Abstract

We present a comparative study between multiwalled carbon nanotubes (MWCNTs) thin films deposited on polyethylene terephthalate (PET) substrates using (i) spin-coating technique and (ii) deposition through a membrane. We deduce from transparence, electrical properties, and AFM image that deposition through membrane presents better properties than spin-coating method. The concentration comparison shows that the optimum result was achieved at a concentration of 1.2 mg·mL−1 corresponding to a resistance () of 180 Ω·cm−2 and an optical transparence of about 81% using a wavelength 550 nm. We will also demonstrate the use of the elaborated electrodes to fabricate the following flexible structure: PET-MWCNTs/MEH-PPV/Al. The series resistance and the ideality factor were calculated.

1. Introduction

Actually, transparent conductive films are extremely common and critically important in optoelectronic devices. They are used as electrodes for E-readers and digital cameras [1], photovoltaic devices such as solar cells [2], and organic light-emitting diodes (OLEDs) [3]. Currently, tin doped indium oxide (ITO) presents the dominant transparent conductive material with growing annual demand at 20% [4]. ITO has been studied and refined for over 70 years. As a result, the material offers many beneficial properties. However, ITO present certain inconveniences, mainly reflected on the depleted supply of raw materials and their brittleness. The supply of indium is constrained by both mining and geopolitical issues, which leads to dramatic price fluctuations over the last decades. The high price of indium determined the high cost of ITO, since they compose nearly 75 wt% of a typical ITO film [5]. The current devices are typically based on rigid substrates. As ITO tends to fracture at strains of 2%, it is completely unsuitable for using in flexible electronics. Therefore, new transparent electrode materials have rapidly emerged in recent years, including carbon nanotubes (CNTs), graphene, and metal nanowires. This material presents good conductivity coupled with aspect ratio which yields films with high transparence, adequate sheet resistance, and high mechanical flexibility [6, 7]. The intrinsically good conductivity coupled with aspect ratio yields films with high transparence, adequate sheet resistance, and high mechanical flexibility. These material properties, combined with the low costs, deposition, and the inexpensive material, make these emerging nanomaterials very attractive. Among the dominant nanoscale materials, CNTs are the most promising and mature materials intensively investigated. Indeed, CNTs exist in two types: the single-walled carbon nanotubes (SWCNTs) and the multiwalled carbon nanotubes (MWCNTs). Much research was devoted to the application of SWCNTs to optoelectronics for their good properties [69]. On the other hand, MWCNTs are easier to process than SWCNTs; MWCNTs are also less liable to form tight clusters and are comparatively cheaper. Also, MWCNTs are even more suitable for the implementation of charge transport and charge transfer, because of the metallic conductivity and the predictable HOMO–LUMO energy levels [10]. They are employed as hole-injecting electrodes or charge transport layers [1116]. For those reasons, we will use MWCNTs as flexible electrodes. We will compare two types of flexible electrodes based on MWCNTs, elaborated by spin coating and through a membrane. Then, we will demonstrate their use as transparent electrodes for the following structure: PET-MWCNTs/MEH-PPV/Al.

2. Experimental Details

MWCNTs were dispersed in deionized water containing 1 wt% sodium dodecyl sulfate (SDS). Then, MWCNTs coated electrodes were prepared using two methods in order to create uniform networks of MWCNTs on flexible substrates. Briefly, MWCNTs were vacuum-filtered on mixed cellulose ester membranes (MCE) and then transferred onto PET substrates (described in our previous work) [16]. MEH-PPV polymer was dispersed in THF solvent at a concentration of 15 mg·mL−1 and then spin-coated on the PET-MWCNTs at an angular speed of 1500 rpm for 10 s. Finally, aluminum top electrodes were deposited by thermal evaporation through a shadow mask.

3. Results and Discussion

3.1. Morphological Study by AFM

In order to explore the topography of the CNTs, AFM studies should be performed. We present in Figure 1 the AFM images of CNTs anode deposited by spin coating at concentrations of (a) 1.2 mg·mL−1 and (b) 2.5 mg·mL−1 and through a membrane adopting the same concentrations (c and d, resp.). The AFM images of spin-coated film show that CNTs are condensed on themselves without any specified direction. In addition, we observe that aggregates domains are formed (Figure 1(b)). This result is due to the Van Der Waals forces among the tubes and their hydrophobicity. We note that interaction between different tubes is the origin of the aggregations observed [17]. This is expected due to the deposition method used to put the CNTs on PET substrate.

Using the second method, CNTs become clearly observed. AFM images exhibit that CNTs networks are more densified and characterized by a network of cylindrical-like features, and the approximate length is about several hundreds of nanometers and the width is much smaller. These well-organized overlapping structures may result from the capillary force and the force induced by the evaporation; it is like the photonic crystals formation mechanism. This alignment can lead to better performance for optoelectronics devices [18]. In addition, the space area unoccupied by CNTs is reduced. This leads to obtaining a solid conductive film and a good optical transparency level. Indeed, for the concentration of 1.2 mg·mL−1, interconnected CNTs looked like a spider network. This could create a rapid charge transport between the interconnecting conductive paths of MWCNTs and a high optical transparency level due to well-developed structures [19].

