ deionized water. 0.05 wt% CNTs were dissolved in deionized water containing 1 wt% DOC, followed by 10 W tip sonication for 20 min. Each cycle of LbL assembly consists of substrate immersion into an aqueous mixture, with rinsing and drying after each deposition, beginning with the cationic solution. Cycles were repeated to deposit the desired number of BLs. All films were stored in a dry box for a minimum of 12 h prior to testing or acid treatment. A schematic of this LbL deposition process is shown in Scheme 1. For acid treatments, assembled CNT films were held in a saturated acid vapor environment. Three types of concentrated acids (fuming sulfuric, nitric, and hydrochloric) in petri dishes were maintained at 70 °C with a water bath (caution: dangerous oxidizing agents). After a 2 to 30 min exposure to acid vapor, the CNT films were rinsed with deionized water and dried with filtered air.

Thin film characterization

Thickness of assemblies was determined by averaging the values of at least five different positions using a PHE-101 Ellipsometer (Microphonics, Allentown, PA) equipped with a 632.8 nm laser. Absorbance and transmittance were measured between 300 and 850 nm with a USB2000 UV-vis spectrometer (Ocean Optics, Dunedin, FL). All absorbance values in the text are for one-sided coating (i.e., absorbance from slides coated on both sides was halved). Sheet resistance of nanotube assemblies on PET was measured with a Signatone Pro4 four-point probe system (Gilroy, CA) with 0.4 mm probe tip diameter and 1.0 mm tip spacing. An Agilent E3644A DC power supply (Santa Clara, CA), a Keithley digital multimeter (Cleveland, OH), and LabVIEW using a SCB-68 shield I/O connecter block (National Instruments Inc., Austin, TX) were used for voltage and current data collection. The Fourier-transform infrared (FT-IR) spectra were obtained with an ALPHA FT-IR spectrometer (Bruker Optics, Billerica, MA). Scanning electron microscopy was performed with a Quanta 600 FE-SEM (FEI Co., Hillsboro, OR) at an operating voltage of 10 kV. Energy dispersive X-ray (EDX) analysis was carried out using an Oxford EDX system attached to the FE-SEM. Thin film cross-sections were imaged with a JEOL 1200 EX TEM (JEOL USA Inc., Peabody, MA) with an operating voltage of 100 kV. For TEM specimens, the thin films on PS substrates were embedded in an epoxy resin comprised of Araldite 502 (modified bisphenol A) and Quetol 651 (ethylene glycol diglycidyl ether), along with dodecenyl succinic anhydride (DDSA) hardener and benzyldimethylamine (BDMA) accelerator (2 : 1 : 1 mole ratio of Araldite 502:Quetol 651:DDSA and 0.2 ml BDMA per 10 g of total weight), which were purchased from Electron Microscopy Sciences (Hatfield, PA). The specimens were sectioned down to ∼90 nm and placed on 300 mesh nickel grids to dry prior to imaging. An Epsilon 851 electrochemical workstation (BASi Instrumentation, West Lafayette, IN) was used for cyclic voltammetry. A CNT- or ITO-coated PET electrode, platinum wire, and an Ag/AgCl (3 M KCl) electrode were used as the working, counter, and reference electrodes, respectively. These measurements were performed at ambient temperature (22 ± 2 °C) in 0.1 M Na₂SO₄ solution that had been bubbled with N₂ gas for more than 20 min prior to measurement. Cyclic voltammetry was performed between two potential limits, −0.2 and 0.8 V, at a scan rate of 100 mV s⁻¹.

Results and discussion

Growth and microstructure of carbon nanotube assemblies

Fig. 1 shows the linear growth of cationic PDDA and anionic DOC-stabilized double-walled carbon nanotubes up to 10 BLs, denoted as [PDDA/(DWNT + DOC)]_n, where n is the number of bilayers deposited. Ellipsometric thickness was obtained by depositing these thin films on silicon wafers. The average thickness of one DWNT-containing bilayer is 4.4 nm, which is much thicker than the individual thickness of a double-walled carbon nanotube (1.5–3 nm), suggesting that the DOC stabilizer and PDDA fully enveloped the DWNT network deposited. In agreement with film thickness, optical absorbance confirms that these CNT-based LbL assemblies grow linearly up to 10 BL. A [PDDA/(DWNT + DOC)]₅ film on PET is highly transparent, with 84.2% transmittance at 550 nm, which is comparable to ITO-coated PET (84.6%T at 550 nm, as shown in Fig. S2 in ESI†). These complementary measurements (ellipsometry and UV-vis) further suggest a constant concentration of DWNTs and polymer in every bilayer, which is in agreement with the PDDA/(SWNT + DOC) system described in a previous study.⁴⁸

Fig. 1 Thickness of PDDA/(DWNT + DOC) thin films as a function of the number of bilayers deposited. The solid lines are linear curve fits. Transmittance was calculated from absorbance data. White points are values after 20 min of exposure to nitric acid vapor.

