polymer-carbon nanotube composites: basic science ...

50 downloads 0 Views 716KB Size Report
Amaratunga, Journal of Applied Physics 93, 1764. (2003). ... E. Kymakis and G. A. J. Amaratunga, Single-wall ... Molecules and Polymers (John Wiley and Sons,.
MATERIALS FORUM VOLUME 32 - 2008 Edited by J.M. Cairney, S.P. Ringer and R. Wuhrer © Institute of Materials Engineering Australasia Ltd

POLYMER-CARBON NANOTUBE COMPOSITES: BASIC SCIENCE AND APPLICATIONS J.M. Bell1, R.G.S. Goh1*, E.R. Waclawik2, M. Giulianini and N. Motta1 1.

Faculty of Built Environment and Engineering, Queensland University of Technology, GPO Box 2434, Brisbane, AUSTRALIA, 4001 2. Inorganic Materials Research Program, School of Physical & Chemical Sciences, Queensland University of Technology, GPO Box 2434, Brisbane, AUSTRALIA, 4001

ABSTRACT Composites of conjugated polymers are becoming increasingly used in organic electronics devices, owing to the electronic and optical properties of the polymers, which are similar to those of semiconductors. However the mechanical properties and processibility of polymers makes them more attractive than crystalline inorganic semiconductors for future applications: their low weight, flexibility and inexpensive preparation procedures are drawing the attention of many researchers. This paper explores some of the fundamental science associated with the use of single-walled and multi-walled carbon nanotubes as additives in poly(alkylthiophene) polymers in order to improve their electrical properties. The particular issues of interest are the dispersion of the carbon nanotubes in the polymers, and the interaction of carbon nanotubes with the polymers, where evidence of templating of the polymers onto the carbon nanotubes has been seen using atomic force microscopy. The impact of the nanotubes on the electrical properties of the composites is also critical to the use of the composites in optoelectronic devices. Recently interest in polymer composites has turned to their potential use in electronic applications, in socalled organic electronics. Not long after their discovery by MacDiarmid, Shirikawa and Heeger3, conjugated or conductive polymer materials were quickly identified as promising candidates for use as the active component of a variety of electronic devices, such as low-cost alternatives to conventional lightemitting diodes (LEDs), photovoltaic cells and disposable electronic chips4. In the case of organic photovoltaics (OPVs), incorporation of a second material or dopant into the conjugated polymer to create a composite has led to great improvements in device performance compared to polymer-only devices5,6. Efficient charge transport in these composites is contingent upon intimate contact between the second phase within the polymer at the nanoscale. The organic semiconductor composite OPV active layer is typically cast in the form of a bicontinuous network so that it operates as an efficient photogenerated-charge separation and transport material, otherwise isolated domains form which trap charge, leading to higher series resistances and thus low photocurrents6. Recently, composites of conjugated polymers and carbon nanotubes (CNT) have been proposed for this application with the aim of improving the performance in electronic applications like organic electronics and photovoltaics7,8 relying on the high electron conductivity of carbon nanotubes. Purified nanotubes have an extremely high surface area ~1600 m2/g and this also offers an opportunity for efficient exciton dissociation. The use of carbon nanotubes as reinforcement filler-components is becoming more widespread for mechanical applications – increasing the strength of polymeric materials – however as elements in electronic

1. INTRODUCTION Polymer composite materials are used in a great variety of applications since they often possess more desirable mechanical properties than granular or ceramic composites, including increased tensile strength, elastic modulus and therefore flexibility. In the area of polymer composite materials, the area of fiberreinforced composites is a long established field1. Conjugated polymer materials are of particular interest to scientists and engineers for a number of reasons, not least because they can be prepared with similar electrical and optical properties to semiconductors or even metals, while still retaining the attractive mechanical properties and processing advantages of polymers2. In the case of pure conjugated polymers, experimental studies have established direct correlations between electrical conductivity and mechanical properties of Young’s modulus and tensile strength. That electrical properties and mechanical properties of conducting polymers improve together, in a correlated manner, as chain extension, chain alignment and inter-chain order are improved, reveals the important influence that intermolecular interactions, self-assembly and nanoscale structure have on such physical properties2. It is therefore not surprising that the properties of composites of conjugated polymers which have been formed by inclusion of either a second macromolecule or nanometer-sized inorganic within the material should be of interest to device scientists and engineers. New nanoscale polymer structures can form in the presence of these inserted materials that lead to new and interesting physical properties.

