Self-Organization of Nanocrystals in Polymer Brushes. Application in

Heterojunction Photovoltaic Diodes. Henry J. Snaith,† Gregory L. Whiting,‡ Baoquan Sun,† Neil C. Greenham,†. Wilhelm T. S. Huck,*,‡ and Rich...
0 downloads 0 Views 337KB Size
NANO LETTERS

Self-Organization of Nanocrystals in Polymer Brushes. Application in Heterojunction Photovoltaic Diodes

2005 Vol. 5, No. 9 1653-1657

Henry J. Snaith,† Gregory L. Whiting,‡ Baoquan Sun,† Neil C. Greenham,† Wilhelm T. S. Huck,*,‡ and Richard H. Friend*,† CaVendish Laboratory, Department of Physics, UniVersity of Cambridge, Cambridge CB3 0HE, U.K., and MelVille Laboratory for Polymer Synthesis, Department of Chemistry, UniVersity of Cambridge, Cambridge CB2 1EW, U.K. Received March 15, 2005; Revised Manuscript Received June 30, 2005

ABSTRACT We present a new approach to achieving order in molecular semiconductors via alignment of polymer chains using surface-initiated polymerization. Polyacrylate brushes grown from transparent conducting electrodes, with triarylamine side groups as hole-transporting components, show characteristics of high mobilities for hole transport. Solution processing a second component with favorable enthalpic interactions can form a composite with mesoscale order and be exploited for heterojunction diodes. We find substantial uptake of CdSe nanocrystals (with diameter in the range 2.5−2.8 nm), and such composites show photovoltaic quantum efficiencies of up to 50%.

Semiconducting polymers hold great promise for the fabrication of organic optoelectronic devices. However, the favorable properties demonstrated in (supra)molecular systems, where all building blocks are positioned in precisely defined geometries, cannot be translated to polymer length scales due to an inherent loss of order. A new paradigm of ordering polymeric materials at the mesoscale is required to bridge the gap between the intrinsic properties of organic materials and device performance in order to fulfill the potential of polymer electronics. Heterostructures formed between electron and hole-accepting molecular semiconductors are generally required for active photovoltaic device structures.1 Solution processing of polymer-polymer, polymer-molecular, and polymer-nanocrystal blends has now been widely developed and allows some control over morphology, both at molecular and at mesoscale order.2-7 We have developed the growth, from transparent conducting electrodes, of films of polyacrylate brushes with triarylamine side groups as holetransporting components.8 Solution processing a second component with favorable enthalpic interactions does not give bilayer structures but instead forms a composite with mesoscale order. We find substantial uptake of CdSe nanocrystals (with diameter in the range 2.5-2.8 nm), with a weight composition of up to 2:1 nanocrystal/polymer. These structures give distinct vertical pathways in each component, show lateral organization on a length scale of tens of * Corresponding authors: Richard H. Friend, e-mail [email protected]; Wilhelm T. S. Huck, e-mail [email protected]. † Cavendish Laboratory, Department of Physics. ‡ Melville Laboratory for Polymer Synthesis, Department of Chemistry. 10.1021/nl0505039 CCC: $30.25 Published on Web 07/28/2005

© 2005 American Chemical Society

nanometers, and are therefore particularly suited for photovoltaic devices, since they provide the correct length scale for migration of photogenerated excitons to heterojunctions and very effective conduction pathways for electrons and holes to charge collection electrodes. This combination of surface-tethered polymers and nanocrystals offers a general strategy for scalable processing of nanoscale-ordered composites. For a surface-initiated polymerization, the polymer chains are grown directly from surface-tethered initiator sites. If the density of attachment sites is sufficiently high, then the polymer chains are forced into a conformation where they stretch away normal to the surface, the “polymer brush” regime.9 Dense polymer brushes have been used to control surface properties such as adhesion, corrosion resistance, and wettability,10 and recently as dielectric films in field-effect transistors.11 At lower grafting densities, the polymers collapse into a noninteracting “mushroom” conformation. We report below the first use of functional semiconducting polymer brushes in composite diodes. The vertically aligned nature of polymer brushes makes them ideal candidates to incorporate as an electroactive component within organic semiconductor devices, potentially improving transport properties and providing clear pathways for charge transport in the direction normal to the substrate. The brushes in this work are composed of poly(triphenylamine acrylate) (PTPAA), which has been shown to exhibit respectable hole mobilities in amorphous films.12,13 The chemical structure of PTPAA is shown as an insert to Figure

Figure 1. AFM image of a 35 nm thick PTPAA brush film grown on an ITO-coated glass substrate. Inset: Chemical structure of poly(triphenylamine acrylate) (PTPAA).