Fig. 1 shows there is very little change in thickness and absorbance following acid vapor exposure. Previous studies indicate that swelling of CNT networks (i.e., an increase in CNT spacing) occurred with acid doping due to denser layers of acidic anions around the individual nanotubes.^55,58 In spite of this swelling effect, some CNT thin films have shown a significant decrease in thickness following acid doping. For example, a sprayed SWNT film was reported to have undergone a 25% thickness reduction due to the removal of bulky surfactant.²⁹ The minimal reduction of thickness for the present LbL assemblies suggests a minimum amount of surfactant was associated with the DWNT during deposition. In some LbL systems, CNT assemblies with a polymeric stabilizer showed more than 30% thickness reduction after sulfuric acid treatment,⁴⁹ suggesting that these films contained excess polymer.

SEM surface images of [PDDA/(DWNT + DOC)]₆ thin films, before and after nitric acid treatment, are shown in Fig. 2. What appears to be a bundled DWNT network is more clearly seen following acid treatment (Fig. 2(B)). DOC exfoliation of individual DWNTs in deionized water ultimately results in a uniform distribution on the substrate.^53,54 During deposition, the nanotubes bundle somewhat due to their high concentration in the deposited film, which creates a strong network and high electrical conductivity. Before treatment (Fig. 2(A)), the surface of the film was covered by an insulating layer (polymer and/or surfactant) that made the DWNT network difficult to observe. After nitric acid treatment for 20 min, an extensive DWNT network was exposed due to removal of PDDA and DOC (Fig. 2(B)). This removal of insulating material enables more direct contact between these highly conductive nanotubes. The bright spots in these images are believed to be metal catalyst particles used to produce DWNT. Removal of insulating organic material during acid treatment more clearly reveals these nanoparticles (Fig. 2(B)). It is also believed that HNO₃ enhances conductivity through the formation of a charge-transfer complex between DWNTs and NO₃⁻ layers around the individual nanotubes, which promote bundling into thicker nanotube ropes and improved alignment of these ropes.^55,58,59TEM cross-sectional micrographs (Fig. 2(C) and 2(D)) illustrate a highly inter-diffused nanostructure, in which the dark and light grey areas are DWNT-rich and PDDA-rich regions, respectively, and the darkest black dots are catalyst impurities from the HIPCO process. In addition to highlighting thin film structure, the TEM images verify the inner diameter of an individual DWNT (∼3 nm) and the negligible change in the [PDDA/(DWNT + DOC)]₁₀ film thickness due to nitric acid treatment.

Fig. 2 SEM surface images of [PDDA/(DWNT + DOC)]₆ on PET (A) before and (B) after 20 min exposure to nitric acid vapor. TEM cross-sections of [PDDA/(DWNT + DOC)]₁₀ (C) before and (D) after 20 min treatment with nitric acid.

Optoelectronic performance of carbon nanotube assemblies

Sheet resistance of [PDDA/(DWNT + DOC)]_n and [PDDA/(SWNT + DOC)]_nassemblies on PET was measured as a function of visible light transmittance, which decreased with increasing the number of BLs deposited, as shown in Fig. 3(A). Sheet resistance significantly decreases with increasing the number of BLs, along with a gradual decrease in %T. [PDDA/(DWNT + DOC)]₁₀ thin films achieve sheet resistance as low as 127 Ω sq⁻¹ (with 67.2% transmittance at 550 nm) at 10 bilayers, but this system is 1180 Ω sq⁻¹ (with 93.5% transparency) at just two BLs. The PDDA/(SWNT + DOC) system exhibits the same optoelectronic trend, with higher sheet resistance than, and similar transmittance to, the DWNT system. This reduction in sheet resistance is due to increased thickness and connectivity of the nanotube network.^45,48 A 4 BL PDDA/(DWNT + DOC) film has the best optoelectronic performance of 413 Ω sq⁻¹ and 87.3% transmittance, which meets the key criteria for touch screens (500 Ω sq⁻¹, 85% transmittance).⁶⁰

Fig. 3 (A) Sheet resistance as a function of transparency, for assemblies of varying thickness, and (B) electrical conductivity as a function of thickness for the [PDDA/(SWNT + DOC)]_n and [PDDA/(DWNT + DOC)]_n systems.