144

applications, CNTs have not found many applications outside of the research laboratory to date. This is due to a number of telling factors and constraints that must first be overcome.

1.2. The problem of Bundles The huge aspect ratios of CNTs results in very large attractive non-covalent forces between CNTs, this causes them to aggregate together into bundles that are difficult to separate. If the polymer composite application requires an even dispersion and distribution of separated CNTs (most applications do) then attention to debundling CNT samples is necessary. Again sonication is the method that has commonly been applied to this task, sometimes in conjunction with surfactant addition in solution and the judicious choice of solvent10.

1.1. Purification and Solubility Depending upon the type of carbon nanotube used, single-walled (SWNT), double-walled (DWNT), or multi-walled (MWNT) and the chemical functionality of the carbon nanotubes both along their basal-plane sidewalls and tip-ends, varying degrees of solubility and dispersion can be attained in common organic solvents. Additionally, longer carbon nanotubes are more difficult to effectively disperse in a solvent using the conventional method of ultrasonic agitation9. Performing chemistry on CNTs in order to better disperse them in a particular solvent for composite casting can lead to a reduction in the desired electrical performance of treated CNTs compared to a theoretical defect-free SWNT. Electronic applications for carbon nanotubes also often require the product to be contaminant-free and defect-free. Most carbon nanotubes are synthesized at high temperature in the presence of a transition metal catalyst such as Fe, Mo, Ni or Co. Regardless of the synthesis technique used – chemical vapour deposition (CVD) and arc-discharge (AD) being the two most common - the raw carbon nanotube-containing soot must be purified of both amorphous, graphitic carbon and the metal catalyst. This usually involves partial digestion of the product in a mixture of sulfuric and nitric acid which leaves the carbon nanotubes somewhat shortened, damaged and end-functionalized with carboxylic acid and amide moieties. High-temperature annealing of the purified product can to some extent repair CNT sidewall damage. All three types of carbon nanotubes SWNT, DWNT, MWNT are thus comparatively intractable materials and careful thought needs to go into selection of the carbon nanotube source, the purification procedure applied and the solvent system used for casting a polymer composite for high-end electronic applications.

1.3. The Polymer-CNT Interfacial Interactions The interaction between nanotubes and polymer at a molecular level is still not completely understood, and several studies are underway to clarify this issue11-14. As nanotube samples are not all equal, but have a distribution of diameters and different chiralities (the direction in which the graphite sheet is rolled), their interaction with a polymer host can vary. This can lead to a range of aggregation and crystallisation behaviours. The ratio of amorphous to crystalline phase in polymer composites is significantly affected by both the processing history and by these interfacial interactions. Depending on the polymer stiffness and nanotube properties, strands of polymer can be wrapped more or less tightly to nanotubes14, or grow as dendrimers on the nanotube surface11. This creates a full range of composites with properties which can be very different from those of the original polymer. Single-walled carbon nanotubes15 have great promise for applications in flexible electronics, being the base of thin film transistors16, FET (CNFET)17 and similar devices, but it has been very difficult, if not impossible, to prepare flexible high-performance integrated circuits based on them. Recently innovative solutions to this problem have been report, by growing dense arrays of aligned nanotubes on a crystalline quartz substrate, and then transferring the arrays onto plastic materials to make flexible, high-performance, high-power electronic devices18, or by using the liquid crystalline behaviour of CNT solutions and a tilted patterned substrate to create aligned stripes of nanotubes16.

CNT Distributions: Because CNTs are macromolecules themselves, synthesis yields a distribution of carbon nanotube lengths and widths, this also has consequences in a polymer composite context. Individual SWNTs are metallic or semiconducting depending on the CNT diameter (and also chiraltiy), so incorporation of a CNT sample into a polymer yields mixtures containing both types. If one CNT-type only is a prerequisite for the electronic application, this needs to be seriously considered. Some methods have been employed to separate metallic from semiconducting SWNTs, separation has been achieved using the physical method of capillary dielectrophoresis, but only small quantities have been produced. Chemical methods to functionalise an entire CNT ensemble and thus impose semiconducting character have been successfully achieved.