1. It was chosen since the acrylate functionality allows surface polymerization via atom transfer radical polymerization (ATRP). Here, the triphenylamine acrylate monomer was either polymerized in the bulk by a free radical polymerization or directly from indium-tin oxide (ITO) modified with an initiator-terminated self-assembled monolayer. An in-depth description of the polymerization technique and characterization of the polymer brush films will be published elsewhere.8 The thicknesses of dry polymer brush films, as measured by ellipsometry on silicon substrates (used only for characterization purposes), varied between 20 and 80 nm depending on polymerization conditions. CdSe nanocrystals were synthesized by a slight modification of the method of Peng and Peng.14 The ligand was replaced with pyridine by refluxing the nanocrystals in pyridine overnight. They were then precipitated with hexane and solvated in a 13:87 by volume pyridine/chloroform mixed solution in an ultrasonic bath for 2 h. Their size was estimated by the absorption spectra. The PTPAA brushcoated substrates were soaked in a solution of nanocrystals (25 g/L 13:87 pyridine/chloroform) for 30 min in a chloroform-saturated atmosphere. After soaking, the substrates were carefully placed on a spin coater and the excess solution was spun off. PTPAA/CdSe nanocrystal blends 1:8 by weight were fabricated by mixing the respective quantities of the mixed solvent nanocrystal solution and a PTPAA solution (25 g/L in chloroform) and films spin-coated on precleaned ITO-coated glass slides. The prepared substrates (coated polymer brush and blend films) were then annealed at 150 °C in nitrogen for 30 min to ensure all the pyridine was removed, allowing the triphenylamine to replace the pyridine as the ligand to the CdSe. The remainder of the device fabrication, characterization, atomic force microscopy, UVvis spectroscopy, and photoluminescence measurements were all performed as previously by the authors.15 An atomic force microscopy (AFM) image of a 35 nm thick brush film grown on an ITO surface is shown in Figure 1654

1. This film appears smooth and continuous over the whole image with no occurrence of pinholes or defects, illustrating that our method is suitable for fabricating very thin continuous films. Electronic characterization of the pristine brush films will be published elsewhere.8 However, we note here that the current density through a PTPAA brush film is found to be approximately 3 orders of magnitude higher than that through a spin-coated film of the same thickness. This suggests that this polymer brush film should be highly efficient at collecting or injecting charge in a diode structure. The ability to infiltrate the polymer brush film with a second component is of central interest. Polymer brush films can swell in a good solvent to many times their original thickness.16 This may generate sufficient free volume to accommodate a significant volume fraction of electrontransporting material. However, since stretching of the polymer brush is entropically unfavorable, a strong favorable interaction between the brush and the coating component is required to aid intercalation into the brush film. Bhat et al. have shown that gold nanoparticles will penetrate into polyacrylamide due to favorable interactions between the nanoparticles and the brush.17 Among the many materials we tried, CdSe nanoparticles proved particularly effective. As we present later, the uptake of the nanoparticles is extremely size sensitive. The size of the nanoparticle can be finely tuned by the reaction time of the synthesis, thus allowing size optimization for interpenetration into the brush system.18 Nanorods (typically 4 nm by 50 nm) and nanocrystals over 3 nm in diameter form a bilayer structure, sitting on top of the brushes. However by using nanocrystals less than 2.8 nm in diameter, infiltration into the brush network is achieved. For the following work we use CdSe nanocrystals of 2.6 nm diameter, with pyridine as the capping ligand. The pyridine ligand was chosen as it enables the nanocrystals to be soluble in organic solvents, but it is relatively volatile and evaporates during the film drying process, allowing the triarylamine to act as a ligand replacing the pyridine through mass action (amines act as strong ligands to inorganic nanoparticles).19 UV-vis absorption spectra of a 45 nm thick PTPAA brush film infiltrated with 2.6 nm CdSe nanocrystals is shown in Figure 2a, along with a pristine brush and nanocrystal film. The absorption of the CdSe-infiltrated brush film at 530 nm (where only the nanocrystals absorb) is identical to that of a 25 nm thick nanocrystal film, suggesting that the polymer brushes adsorb the equivalent of a 25 nm thick nanocrystal film. The total thickness of the nanocrystal infiltrated brush film is approximately 70 nm. This demonstrates that the polymer brushes swell, leading to a final film with over half the original polymer volume replaced by nanocrystals, and an ensuing composition of approximately 1:2 PTPAA/CdSe by weight (assuming the density of PTPAA to be 1 g cm-3 and CdSe nanocrystals to be approximately 4 g cm-3).20,21 Figure 2b shows an AFM phase image of a 45 nm thick PTPAA brush film infiltrated with CdSe nanocrystals (equivalent to 25 nm thick). The phase image in tapping mode AFM gives information concerning the loss in energy of the tip upon contact with the surface. This is highly Nano Lett., Vol. 5, No. 9, 2005