Electrical conductivity of these films, shown in Fig. 3(B), was obtained by multiplying the inverse of sheet resistance by ellipsometric film thickness. The gradual increase from 909 to 1810 S cm⁻¹, for the PDDA/(DWNT + DOC) system, suggests an increase in the density of intersecting pathways for electron transport (due to interconnection between DWNTs in upper and lower layers) as the nanotube network transitions from two to three-dimensional with increasing layers. DWNT-based assembly conductivity is expected to converge on a constant value beyond 10 bilayers, much like the SWNT film. The PDDA/(SWNT + DOC) system also shows a gradual increase in conductivity up to 10 BLs and then levels off at ∼1050 S cm⁻¹. The conductivity of 1810 S cm⁻¹ is low relative to a vacuum filtered SWNT film (6700 S cm⁻¹)²⁵ and sprayed SWNT film (5500 S cm⁻¹),²⁹ but it is higher than any other CNT-based LbL assembly.^45–47,50 This relatively high conductivity is due to thin PDDA deposition and the relatively mild sonication used to disperse the nanotubes. Acid treatment further improves the conductivity of the present assemblies, making them comparable to the highest conductivity CNT thin films.

Influence of acid treatment on sheet resistance

Even when SWNT is replaced by DWNT, the sheet resistance of as-assembled nanotube films is higher than the requirement for replacing the best ITO. It is believed that excessive polymer and surfactant disrupt connections between CNTs. Several studies have reported improvements in electrical conductivity of SWNTs with acid treatment,^29,30,49,55 but the influence of acid type and exposure time on the transparency and electrical conductivity of CNT assemblies has never been examined, especially when using surfactant-stabilized nanotubes. As previously mentioned, strong acidic anions form a charge-transfer complex around individual CNTs. Additionally, similar to previous surfactant/CNT thin films,^29,57,61 strong acid removes insulating material from these PDDA/(CNT + DOC) films.

Most prior acid treatment studies have been performed with dip treatment (i.e., immersing the films into a strong acid bath).⁶¹ Dip treating makes PET substrates brittle or even causes shrinkage. Moreover, dip treated CNT thin films readily de-adhere from substrates during rinsing. This instability due to acid treatment has been reported in multiple studies.^61,62 In the present study, films were exposed to acid vapor by heating the acid solution in a 70 °C water bath. Fig. 4(A) shows sheet resistance of DWNT assemblies after exposure to nitric, hydrochloric, or sulfuric acid vapor for increasing time periods. Nitric acid (b–e in Fig. 4(A)) dramatically reduced the sheet resistance of as-assembled 5 BL PDDA/(DWNT + DOC) thin films, from 309 to 107 Ω sq⁻¹, after a 10 min exposure. Acid type plays a significant role in reducing sheet resistance. Nitric acid treatment provides the lowest resistance, while hydrochloric and sulfuric acids result in values that are nearly double that of HNO₃.

Fig. 4 (A) Sheet resistance of [PDDA/(DWNT + DOC)]₅ after 2 to 20 min exposure to nitric acid, hydrochloric acid, and sulfuric acid. (a) is the sheet resistance of an as-assembled 5 BL DWNT thin film. (B) EDX spectrum of [PDDA/(DWNT + DOC)]₅ before acid treatment. Platinum was from ~10 nm thin coating for SEM images. The inset contains spectra of the 5BL DWNT films after 20 min nitric acid, hydrochloric acid and sulfuric acid treatments.

Uncertainty remains regarding why nitric acid is the most effective dopant, although it can be reasonably assumed that it is related to the intensity of doping of individual nanotubes and degradation of the surrounding insulating material. Dipped films have shown that H₂SO₄ is a stronger dopant than HNO₃ (and much more than HCl).^56,62 The use of vapor in the present work is a significant difference. In dip treatment, the concentration of acidic anions is that of the solution, while the vapor pressure is a more important factor for vapor treatment. The vapor pressure is 2, 8, and 167 mmHg at room temperature for fuming H₂SO₄, HNO₃, and HCl, respectively.⁶³Thesulfate anion is bigger than the nitrate anion and much bigger than the chloride ion. Additionally, it has been suggested that removal of insulating surfactant is more important than classical doping for these acid treatments.²⁹Nitric acid more effectively removes residual surfactant than sulfuric acid.

To better understand the effect of each acid treatment, doped assemblies were characterized by EDX, FT-IR, and SEM. Fig. 4(B) shows the EDX analysis, which reveals the composition of each thin film. Before acid treatment, 5 BL DWNT films were composed of 93.4 wt% carbon and 2.76 wt% oxygen. Acid exposure caused the content of carbon to increase, and that of oxygen to decrease. Table 1 summarizes these changes. Nitric acid treatment for 20 min resulted in a 0.89% increase in carbon content (by weight) and a 29.7% decrease in oxygen content. There were 0.51% and 0.12% increases and 17.1% and only 4.0% decreases after hydrochloric and sulfuric acid treatments, respectively. Of the ingredients in these assemblies, only DOC contains oxygen. This result suggests that PDDA and DOC molecules are more effectively removed by nitric acid vapor than the other two acids. The DWNT networks after each acid treatment were also characterized by SEM (see Fig. S3 in ESI†). 20 min nitric acid treatment exposes an extensive DWNT network, while more hazy DWNT networks are observed after 20 min hydrochloric and sulfuric acid treatments due to less removal of insulating molecules. EDX and SEM together suggest that the removal of insulating material is the main factor in resistance reduction of DWNT-based assemblies, rather than a true acid doping effect. Nitric acid seems to be the most effective at degrading PDDA and DOC, despite having a lower vapor pressure than HCl. The low solution concentration of hydrochloric acid (37%) may account for this discrepancy.