In this paper, we review our experience with processing disordered CNT composites containing different loadings of MWNTs cast from solution in the presence of the polymer host poly(3-hexyl thiophene), or P3HT. Even in its undoped form, regioregular poly(3-hexyl thiophene), rrP3HT can have quite a high charge mobility 0.1 V/cm2 making it a promising candidate material for organic electronic device applications. Early reports of relatively efficient photovoltaic sunlight-to electrical energy conversion measured in a SWNT-poly(3-octyl thiophene) thin film OPV has led a number of groups to explore these types of systems for OPV applications19-21. In an effort to produce an effective CNT-rrP3HT composite for an OPV application we have studied the influence of CNT 145

purity, CNT-debundling, CNT-loading and solution processing conditions on bulk electrical conductivity in these composites. In order to better understand the structure-property relationship between the components we have examined structure properties of this system at the nanoscale and spectroscopic and thermal properties of macroscopic samples.

b)

2. CHARACTERISATION AND PURIFICATION OF CNT SAMPLES Two CNT types were investigated, SWNTs and MWNTs. SWNT samples were purchased from Carbolex. Thermoanalytical and electron microscopic studies were used as characterisation tools for the determination of the composition of the single walled carbon nanotube samples. A purification method proved to be effective, resulting in a three fold increase in the percentage of SWNTs present in the purified product as determined by thermogravimetric analysis. The as bought carbon nanotubes employed in this work were of stated 50-70 vol% purity. However, it is reported to be very difficult22 to obtain after the purification procedures, the expected amount of nanotube from the commercial raw material. The purpose of this work was to attain a better understanding of the exact composition of purified and unpurified SWNT matrices through thermogravimetric analysis. The TG method of high resolution thermogravimetry (HRTG)23,24 were used to this end.

c)

Figure 1: SEM images of a) and b) as-prepared and c) purified SWNT. The SEM Figure 1(a) focuses on a gap in the carbonaceous film of an unpurified SWNT sample. The carbon nanotube bundles imaged in Figure 1(a), appear to be very thin (or even isolated nanotubes) and span the micron-sized crack in the film. Figure 1(b) shows a region with evidence of larger particles and aggregates that appear to have clumped together into nanotube ropes. These are similar to typical images found in literature of as-prepared soot of SWNT.

Purification of single walled carbon nanotubes was undertaken using a method adopted from Furtado et al2521. The as-prepared SWNTs (Carbolex) was heated in a furnace under air at 395 °C for 20 minutes. After this stage of dry oxidation the material was refluxed in 3.0 M HCl for 4 hours. The acidified dispersion was then filtered through track-etched polycarbonate membranes and copiously washed with ultra-pure water (Milli-Q). The filtrate was twice resuspended in water, sonicated, filtered and washed with ultra-pure water. These procedures led to loss of material at each stage. The yield of the purification was 6 wt%. At least four different components are present in the raw material: amorphous and graphitic forms of carbon, carbon nanotubes and metal particles that are encapsulated by carbon shells. The SEM images of the as-prepared SWNTs in Figure 1 show two typical regions.

Conventional and high resolution thermogravimetry of an unpurifed SWNT and purified SWNT matrices revealed a large amount of information regarding the accurate quantitative measurements of the content of carbonaceous by-products, carbon nanotubes, and residual metal catalyst present in the materials. Purification of the SWNTs was found to increase the content of SWNTs in the matrix by nearly 300%. Short, thin MWNT samples purchased from Nanocyl (Nanocyl ® - 3150) used in this study were examined using Raman spectroscopy, transmission electron microscopy and thermogravimetry (TGA) to confirm the manufacturer’s stated purity. Analysis of transmission electron microscopy images of the MWNT samples confirmed the distribution of length, l and diameter, d in the material ( l = 0.1 − 10 µm, d = 10 nm , see Figure 2). Figure 2 and the analysis of approximately 50 tubes in several other TEM images confirm the supplier’s specification of the diameter distribution centred at 10 nm (Figure 2). Figure 2 shows the MWNT structure and the analysis of 15 tubes lead to the determination of distribution of number of wall in this sample: 8±3. Figure 3 and the analysis of other SEM images allow

a)

146

TGA analysis. Nanocyl MWNT samples used to make polymer composites were thus used without the requirement for further modification or purification.

the evaluation of the MWNT length as distributed around 1 µm. The manufacturers stated C purity > 95% and metal oxide impurity < 5% were confirmed by

a)

b)

c)

d) Figure 2: TEM images of MWNT and diameter distribution.