Figure 2. (a) Absorption spectra as measured by UV-vis spectroscopy of a 25 nm thick film of CdSe nanocrystals spin-coated on an ITO slide (dashed line), 45 nm thick PTPAA brushes grown from an ITO slide (dotted line), and the PTPAA brush film infiltrated with CdSe nanocrystals (solid line). (b) AFM phase image of a 45 nm PTPAA brush film grown from a glass substrate infiltrated with 2.6 nm diameter CdSe nanocrystals. (c) Current-voltage characteristics for diodes fabricated with ITO anodes and Al cathodes (bipolar diodes) with active layers of PTPAA brushes (45 nm) infiltrated with 2.6 nm diameter CdSe nanocrystals (solid line), PTPAA/CdSe nanocrystals blend 1:8 by weight (70 nm thick) (dashed line) and PTPAA brushes (45 nm) coated with 3.0 nm diameter CdSe nanocrystals (dot-dashed line). Also shown hole-only diodes consisting of ITO/PTPAA brushes (35 nm)/PEDOT:PSS/Au (dotted line). (d) Cartoon of inferred structure for CdSe nanocrystal infiltrated polymer brush photovoltaic device. From bottom to top: ITO-coated glass slide modified by surface attachment of a bromine end-caped trichlorosilane self-assembled-monolayer (SAM) (blue squares), polymer brushes grown from the SAM (red lines), CdSe nanocrystals infiltrated into the brush network exhibiting some degree of phase separation in the plane of the film (small black circles), and caped with an aluminum cathode.

influenced by the elasticity of materials, with hard materials exhibiting a low phase and soft materials exhibiting a high phase. From the image in Figure 2b it is apparent that there are two materials present, or at least two phases, with the darker regions corresponding to the hard nanocrystals, or nanocrystal-rich phases, and the lighter regions corresponding to the soft polymer, or polymer-brush-rich phases. We note that the length scale of phase separation is on the order of 10 nm, matching the short exciton diffusion length in polymeric semiconductors.22 Furthermore the presence, in the AFM image, of polymer brushes at the surface implies that the nanocrystals have penetrated into the film. Polymer brush/nanocrystal composite diodes were fabricated by capping the nanocrystal-coated PTPAA brush films with aluminum as the cathode, forming bipolar diodes. Diodes fabricated from blends of CdSe nanocrystals in the Nano Lett., Vol. 5, No. 9, 2005