Table 1 Carbon and oxygen contents in DWNT and SWNT films before and after nitric acid treatment

System	Acid Treatment	C Content (wt%)	C Increase (%)	O Content (wt%)	O Decrease (%)
[PDDA/(DWNT + DOC)]5	No	93.4	—	2.76	—
	2 min HNO3	93.6	0.16	2.53	5.6
	5 min HNO3	94.0	0.63	2.09	22.0
	20 min HNO3	94.2	0.89	1.96	29.7
	20 min HCl	93.8	0.51	2.29	17.1
	20 min H2SO4	93.5	0.12	2.65	4.0
[PDDA/(SWNT + DOC)]10	No	97.0	—	2.97	—
	20 min HNO3	98.4	1.41	1.60	46.1

---

Fig. 5(A) shows the effect of nitric acid treatment on [PDDA/(SWNT + DOC)]₅, [PDDA/(DWNT + DOC)]₅, and [PDDA/(DWNT + DOC)]₁₀ as a function of exposure time (from 2 to 20 min). In just two minutes, all three films exhibit a remarkable decrease in sheet resistance. Further acid treatment, however, results in only a modest decrease in sheet resistance, relative to the first 2 min, especially beyond five minutes of exposure. 20 min nitric acid treatment did not achieve significantly lower sheet resistance than a 10 min treatment. It is likely that any insulating material present is degraded very quickly due to the very thin nature of these films (<50 nm).

Fig. 5 (A) Sheet resistance of [PDDA/(SWNT + DOC)]₅, [PDDA/(DWNT + DOC)]₅, and [PDDA/(DWNT + DOC)]₁₀ as a function of nitric acid treatment time. (B) FT-IR spectra of [PDDA/(DWNT + DOC)]₅ before and after nitric acid treatment, from 2 to 20 min. These spectra are magnified and overlaid with arbitrary offset for clarity.

To better understand the compositional changes due to nitric acid treatment, FT-IR analysis of [PDDA/(DWNT + DOC)]₂₅ films was performed, as shown in Fig. 5(B). The FT-IR spectrum of the film before nitric acid treatment is similar to that of a PDDA + MWNT film, whose study provided the peak assignments used here.^64,65 Absorbance peaks at 3386, 2928, and 2864 cm⁻¹ in the [PDDA/(DWNT + DOC)]₂₅ film were also observed in the (PDDA + MWNT) film. The 3386 cm⁻¹ peak is attributed to a strong O–H stretching vibration in DOC. As expected, this peak disappeared from the spectrum after nitric acid treatment, along with a decrease in the intensity of the peaks at 2928 and 2864 cm⁻¹, both attributed to CH stretching from PDDA and DOC. Additionally, the C=O in DOC exhibits a sharp band at 1715 cm⁻¹ before acid treatment, but this became a broad band centered at 1695 cm⁻¹ afterward, which indicates that this vibration band changed due to PDDA and DOC removal. FT-IR spectra after nitric acid treatment, from 2 to 20 min, did not show any significant differences, indicating removal of insulating materials due to nitric acid vapor was largely completed in the first two minutes. As a complementary result, EDX (Fig. S4(A) in ESI†) shows no significant decrease in oxygen concentration from 5 to 20 min exposure. As shown in Table 1, nitric acid treatment for 5 min caused a 22.0% decrease in oxygen content, which is almost the same after 20 min. FT-IR and EDX analyses suggest that the penetration of acid vapors through the film is nearly instantaneous. Furthermore, the SEM surface images in Fig. 6 support this assertion by showing that most insulating materials were removed after 2 min exposure to nitric acid vapor (and a 5 min treatment exhibited the same DWNT network as a 20 min treatment). Additionally, acid treating beyond 10 min produced no additional improvement for DWNT films and an increased sheet resistance for SWNT films. It is believed that some modest damage of CNTs occurs with excessive acid treatment (e.g., sidewall oxidation),^58,66 causing the observed increase in sheet resistance.

Fig. 6 SEM surface images of [PDDA/(DWNT + DOC)]₅ on PET (A) before and (B) after 2 min, (C) 5 min, and (D) 20 min HNO₃ exposure.