Figure 3: SEM images of MWNT

3. P3HT NANOTUBE MIXTURES – BASIC INTERACTIONS

molar mass and polydispersity, to name a few, are often linked together in non-trivial ways.

Experimental studies26 have shown that the electrical and mechanical properties of conducting polymers improve together in a correlated manner, as chain extension, chain alignment and inter-chain order are improved. Also the optoelectronic properties of P3HT films are highly sensitive to microstructure and film morphology which depend on many materials and processing variables. These variables, e.g. solvent type, temperature, film-casting technique, P3HT average

Scanning tunnelling microscopy (STM) analysis may therefore be used to reveal the significant influence of intermolecular interactions, self-assembly and nanoscale structure on important physical properties. It is imperative that close packing of the nanotube with conjugated polymer occurs at a molecular level in order to achieve efficient charge transport. The way in which polymers are attached to carbon nanotubes has been investigated by various authors, 147

leading to two contradictory conclusions: Lei et al.27 suggests that the arrangement is dominated by the matching of the graphene hexagonal structure with the detailed structure of the conjugated polymer, while Coleman and co-workers17,28 have pointed out that geometric factors and constraints imposed by a narrow cylindrical structure, like a SWNT on a conjugated polymer chain, may outweigh any lattice-resolved contributions to polymer ordering. To better understand this issue, we have studied by STM the organization of the conjugated polymer rrP3HT on SWNT and on highly oriented pyrolitic graphite (HOPG).

Figure 4 STM image of polycrystalline domains of rrP3HT on HOPG,Vbias = 0.814V, I = 0.100Na

3.1 Scanning tunnelling microscopy of rrP3HT on HOPG

Since a SWNT can be considered as a seamless rolledup sheet of rapheme, a comparison between STM measurements of P3HT/SWNT composites and P3HTcoatings on HOPG should be used to separate atomic surface structure from geometric factors ( rapheme curvature) in P3HT ordering. Hence, in order to understand the basic physics underlying polymer-CNT interactions we performed STM investigations of rrP3HT films cast from chloroform solutions onto HOPG14 Figure 4 presents an STM image that is recorded in the sub-micron range showing a large polycrystalline domain of rrP3HT on HOPG. Areas of rrP3HT occupy most of the image in a light contrast, whilst the darker areas of the image are most likely due to an amorphous/disordered polymer layer.

Figure 5. (top) image of rrP3HT monolayers on HOPG, (bottom) line profile analyses of the respective rrP3HT monolayers.

All mono-domains of polymer in the image appear interconnected, which is supported by the findings of Grevin and co-workers29.

4. PREPARATION OF SWNT/RRP3HT FILMS FOR STM

Figure 5(a) shows a rrP3HT monolayer film cast from a dilute chloroform solution where the bright domains are attributed to the conjugated backbone of the polymer, aligned clearly in rows on the HOPG surface.

Single-walled carbon nanotubes purchased from Carbolex have a mean diameter of 1.4nm and are reported to aggregate into bundles of typically 100-400 SWNTs that are approximately 1-5µm long. Carbon nanotube samples were purified using the method illustrated above and mixed to poly(3-hexylthiophene) (P3HT) by sonication in chloroform solutions.

The measured chain-to-chain distance obtained from the profile is 1.45 nm, within the range of those measured by Mena-Osterlitz et al.30.

We performed STM investigations of rrP3HT films cast from these solutions onto HOPG14. We investigated drop-cast films of rrP3HT solutions with low (~1µg/ml) and high (~4mg/ml) polymer

148

Figure 6. STM image of cast film shows nanotubes dispersed throughout the polymer matrix. (a) A 70x50 nm image reveals an ordered pattern on the nanotube body, along with a short tube of about 30nm; Vbias = 0.661V; I= 0.093Na . (b) Zoom of the short nanotube in (a) with cross section: the repeat distance along the SWNT long-axis is 1.66 nm. 5 RESULTS

Concentration, the latter being a concentration equivalent to those used for the rrP3HT/SWNT composites. Figure 6 shows a P3HT/SWNT composite imaged at various length-scales using STM. The high aspect ratio of SWNTs allows them to be easily located by STM, even when covered with rrP3HT polymer.