3.2. Electrical and Optical Properties

An encouraging optical properties’ result, comparable to that obtained for the ITO, was found with 0.8 mg·mL−1 and 1.2 mg·mL−1. However, the CNTs network retains high transparency to the visible and near infrared range of the electromagnetic spectrum. Obviously, we found that CNTs transmission is mainly due to the absorbance inside film [20]. Indeed, for the other two concentrations, transparency measurement shows that these films are opaque. This interpretation is correlated with AFM morphology. Accordingly, these structures can be good candidates to be integrated as anodes in photovoltaic devices. The variation of the surface resistance and the transmission are shown in Figure 2, for MWCNTs anode deposited by spin coating. Obviously, it is important to note that the distribution of MWCNTs plays a crucial role for optical transparency measurements and for networks electrical conductivity.

Indeed, the increase of MWCNTs concentration leads to the decrease of ; thicker layers (low transmission) exhibit better electrical properties due to the increase of the percolation path number so that electrons can move through. However, for the concentration 0.8 mg·mL−1, PET-MWCNTs films exhibit a high resistance in the order of 2.21 KΩ·cm−2 and an optical transmission of 91% at 550 nm. For the concentration 2.5 mg·mL−1, decreased significantly to 580 Ω·cm−2 with an optical transmission of 20%. The resistance decrease was effective because of charge transport percolation paths through the well-interconnected networks [21]. The resistance is high compared to the result obtained with the method using an ITO substrate. Consequently, we cannot use them as transparent electrodes in optoelectronic devices.

Transparency and conductivity variations of flexible anode elaborated through a membrane are shown in Figure 2. For the concentration 2.5 mg·mL−1, PET-MWCNTs films present a high resistance in the order of 587 Ω.cm−2 and an optical transmission of 43%. For the concentration 1.2 mg·mL−1, sheet resistance decreased significantly to 180 Ω·cm−2 with an optical transmission approximately 81% at 550 nm. This can satisfy the requirement for the application in touch screen [22] and is comparable to most of the literature results [23]. Obviously, it is noted that the transmission and the surface resistance depend mainly on the distribution of MWCNTs on the substrate. Indeed, it is clear that increase of MWCNTs concentration leads to the decrease of . This can be attributed to three major factors [24, 25]: (i) the intrinsic conductivity of MWCNTs, (ii) contact resistance tube-tube, and (iii) the concentration and the distribution of MWCNTs on the PET substrate.

The SDS covering the MWCNTs surface increases the contact resistance. Hence, it is required to remove SDS residues after films elaboration. Cui et al and Wang et al. [26, 27] studied the effect of the removal of residual SDS from CNTs network and they showed that the conductivity increase is mainly due to the effective removal of residual SDS and not from a chemical doping effect. In contrast, we made a similar experience with our film. Indeed, we washed the films with chlorobenzene, acetone, and pure water. Thus, we note an improvement in conductivity due to the better connection between the MWCNTs. Our results are comparable to other similar works reported in the literature. Ko et al. [21] and Castro and Schmidt [28] found an resistance of 370 Ω·cm−2 with a transmission of 77% and of 20 KΩ·cm−2 with a transmission of 78%, respectively.

3.3. Electrical Properties of PET-MWCNTs/MEH-PPV/Al Structures
3.3.1. Current Density-Voltage Characteristic

In order to demonstrate the efficiency of our flexible electrodes, we elaborate the following structure: PET-MWCNTs/MEH-PPV/Al. Figure 3 shows their energy diagram with a picture of the elaborated structure.

Current density-voltage characteristic of PET-MWCNTs/MEH-PPV/Al is shown in Figure 4. The applied voltage during the measurement varies from −3 to +3 V. The positive polarity corresponds to side of the PET-MWCNTs electrode. Consider the output work values of the MWCNTs and aluminum, which are, respectively, about 4.9 eV and 4.3 eV and from the gap energy of the MEH-PPV conjugated polymer ( = 5 eV and = 2.8 eV). We conclude that the holes injection barrier (~0.1 eV) in the MEH-PPV polymer is localized at the MWCNTs/MEH-PPV interface. This barrier height allows charge injection by thermoionic emission. An Ohmic behavior is obtained for the structure elaborated on the MWCNTs using spin coating. These results could be explained by the low thickness of the active layer which allows direct connections (short circuit) between the cathode and the anode. However, we always find the same Ohmic behavior when using other film thicknesses obtained with different deposition rates. This can also be related to the MWCNTs orientations, which present a several micrometers of length. On the other hand, for the structure prepared on the MWCNTs elaborated through membrane, the curve is nonlinear and it has a rectifier behavior due to the injection charge carriers from MWCNTs electrode to the active layer. We obtain an organic diode behavior with a threshold voltage close to 1.9 V. Our results are comparable to many other published studies [29].