In addition to exposure time, CNT type alters the impact of nitric acid exposure on sheet resistance. In the first five minutes, the 5 BL SWNT film exhibits an order of magnitude decrease in sheet resistance, while the 5 BL DWNT film decreases by only a factor of three. A 10 BL DWNT film exhibits the same trend in sheet resistance as the 5 BL DWNT film. EDX analysis, in Fig. S4(B) and Table 1, helps to explain this trend by revealing a 46.1% decrease in oxygen weight of the 10 BL SWNT film and a 29.7% decrease in the 5 BL DWNT film, suggesting more significant removal of insulating materials from the SWNT films. SWNT films likely contain more insulating material than DWNT, in as-deposited assemblies, due to enhanced surface area (smaller diameter tubes and greater exfoliation). The SEM surface images (Fig. S5 in ESI†) support this assertion by showing blurrier SWNT film surface than DWNT (Fig. 6(A)) before acid treatment.

[PDDA/(SWNT + DOC)]₅, [PDDA/(DWNT + DOC)]₅, and [PDDA/(DWNT + DOC)]₁₀ achieved 227, 107, and 43 Ω sq⁻¹ sheet resistance, respectively, after 10 min exposure to HNO₃ vapor. These values are low enough to use for many optoelectronic applications. Visible light transmittance of these films is 86.8%, 84.0%, and 67.1%, respectively, as shown in Fig. 7. The influence of acid treatment on the optoelectronic performance of several other CNT-based LbL films is also included in Fig. 7, which emphasizes the relationship between sheet resistance and transmittance. Sheet resistance of SWNT films, due to nitric acid treatment, is consistently reduced by a factor of five in this study and others,^48,49 regardless of number of bilayers, while that of DWNT films was reduced by a factor of three. DWNT films with more than five bilayers exhibit sheet resistance below 100 Ω sq⁻¹, which is the lowest currently reported in the literature. This value is comparable with the best SWNT-based films produced using any other technique and is competitive with high transparency (∼85%T) ITO coatings.²⁹ The transmittance of the 5 BL DWNT film (>84%) is unaltered by the acid treatment (see Fig. S2 in ESI†). This optoelectronic performance demonstrates the ability of LbL assemblies to simultaneously achieve high transparency and low sheet resistance. Taking the inverse of the product of sheet resistance and thickness shows that the electrical conductivity of 5 BL and 10 BL DWNT films reach 4100 and 5300 S cm⁻¹, respectively, after nitric acid treatment for 10 min. This conductivity is higher than most CNT films made by any method, except for vacuum filtration whose films have relative low transparency (∼70%) that eliminates their ability to be used as ITO replacement.²⁵

Fig. 7 Optoelectronic performance of several CNT LbL systems before and after acid treatment. Points with the black outline are values of as-assembled LbL films and those with no outline are acid-treated.

Mechanical and electrochemical stability of carbon nanotube assemblies

The bending stability of transparent electrodes is becoming more and more important as flexible electronic applications continue to grow. Nitric acid-treated 5 BL DWNT LbL films were compared to a commercial 100 Ω sq⁻¹ ITO-coated PET film as a function of the number of bending cycles, as shown in Fig. 8. These films were repeatedly bent to a 1.0 cm radius of curvature (inset of Fig. 8), with sheet resistance measured at the center of each specimen after each bend. DWNT films maintained constant resistance up to 100 bending cycles. In contrast, the ITO-coated PET exhibited a two order of magnitude increase in sheet resistance after 100 bending cycles. ITO and other metal oxide semiconductors readily crack due to their ceramic brittleness, while DWNT LbL films have polymeric behavior, which is ductile enough to withstand the strain associated with severe bending. This exceptional mechanical stability in LbL films is potentially useful for flexible electronics.

Fig. 8 Sheet resistance as a function of bending cycles (to 1 cm radius of curvature) for 5 BL DWNT and 100 Ω sq⁻¹ ITO-coated PET films. R_So indicates the resistance (∼100 Ω sq⁻¹) prior to any bending.

Cyclic voltammograms (CV) of an HNO₃-treated 5 BL DWNT film and an ITO film (both on PET) were used to compare electrochemical stability, as shown in Fig. 9(A). The potential was cycled between −0.2 and 0.8 V vs. an Ag/AgCl reference electrode, with a scan rate of 100 mV s⁻¹, in 0.1 M Na₂SO₄ aqueous solution. ITO shows oxidation–reduction (redox) peaks at 0.1 and 0.3 V, respectively, indicating that ITO undergoes chemical changes under potentiodynamic stresses. Similar peaks were found in a previous study of ITO coated glass.⁶⁷ A relatively large peak-to-peak separation of 0.2 V (between 0.1 and 0.3 V) is due to this transformation (irreversible change) that occurs in the ITO film during electrochemical cycling. The DWNT LbL films exhibit no redox peaks, suggesting they are electrochemically stable in this potential range. Cyclic voltammograms of nitric acid-treated DWNT films were also collected as a function of the number of BLs deposited, as shown in Fig. 9(B). These rectangular shaped curves are indicative of the capacitive behavior found in carbon materials.⁵⁰ These voltammograms show little change before and after acid treatment, which suggests there is no chemical damage of the nanotubes. Integrated charge density from adsorbed and desorbed ions on the DWNT film was calculated by integrating these CV curves with potential differentials. Fig. 9(C) shows the linear increase in charge density as a function of the film thickness. This thickness-dependent electrochemical behavior demonstrates the tailorability of electrochemical behavior in these LbL assemblies by varying the number of BLs.