The device architecture of early OPVs closely mimicked inorganic heterojunction devices, they consisted of a bilayer of p-type and n-type organic semiconducting polymers sandwiched between two metal contacts. Charge separation of excitons photogenerated near the p-n junction interface ultimately led to a photocurrent. Unfortunately the sunlight-to-electrical conversion efficiency of bilayer OPVs were disappointingly low. The reasons for this were twofold. Unlike inorganic semiconductor devices which are doped with heteroatoms to produce majority and minority carriers, organic semiconductors are either intrinsically p-, or n-type. Since they are not doped, no internal field develops at the heterojunction, the requirement of an internal electric field thus requires metals (often oxygen-sensitive metals) possessing different workfunction to be deposited either side of the bilayer device6. The second and more important difference is that the molecular nature of the organic semiconductor leads to a HOMO/LUMO picture of electronic energy levels rather than extended valence and conduction bands. Excitons are strongly coupled to the molecular geometry of the polymer and possess a strong Coulomb interaction or exciton binding energy. Organic excitons are thus Frenkel excitons, the size of which rarely extend further than approximately three polymer repeat units (~10nm typically), the diffusion radius of these excitons is quite short (~5nm)31. A direct consequence of the short exciton diffusion radius is that only excitons generated within a few nanometers of the polymer bilayer interface have a high probability of being separated into carriers in these OPVs. The external quantum efficiencies (EQEs) – the efficiency of the process of turning incident light into electric current – of these devices are thus severely limited32. The ability to

4.1 Optimization of carbon nanotube dispersion in polymers

We have used Raman spectroscopy, thermogravimetric analysis, and STM measurements of absorbed conductive polymer monolayer structures on multi walled carbon nanotubes MWNTs and SWNTs and HOPG to test the dispersion of the nanotubes in the polymer matrix and to evaluate the influence of the substrate curvature on the self-assembly process. Thin and short multi walled carbon nanotubes (MWNTs) were used to prepare nanocomposites based on poly(3-hexylthiophene) (P3HT). MWNTs were characterized by TG, SEM, TEM and Raman spectroscopy. Stable dispersions of MWNT in chloroform were obtained. Non-covalent interaction between MWNT and P3HT dissolved in chloroform allowed the preparation of composites by dissolution followed by precipitation with methanol. Composite thermal events such as glass transition, melting temperature and heat of fusion were investigated and compared with pure polymer. P3HT may interact with the rapheme structure of the tubes producing a wrapping structure. The conductivity of 1wt% MWNT/P3HT composite is approximately 2x10-3S/cm.

149

process conjugated polymers from solution provides a way to overcome the problem of short exciton diffusion length. An electron donor can be combined with an electron acceptor to create a composite material which can operate as the OPV active layer. This distributed p-n junction approach to organic photovoltaic device engineering has been called the bulk heterojunction approach6.

is a second effect which can influence polymer crystal structure and that is the interactions between the dopant and polymer. If there is large mismatch between the polymer’s preferred crystal lattice structure and the crystal structure of the dopant material, the second phase can be a source of defects in polymer nanostructure. Where this occurs, high loadings of the non-polymeric component within the composite may reduce the conjugated polymer’s overall crystallinity and thereby decrease charge carrier mobility within the composite material. A high weight percent of the nonpolymeric component within the composite can also affect charge carrier mobility in other ways. Aggregation of the second phase can occur at high dopant concentrations when spin cast from solution35. It is thus quickly apparent that competing factors must be considered when designing organic semiconductor composites for use in bulk heterojunction OPVs. High loadings of the non-polymer component are needed to assure percolation of charge through the bicontinuous network, however the loadings must be low enough such that neither isolated aggregates of one component form, nor breakdown in polymer crystalline order occurs either.

Another approach to creating a heterojunction OPV is to cast a semiconducting polymer layer containing an inorganic material, as illustrated in Figure 7. A number of different polymer/inorganic material combinations have been investigated as heterojunction OPVs 4. To date the most efficient such bulk heterojunction OPVs have been formed from poly((2-methoxy-5-(2'ethylhexoxy)-p-phenylene) vinylene) (MEH-PPV) and soluble fullerene derivatives10. Optimum performance in these devices occurs at high fullerene loads (80%), which indicates that percolation is easier to achieve in the polymeric component than the non-polymeric component. Quantum efficiencies approaching 6% have been reported33.