The conduction mechanisms that control OLEDs behavior are determined from forward bias. The current density-voltage characteristics in logarithmic scale for PET-MWCNTs/MEH-PPV/Al structures are shown in Figure 4. Indeed, the structure elaborated on the MWCNTs elaborated by spin coating shows a single linear region. The slope value is one, indicating an Ohmic conduction. Therefore, we cannot use the PET-MWCNTs electrode prepared by spin coating as anode. On the other hand, for the structure using PET-MWCNTs anode elaborated through membrane, the curve shows the presence of two regions. We note a transition from a linear to a quadratic variation. In the first region ( V), conduction is due to the thermally generated intrinsic charge carriers and the slope value equal to 1 indicates an Ohmic conduction. The second region with a slope value equal to 2 indicates that the current is limited by space charge (SCLC) with a single level traps.

3.3.2. Calculation of Series Resistance by J-J (dV/dJ) Method

J-V characteristic curvature (Figure 4) for forward bias and sufficiently large tension is caused by the presence of series resistance [30]. The series resistance and the ideality factor are determined from Figure 5 using a method developed in [31].

At the Schottky contact formed in the semiconductor/metal interface, the current is supposed due to thermoionic emission. By adopting this theory, the current is expressed asThe following equation presents voltage versus current:Taking into account the voltage expression, the voltage derivative with respect to current is given byFor large values, the total current is greater than the saturation current (). So leads toWith these assumptions, depends linearly on . Therefore, the series resistance can be easily obtained from the slope of the region corresponding to the high current by a linear fit. characteristic is shown in Figure 5. We obtained the following parameters: for the structure prepared on the MWCNTs elaborated by spin coating, the series resistance = 312 Ω and ideality factor . For the structure prepared on the MWCNTs elaborated through membrane, the series resistance  Ω and ideality factor .

For the structure prepared on the MWCNTs elaborated by spin coating, the series resistance is approximately in a high order. This value is attributed to the anode morphology, the contact resistance and the nature of the PET-MWCNTs/MEH-PPV interface. For the structure fabricated on the MWCNTs elaborated through membrane, the series resistance is approximately in a middle order. This value is allocated to (i) more surface heterojunction due to the interpenetration of MEH-PPV and MWCNTs, and (ii) the charge transporting paths are relatively long [32]. In addition, the structure was prepared in air and left in ambient conditions. Therefore, the series resistance value can be affected by many factors, such as reduced mobility and the changes in the contact barrier located at the space charge regions [33, 34]. The ideality factor is an important parameter in the specification of the electrical behavior of the diodes which is assigned to the voltage drop in the interfacial layer [35]. This is probably due to the charge overcrowding and the high probability of electrons and holes recombination in the depletion region [36].

3.4. Overview

Our results show that the anode/organic interface plays a crucial role to determine the devices performance. Holes injection process at PET-MWCNTs/MEH-PPV interface can be considered as one of the main factors that control the electrical characteristics and performance of the device. It has been shown that holes injection depends on several factors such as the height energy barrier at the anode/organic interface. This energy is determined by the difference between the anode work function and the HOMO level of the organic layer according to the energy levels of the diagram given in Figure 1. After the introduction of the PET-MWCNTs electrode interface, the decrease of holes injection barrier height justifies the improvement of current-voltage characteristics observed in the modified structure. Therefore, we obtain better diode performance. It is also interesting to note that the improvement of electrical properties highlights the lack of dipoles that can be formed at the anode/organic interface and reduces the diode efficiency [37]. The injection barrier height is not the only factor that controls the holes injection. Indeed, the variation of the anode surface morphology can be also considered as an important factor: a rougher interface can facilitate the charge carriers exchange. Therefore, we can recapitulate that the nature of the surface morphology dependent on MWCNTs film deposits method. It is directly correlated with the decrease of the barrier height. The improvement of the interface electrical parameters using a PET-MWCNTs electrode, elaborated through a membrane, is attributed to the good surface morphology and the contact resistance. In fact, the contact resistance is a significant factor contributing to the decrease of the series resistance. This is correlated with the reduction of the hole barrier height.

4. Conclusion

The AFM analysis made on CNTs deposited by both methods allowed us to conclude that adopting the second deposition procedure we obtain an oriented and homogeneously distributed CNTs network. The measurement results of the resistivity and the optical transmission show that these electrodes can be used as anodes in flexible optoelectronic devices. Using the method, we calculated the ideality factor and the series resistance. The results indicate that the holes injection process at the PET-MWCNT/MEH-PPV interface is effective. The origin of these improvements is mainly explained by the band structure alignment leading to reduction of the height barrier.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.