Fig. 9 (A) Cyclic voltammograms of a [PDDA/(DWNT + DOC)]₅ thin film on PET and a commercial 100 Ω sq⁻¹ ITO-coated PET. (B) Cyclic voltammograms of [PDDA/(DWNT + DOC)]_nassemblies deposited on PET. (C) Integrated charge density of each film, as a function of film thickness, was calculated from cyclic voltammograms and electrode area.

Conclusions

Highly transparent thin film electrodes were assembled through the alternate exposure of flexible PET substrates to aqueous mixtures of positively-charged poly(diallyldimethyl ammonium chloride) and CNTs, stabilized with negatively-charged deoxycholate. Double-walled carbon nanotubes were substituted for SWNTs in an effort to achieve lower sheet resistance in these flexible transparent electrodes. The layer-by-layer film growth is linear for these nanotube-based thin films. DWNT thin films exhibit a significant increase in electrical conductivity, and simultaneous decrease in sheet resistance, after exposure to nitric acid vapor due to removal of insulating material (polymer and surfactant). Additionally, these DWNT LbL coatings on PET substrates have excellent flexibility without any loss of conductivity after 100 bending cycles, unlike commonly used ITO. The sheet resistance, transparency, mechanical flexibility, and electrochemical stability of these CNT-based LbL assemblies meet the criteria for ITO replacement for most electronics applications (%T >85 and R_S <500 Ω sq⁻¹). Achieving these exceptional properties with fewer than five bilayers is also important in terms of being commercially viable.

Acknowledgements

The authors thank Bayer MaterialScience for financial support of this work. Ann Ellis, at the Microscopy and Imaging Center (MIC) at Texas A&M University, is paid special thanks for assistance with TEM sample preparation and imaging. The FE-SEM acquisition was supported by the NSF grant DBI-0116835, the VP for Research Office, and the Texas Engineering Experiment Station.