Third, in order to obtain good Electron Quantum Efficiency (EQE) from a bulk heterojunction OPV, good overlap between the absorption spectrum of the conjugated polymer and the solar spectrum is desirable. The optical bandgap of a typical organic semiconducting polymer such as poly(phenylenevinylene or poly(3-hexylthiophene) occurs in the range 600 to 300 nm, where the electronic excitation is often coupled to a vibrational mode in the polymer36. The wavelength range over which light absorption by conjugated polymers occurs depends on the amount of conformational disorder of individual polymer chains within the cast films. A conjugated polymer can be considered to consist of a string of effectively conjugated segments (ECS) over which the polymer’s π-electron cloud maintains coherence. At the end of each effectively conjugated segment, one photophysical unit ends and another begins. The length of effective conjugated segments in a conjugated polymer can often be limited by conformational disorder, although this is not the only limiting factor. Entropy considerations mean that for large molecular weight polymers no individual polymer main-chain can have a perfectly ordered backbone, there will be some degree of conformational disorder (eg. chain-twists or polymer chain hairpin folds). Consequently, absorption spectra are a superposition of the spectra of ECSs of slightly different lengths. This causes inhomogeneous broadening of vibronic bands, increasing overlap with the solar spectrum6.

Figure 7. Scheme of the polymer-nanotube solar cell 5.1 Factors affecting external quantum efficiency in bulk heterojunction OPVs

Organic conjugated polymers are not completely crystalline; spin cast films contain both crystalline and amorphous regions. This can clearly be observed using x-ray diffraction, where the spectrum of these partially crystalline polymers display sharp peaks due to crystalline scattering superimposed over a broad background assigned to amorphous scattering34. Since defect-free, crystalline conjugated polymers have the higher charge transport mobility, maximizing effective polymer crystallinity is of paramount importance in OPV device applications. Film morphology and the ratio of crystalline to amorphous polymer strongly depend on spin casting conditions (eg. speed, temperature) and the choice of solvent used to cast films. Annealing of films just below the glass transition temperature of the polymer can “iron-out” defects in individual polymer chains which has a beneficial effect when trying to increase polymer crystallinity35. In the case of organic conjugated polymer composites, there

5.2 Polymer/single walled carbon nanotube composites for bulk heterojunction diodes

The electrical and optical properties of composites formed between carbon nanotubes and the conjugated polymer poly-(alkylthiophenes) makes them very promising candidates for use in organic photovoltaic 150

devices37. By including carboxylic acid functionalized, thin, short, multiwalled carbon nanotubes, we readily dispersed the two components in a common organic solvent suitable for spin casting the composite films13,14. Interestingly most conjugated polymers are also insoluble in common solvents. This is because a rigid polymer backbone is often the requirement to maintain significant π-electron delocalization along the chain. However an insoluble conjugated polymer such as polythiophene can be granted solubility by grafting flexible side chains to the main polymer backbone. This increases the entropy of solution and entropy of melting of the material making it soluble and tractable but without losing main-chain conjugation. The solubility requirement for side-chain decoration on the polymer can have a profound effect on polymer crystalline structure, optical and electrical properties, depending on the nature of the side-chain and its position of attachment. This is clearly demonstrated in the case of poly(3-hexylthiophene) (P3HT) by comparing the charge transport mobility of the regioregular form of P3HT (rrP3HT mobility = 0.1 cm2/Vs) with regiorandom P3HT of the same average molecular weight (rraP3HT mobility = 10-7–10-4 cm2/Vs)16. The only difference between these two forms of the same material is that the former possesses a hexyl side-chain attachment on the 3-position of the thiophene ring while the latter contains hexyl side chains randomly attached at either the 3 or 5 position of the thiophene units of the polymer. When cast into a film, the regioregular P3HT is more likely to form an ordered, head-to-tail coupled solid, with fewer amorphous regions compared to the regiorandom form38. P3HT is a semi-crystalline polymer and in the crystalline region is believed to conduct current through both inter and intra-chain transport, whereas in the amorphous region, transport is through hopping or tunneling processes only35. The significant influence of non-covalent, intermolecular interactions within the polymer upon electrical charge transport is clearly demonstrated by this comparison.