PTPAA polymer show much higher current densities than the hole-only diodes fabricated from the pure polymer or brush film, Figure 2c. We consider that this is due to electrontransport through the nanocrystal phase in addition to holetransport through the polymer phase.23 In contrast, when 3 nm diameter CdSe nanocrystals are used, the current density drops when compared to that of the pristine brush film. This is characteristic of the bilayer structure formed here where the nanocrystals sit on top of the brush film, which does not give direct paths for holes or electrons right through the device. However, when the brush film is coated with 2.6 nm diameter nanocrystals, the current density is approximately 3 orders of magnitude larger than that obtained through the pure PTPAA brush hole-only diode. As the charge mobility through the nanocrystal phase is considerably higher than that through the polymer phase, this provides 1655

direct evidence that these small nanocrystals interpenetrate the polymer brush film from the top, right down to the ITO anode. The current density through the brush-nanocrystal composite film is also much higher than that seen through the standard spin-coated blend film. This requires very good connectivity within the adsorbed nanocrystal phase as would result from formation of near-pristine columns of nanocrystals within the brush network. These results demonstrate that the polymer brushes swell when infiltrated with small nanocrystals and that the nanocrystals completely interpenetrate the film from top to bottom resulting in efficient electron transport pathways along nanocrystal “channels”. From the AFM surface image (Figure 2b) there appears to be phase separation within the plane of the film suggesting that these nanocrystal “channels” are in the order of 10 nm in diameter. A schematic representation of the described structure is shown in Figure 2d. We present photovoltaic action spectra for CdSe nanocrystal polymer brush diodes in Figure 3a. For comparison we studied blend devices over a range of blend compositions (2:1 to 1:10 PTPAA/CdSE nanocrystals by weight) and present the data for the best performing blend device comprising 1:8 PTPAA/CdSe. The polymer brush device shows significantly enhanced performance, with the external quantum efficiency at 400 nm approximately twice that of the blend device. The maximum external quantum efficiency of the brush device occurs where the polymer absorbs and is approximately 20%. The internal quantum efficiency (as the fraction of collected electrons to absorbed photons) was calculated for the brush and blend devices by dividing the external quantum efficiency by the fraction of incident light absorbed in the active layer. This absorption was calculated taking full account of optical interference effects. Optical constants for the brush and spin-coated layers were determined by fitting a combination of reflection ellipsometry and transmission data. We find the brush device is most efficient at converting absorbed photons to electrons where the nanocrystals absorb. Most significantly the internal quantum efficiency is as high as 50%. This is considerably more than that of the blend device (20%); therefore we have a structure in the brush device which is well-matched for charge collection. Due to the large dark current through these brush composite devices, it is only possible to generate a small open circuit voltage (50-100 mV) when using ITO as the anode. However, by the use of an electron blocking layer a substantial open-circuit voltage can be generated. It is possible to grow PTPAA brushes from modified PEDOT/ PSS substrates.8 This technique has not as yet been perfected; however initial results are promising. Figure 3b shows the current voltage characteristics of a device comprising a 20 nm thick PTPAA brush film grown from a modified PEDOT/ PSS substrate infiltrated with 2.8 nm diameter CdSe nanocrystals and capped with an aluminum cathode. We obtain an open-circuit voltage of approximately 850 mV and a peak external quantum efficiency of 8%. This demonstrates that the brush composite films are highly suitable for photovoltaic operation. 1656

Figure 3. (a) Photovoltaic action spectra for a CdSe nanocrystal (2.6 nm diameter) infiltrated PTPAA brush (45 nm thick) device (open diamonds) and for PTPAA/CdSe nanocrystal blend device 1:8 by weight (solid circles), with external quantum efficiency (dotted lines) and internal quantum efficiency (solid lines). (b) Current-voltage characteristics for a device comprising a 20 nm thick PTPAA brush film grown from a modified PEDOT:PSS coated ITO slide, infiltrated with 2.8 nm diameter CdSe nanocrystals and capped with an aluminum cathode, in the dark (dashed line) and under 400 nm incident illumination (solid line).