References

1. C. G. Granqvist and A. Hultaker, Thin Solid Films, 2002, 411, 1–5.
2. L. Hu, D. S. Hecht and G. Gruner, Chem. Rev., 2010, 110, 5790–5844.
3. C. G. Granqvist, Sol. Energy Mater. Sol. Cells, 2007, 91, 1529–1598.
4. E. Valkonen, B. Karlsson and C. G. Ribbing, Sol. Energy, 1984, 32, 211–222.
5. C. G. Granqvist, Appl. Phys. A: Solids Surf., 1993, 57, 19–24.
6. X. Crispin, S. Marciniak, W. Osikowicz, G. Zotti, A. W. D. van der Gon, F. Louwet, M. Fahlman, L. Groenendaal, F. De Schryver and W. R. Salaneck, J. Polym. Sci., Part B: Polym. Phys., 2003, 41, 2561–2583.
7. S. Marciniak, X. Crispin, K. Uvdal, M. Trzcinski, J. Birgerson, L. Groenendaal, F. Louwet and W. R. Salaneck, Synth. Met., 2004, 141, 67–73.
8. Y. Igasaki and H. Saito, J. Appl. Phys., 1991, 69, 2190–2195.
9. H. S. Randhawa, M. D. Matthews and R. F. Bunshah, Thin Solid Films, 1981, 83, 267–271.
10. H. Nanto, T. Minami, S. Shooji and S. Takata, J. Appl. Phys., 1984, 55, 1029–1034.
11. F. L. Wong, M. K. Fung, S. W. Tong, C. S. Lee and S. T. Lee, Thin Solid Films, 2004, 466, 225–230.
12. Y. Hoshi and T. Kiyomura, Thin Solid Films, 2002, 411, 36–41.
13. Y. F. Lan, W. C. Peng, Y. H. Lo and J. L. He, Mater. Res. Bull., 2009, 44, 1760–1764.
14. P. Nath, R. F. Bunshah, B. M. Basol and O. M. Staffsud, Thin Solid Films, 1980, 72, 463–468.
15. R. G. Gordon, MRS Bull., 2000, 25, 52–57.
16. H. R. Fallah, M. Ghasemi, A. Hassanzadeh and H. Steki, Mater. Res. Bull., 2007, 42, 487–496.
17. M. Boehme and C. Charton, Surf. Coat. Technol., 2005, 200, 932–935.
18. Y. C. Lin, W. Q. Shi and Z. Z. Chen, Thin Solid Films, 2009, 517, 1701–1705.
19. Y. F. Lan, W. C. Peng, Y. H. Lo and J. L. He, Org. Electron., 2010, 11, 670–676.
20. G. Gu, Z. L. Shen, P. E. Burrows and S. R. Forrest, Adv. Mater., 1997, 9, 725–728.
21. D. R. Cairns, R. P. Witte, D. K. Sparacin, S. M. Sachsman, D. C. Paine, G. P. Crawford and R. R. Newton, Appl. Phys. Lett., 2000, 76, 1425–1427.
22. Y. Leterrier, L. Medico, F. Demarco, J. A. E. Manson, U. Betz, M. F. Escola, M. K. Olsson and F. Atamny, Thin Solid Films, 2004, 460, 156–166.
23. S. K. Park, J. I. Han, D. G. Moon and W. K. Kim, Jpn. J. Appl. Phys., Part 1, 2003, 42, 623–629.
24. S. Iijima, Nature, 1991, 354, 56–58.
25. Z. C. Wu, Z. H. Chen, X. Du, J. M. Logan, J. Sippel, M. Nikolou, K. Kamaras, J. R. Reynolds, D. B. Tanner, A. F. Hebard and A. G. Rinzler, Science, 2004, 305, 1273–1276.
26. D. H. Zhang, K. Ryu, X. L. Liu, E. Polikarpov, J. Ly, M. E. Tompson and C. W. Zhou, Nano Lett., 2006, 6, 1880–1886.
27. Q. L. Williams, X. Liu, W. Walters, J. G. Zhou, T. Y. Edwards and F. L. Smith, Appl. Phys. Lett., 2007, 91, 143116.
28. W. J. Ma, L. Song, R. Yang, T. H. Zhang, Y. C. Zhao, L. F. Sun, Y. Ren, D. F. Liu, L. F. Liu, J. Shen, Z. X. Zhang, Y. J. Xiang, W. Y. Zhou and S. S. Xie, Nano Lett., 2007, 7, 2307–2311.
29. H. Z. Geng, K. K. Kim, K. P. So, Y. S. Lee, Y. Chang and Y. H. Lee, J. Am. Chem. Soc., 2007, 129, 7758–7759.
30. B. Dan, G. C. Irvin and M. Pasquali, ACS Nano, 2009, 3, 835–843.
31. X. Yu, R. Rajamani, K. A. Stelson and T. Cui, Surf. Coat. Technol., 2008, 202, 2002–2007.
32. P. Bertrand, A. Jonas, A. Laschewsky and R. Legras, Macromol. Rapid Commun., 2000, 21, 319–348.
33. P. T. Hammond, Adv. Mater., 2004, 16, 1271–1293.
34. K. Ariga, J. P. Hill and Q. M. Ji, Phys. Chem. Chem. Phys., 2007, 9, 2319–2340.
35. G. Decher and J. B. Schlenoff, Multilayer thin films: sequential assembly of nanocomposite materials, Wiley-VCH, Weinheim, 2003.
36. H. S. Seo, S. E. Kim, J. S. Park, J. H. Lee, K. Y. Yang, H. Lee, K. E. Lee, S. S. Han and J. Lee, Adv. Funct. Mater., 2010, 20, 4055–4061.
37. J. Park, L. Fouché and P. Hammond, Adv. Mater., 2005, 17, 2575–2579.
38. H. Koo, D. Yi, S. Yoo and D. Y. Kim, Adv. Mater., 2004, 16, 274–277.
39. Y. T. Park and J. C. Grunlan, Electrochim. Acta, 2010, 55, 3257–3267.