the efficiency of our cells is only 0.01%, and ranges up to 0.1% for other work7,19-21. Poor order in the microscopic arrangement, oxidation and pollution of the compound, presence of impurities or low grade of carbon nanotubes purity can all be possible causes of these low efficiencies. If changes in molecular order strongly affect P3HT film conductivity, rrP3HT structures formed at P3HT/SWNT interfaces throughout the composite are likely to be important in determining the overall efficiency of heterojunction OPVs made from this material. We have recently demonstrated that the inclusion of nanotubes in P3HT forms a composite with enhanced electrical conductivity and increased band-edge absorption. We further arrived at the surprising results that carbon nanotube inclusion in polymer templates P3HT organisation and resulted in a highly ordered material14. This is highly significant since an important requirement for efficient electron-injection from rrP3HT to carbon nanotubes in an OPV constructed from this composite material is close physical contact of the P3HT, preferably in some crystalline or ordered monolayer form at the SWNT interfaces throughout composite film. A further motivation for this investigation relates to an ultimate use for rrP3HTwrapped SWNTs in future nanotechnology applications. There certainly are tremendous opportunities for improving the performance of the solar cell based on this material described here, and we for one are greatly excited. Acknowledgements

This work is supported by Air Force Office of Research under grant no. US AirForce AOARD-064041 and the Australian Research Council (ARC) Linkage Grant LX0561885. R.G. acknowledges the QLD Government for a Smart State Award References

1.

2.

3. 4. 5. Figure 8. Organic solar cell (QUT, Brisbane)

6.

To date our results (as well as those of other groups e.g. Kymakis et al7) have not been satisfactory. While cells can be relatively easily fabricated (see Figure 8), 151

A. B. Dalton, S. Collins, E. Munoz, J. M. Razal, V. H. Ebron, J. P. Ferraris, J. N. Coleman, B. G. Kim, and R. H. Baughman, Nature (London) 423, 703 (2003). A. J. Heeger, in Proceedings of the Eighty-first Nobel Symposium., edited by W. R. Salaneck, I. Lundström and B. Rånby (Oxford University Press, 1993, Luleå, Sweden, 1991), p. 27 H. Shirakawa, E. J. Louis, A. G. MacDiarmid, C. H. Chiang, and A. J. Heeger, J. Chem. Soc., Chem. Commun, 578 (1977). A. Star, Y. Lu, K. Bradley, and G. Gruner, Nano Letters 4, 1587 (2004). H. Spanggaard and F. Krebs, C., Solar Energy Materials & Solar Cells 83, 125 (2004). M. Grell, in Nanoscale Science and Technology, edited by R. Kelsall, W., I. Hamley, W. and M. Geoghegan (John Wiley & Sons, Chichester, 2005), p. 282.

7. 8. 9.

10.

11.

12. 13. 14. 15. 16. 17. 18. 19.

20.

21.

22. 23. 24.