It is apparent that the structure obtained when the polymer brush film is coated with CdSe nanocrystals of the correct size is well-suited to that necessary for photovoltaic operation. As the polymer chains are tethered to the anode and have a high degree of alignment perpendicular to the substrate, any hole generated within the device has a direct path to the collection electrode. Furthermore the transport in the nanocrystal phase is dramatically improved as compared to that of a polymer/nanocrystal blend film, facilitating highly efficient electron collection. In conclusion, we have achieved a new level of control over the active components of organic optoelectronic devices using polymer brushes and self-organization of hybrid materials. We consider this to be a general and widely applicable strategy for controlled assembly of nanoparticles to form functional structures. Nano Lett., Vol. 5, No. 9, 2005

Acknowledgment. This work was funded by the Engineering and Physical Sciences Research Council. References (1) Tang, C. W. Appl. Phys. Lett. 1986, 48 (2), 183-185. (2) Arias, A. C.; Corcoran, N.; Banach, M.; Friend, R. H.; MacKenzie, J. D.; Huck, W. T. S. Appl. Phys. Lett. 2002, 80 (10), 1695-1697. (3) Brabec, C. J.; Sariciftci, N. S.; Hummelen, J. C. AdV. Funct. Mater. 2001, 11 (1), 15-26. (4) Halls, J. J. M.; Arias, A. C.; MacKenzie, J. D.; Wu, W.; Inbasekaran, M.; Woo, E. P.; Friend, R. H. AdV. Mater. 2000, 12 (7), 498-502. (5) Halls, J. J. M.; Walsh, C. A.; Greenham, N. C.; Marseglia, E. A.; Friend, R. H.; Moratti, S. C.; Holmes, A. B. Nature 1995, 376 (6540), 498-500. (6) Sun, B.; Marx, E.; Greenham, N. C. Nano Lett. 2003, 3 (7), 961963. (7) Yu, G.; Gao, J.; Hummelen, J. C.; Wudl, F.; Heeger, A. J. Science 1995, 270 (5243), 1789-1791. (8) Whiting, G. L.; Snaith, H. J.; Khodabakhsh, S.; Andreasen, J. W.; Nielsen, M. N.; Friend, R. H.; Huck, W. T. S. To be submitted. (9) Milner, S. T. Science 1991, 251 (4996), 905-914. (10) Jones, D. M.; Smith, J. R.; Huck, W. T. S.; Alexander, C. AdV. Mater. 2002, 14 (16), 1130-1134. (11) Rutenberg, I. M.; Scherman, O. A.; Grubbs, R. H.; Jiang, W. R.; Garfunkel, E.; Bao, Z. J. Am. Chem. Soc. 2004, 126 (13), 40624063.

Nano Lett., Vol. 5, No. 9, 2005

(12) Stolka, M.; Pai, D. M.; Renfer, D. S.; Yanus, J. F. J. Polym. Sci., Part A: Polym. Chem. 1983, 21 (4), 969-983. (13) Tamada, M.; Koshikawa, H.; Hosoi, F.; Suwa, T.; Usui, H.; Kosaka, A.; Sato, H. Polymer 1999, 40 (11), 3061-3067. (14) Peng, Z. A.; Peng, X. G. J. Am. Chem. Soc. 2002, 124 (13), 33433353. (15) Snaith, H. J.; Arias, A. C.; Morteani, A. C.; Silva, C.; Friend, R. H. Nano Lett. 2002, 2 (12), 1353-1357. (16) Biesalski, M.; Ruhe, J. Macromolecules 2002, 35 (2), 499-507. (17) Bhat, R. R.; Genzer, J.; Chaney, B. N.; Sugg, H. W.; LiebmannVinson, A. Nanotechnology 2003, 14, 1145-1152. (18) Shull, K. R. J. Chem. Phys. 1991, 94 (8), 5723-5738. (19) Comparelli, R.; Zezza, F.; Striccoli, M.; Curri, M. L.; Tommasi, R.; Agostiano, A. Mater. Sci. Eng., C 2003, 23 (6-8), 1083-1086. (20) Weast, R. C. Handbook of Chemistry and Physics, 66th ed.; Chemical Rubber: Boca Raton, FL, 1995; Vol. E-86. (21) For the density of the nanocrystals, we take the bulk density of CdSe to be 5.81 and assume them to be close packed spheres with a packing fraction of 0.74. (22) Stevens, M. A.; Silva, C.; Russell, D. M.; Friend, R. H. Phys. ReV. B 2001, 6316 (16), art. no.-165213. (23) Ginger, D. S.; Greenham, N. C. J. Appl. Phys. 2000, 87 (3), 1361-1368.

NL0505039

1657