40. T. J. Dawidczyk, M. D. Walton, W. S. Jang and J. C. Grunlan, Langmuir, 2008, 24, 8314–8318.
41. Y.-C. Li, J. Schulz, S. Mannen, C. Delhom, B. Condon, S. Chang, M. Zammarano and J. C. Grunlan, ACS Nano, 2010, 4, 3325–3337.
42. M. A. Priolo, D. Gamboa and J. C. Grunlan, ACS Appl. Mater. Interfaces, 2010, 2, 312–320.
43. T. R. Hendricks, J. Lu, L. T. Drzal and I. Lee, Adv. Mater., 2008, 20, 2008–2012.
44. M. Osada, Y. Ebina, H. Funakubo, S. Yokoyama, T. Kiguchi, K. Takada and T. Sasaki, Adv. Mater., 2006, 18, 1023–1027.
45. Y. T. Park, A. Y. Ham and J. C. Grunlan, J. Phys. Chem. C, 2010, 114, 6325–6333.
46. B. S. Shim, Z. Y. Tang, M. P. Morabito, A. Agarwal, H. P. Hong and N. A. Kotov, Chem. Mater., 2007, 19, 5467–5474.
47. H. J. Park, K. A. Oh, M. Park and H. Lee, J. Phys. Chem. C, 2009, 113, 13070–13076.
48. Y. T. Park, A. Y. Ham and J. C. Grunlan, J. Mater. Chem., 2011, 21, 363–368.
49. B. S. Shim, J. Zhu, E. Jan, K. Critchley and N. A. Kotov, ACS Nano, 2010, 4, 3725–3734.
50. S. W. Lee, B. S. Kim, S. Chen, Y. Shao-Horn and P. T. Hammond, J. Am. Chem. Soc., 2009, 131, 671–679.
51. A. A. Green and M. C. Hersam, Nat. Nanotechnol., 2009, 4, 64–70.
52. T. Shimada, T. Sugai, Y. Ohno, S. Kishimoto, T. Mizutani, H. Yoshida, T. Okazaki and H. Shinohara, Appl. Phys. Lett., 2004, 84, 2412–2414.
53. W. Wenseleers, I. I. Vlasov, E. Goovaerts, E. D. Obraztsova, A. S. Lobach and A. Bouwen, Adv. Funct. Mater., 2004, 14, 1105–1112.
54. R. Haggenmueller, S. S. Rahatekar, J. A. Fagan, J. H. Chun, M. L. Becker, R. R. Naik, T. Krauss, L. Carlson, J. F. Kadla, P. C. Trulove, D. F. Fox, H. C. DeLong, Z. C. Fang, S. O. Kelley and J. W. Gilman, Langmuir, 2008, 24, 5070–5078.
55. L. M. Ericson, H. Fan, H. Q. Peng, V. A. Davis, W. Zhou, J. Sulpizio, Y. H. Wang, R. Booker, J. Vavro, C. Guthy, A. N. G. Parra-Vasquez, M. J. Kim, S. Ramesh, R. K. Saini, C. Kittrell, G. Lavin, H. Schmidt, W. W. Adams, W. E. Billups, M. Pasquali, W. F. Hwang, R. H. Hauge, J. E. Fischer and R. E. Smalley, Science, 2004, 305, 1447–1450.
56. W. Zhou, J. Vavro, N. M. Nemes, J. E. Fischer, F. Borondics, K. Kamars and D. B. Tanner, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 205423.
57. H.-Z. Geng, K. K. Kim, C. Song, N. T. Xuyen, S. M. Kim, K. A. Park, D. S. Lee, K. H. An, Y. S. Lee, Y. Chang, Y. J. Lee, J. Y. Choi, A. Benayad and Y. H. Lee, J. Mater. Chem., 2008, 18, 1261–1266.
58. X. Zhang, T. V. Sreekumar, T. Liu and S. Kumar, J. Phys. Chem. B, 2004, 108, 16435–16440.
59. U. Dettlaff-Weglikowska, V. Skákalová, R. Graupner, S. H. Jhang, B. H. Kim, H. J. Lee, L. Ley, Y. W. Park, S. Berber, D. Tománek and S. Roth, J. Am. Chem. Soc., 2005, 127, 5125–5131.
60. M. Kaempgen, G. S. Duesberg and S. Roth, Appl. Surf. Sci., 2005, 252, 425–429.
61. R. Wang, J. Sun, L. Gao and J. Zhang, ACS Nano, 2010, 4, 4890–4896.
62. R. Graupner, J. Abraham, A. Vencelova, T. Seyller, F. Hennrich, M. M. Kappes, A. Hirsch and L. Ley, Phys. Chem. Chem. Phys., 2003, 5, 5472–5476.
63. In Material Safety Data Sheet, Sigma-Aldrich Corporation, http://www.sigmaaldrich.com/safety-center.html, accessed 02/21/2011.
64. D. Q. Yang, J. F. Rochette and E. Sacher, J. Phys. Chem. B, 2005, 109, 4481–4484.
65. H. J. Chen, Y. L. Wang, Y. Z. Wang, S. J. Dong and E. K. Wang, Polymer, 2006, 47, 763–766.
66. T. J. Simmons, J. Bult, D. P. Hashim, R. J. Linhardt and P. M. Ajayan, ACS Nano, 2009, 3, 865–870.
67. M. Senthilkumar, J. Mathiyarasu, J. Joseph, K. L. N. Phani and V. Yegnaraman, Mater. Chem. Phys., 2008, 108, 403–407.

---

Footnotes

† Electronic Supplementary Information (ESI) available: Photograph of CNT dispersion by DOC, photograph of the DWNT LbL films on PET with different bilayers, UV-vis analysis, SEM images of DWNT and SWNT films, EDX spectra of DWNT and SWNT films, before and after acid treatment. See DOI: 10.1039/c1ra00225b

‡ Present address: Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, MN 55455, USA.

---