E. Kymakis, I. Alexandrou, and G. A. J. Amaratunga, Journal of Applied Physics 93, 1764 (2003). E. Kymakis, E. Koudoumas, I. Franghiadakis, and G. A. J. Amaratunga, Journal of Physics D: Applied Physics 39, 1058 (2006). Anthony W. Musumeci, Glaura G. Silva, JiangWen Liu, Wayde N. Martens and Eric R. Waclawik, “Structure and conductivity of multiwalled carbon nanotube/poly(3hexylthiophene) composite films”, Polymer, 48, 1667-1678 (2007) M. J. O'Connell, P. Boul, L. M. Ericson, C. Huffman, Y. Wang, E. Haroz, C. Kuper, J. Tour, K. D. Ausman, and R. E. Smalley, Chem. Phys. Lett., 342 (3-4), 265-271 (2001) B. McCarthy, J. N. Coleman, R. Czerw, A. B. Dalton, M. i. h. Panhuis, A. Maiti, A. Drury, P. Bernier, J. B. Nagy, B. Lahr, H. J. Byrne, D. L. Carroll, and B. W. J., J. Phys. Chem. B 106, 2210 (2002). D. Baskaran, J. W. Mays, and M. S. Bratcher, Chemistry of Materials 17, 3389 (2005). R. G. S. Goh, J. M. Bell, N. Motta, and E. R. Waclawik, Journal of Nanoscience and Nanotechnology 6, 3929 (2006). R. G. S. Goh, N. Motta, J. M. Bell, and E. R. Waclawik, Applied Physics Letters 88 (2006). H. Dai, A. Javey, E. Pop, D. Mann, and Y. Lu, NANO: Brief Reports and Reviews 1, 1 (2006). H. Ko and V. Tsukruk, Nano Lett 6, 1443 (2006). P. Avouris, Acc. Chem. Res 35, 1026 (2002). S. J. Kang, C. Kocabas, T. Ozel, M. Shim, N. Pimparkar, M. A. Alam, S. V. Rotkin, and J. A. Rogers, Nature Nanotechnology 2, 230 (2007). Savita P. Somani, Prakash R. Somani, M. Umeno and E. Flahaut, “Improving photovoltaic response of poly(3-hexylthiophene)/n-Si heterojunction by incorporating double walled carbon nanotubes, Applied Physics Letters, 89, 223505-1 - 223505-3, (2006). Brian J. Landi, Ryne P. Rafaelle, Stephane L. Castro and Sheila G. Bailey, Single-wall Carbon Nanotube-Polymer Solar Cells, Progress in Photovoltaics: Research and Applications, 13, 165 – 172 (2005) E. Kymakis and G. A. J. Amaratunga, Single-wall carbon nanotube/cpnjugated polymer photovoltaic devices, Applied Physics Letters, 80, 112 – 114 (2002) A. C. G. Dillon, T., K. M. Jones, J. L. Alleman, P. A. Parilla, and M. J. Heben, Adv. Mater. 11, 1354 (1999). P. S. Gill, S. R. Sauerbrunn, and B. S. Crowe, Journal of Thermal Analysis and Calorimetry 38, 255 (1992). Y. Xi, W. Martens, H. He, and R. L. Frost., Journal of Thermal Analysis and Calorimetry 81, 91 (2005).

25. C. A. Furtado, U. J. Kim, H. R. Gutierrez, L. Pan, E. C. Dickey, and P. C. Eklund, J. Am. Chem. Soc. 126, 6095 (2004). 26. H. S. Nalwa, Handbook of Organic Conductive Molecules and Polymers (John Wiley and Sons, New York, 1997). 27. S.-B. Lei, L.-J. Wan, C. Wang, and C.-L. Bai, Adv. Mater. 16, 828 (2004). 28. J. N. Coleman and M. S. Ferreira, Applied Physics Letters 84, 798 (2004). 29. B. Grevin, P. Rannou, R. Payerne, A. Pron, and J. P. Travers, J. Chem. Phys. 118, 7097 (2003). 30. E. Mena-Osteritz, A. Meyer, B. M. W. LangeveldVoss, R. A. J. Janssen, E. W. Meijer, and P. Bauerle, Angewandte Chemie-International Edition 39, 2680 (2000). 31. B. Gregg, J. Phys. Chem. B 107, 4688 (2003). 32. C. Brabec, J., Solar Energy Materials & Solar Cells 83, 273 (2004). 33. K. Kim, J. Liu, M. A. G. Namboothiry, and D. L. Carroll, Appl. Phys. Lett. 90, 163511 (2007). 34. E. J. Samuelsen and J. Mårdalen, in Handbook of Organic Conductive Molecules and Polymers, edited by H. S. Nalwa, 1997), Vol. 3, p. 87. 35. B. J. Schwartz, Ann. Rev. Phys. Chem. 54, 141 (2003). 36. J. Ruhe, N. F. Colaneri, D. D. C. Bradley, R. Friend, H., and G. Wegner, J. Phys.: Condens. Matter 2, 5465 (1990). 37. E. Kymakis, I. Alexandou, and G. Amaratunga, Synthetic Metals 127, 59 (2002). 38. H. Sirringhaus, P. J. Brown, R. H. Friend, M. M. Nielsen, K. Bechgaard, B. M. W. Langeveld-Voss, A. J. H. Spiering, R. A. J. Janssen, E. W. Meijer, P. Herwig, and D. M. De Leeuw, Nature (London) 401, 685 (1999).

152