A New Supramolecular Route for Using Rod-Coil Block Copolymers in Photovoltaic Applications

A New Supramolecular Route for Using Rod-Coil Block Copolymers in Photovoltaic...

(Parte 1 de 2)

A New Supramolecular Route for Using Rod-Coil Block Copolymers in Photovoltaic Applications

By Nicolas Sary, Fanny Richard, Cyril Brochon, Nicolas Leclerc, Patrick Leveque, Jean-Nicolas Audinot, Solenn Berson, Thomas Heiser,* Georges Hadziioannou,* and Raffaele Mezzenga*

The growing interest for renewable energy technologies, such as photovoltaic (PV) devices, combined with the need for low-cost processing, have contributed to the quick expansion of organic PVs.[1] Since the pioneering work of Tang[2] on electron-donor (D)/electron-acceptor (A) double-layer devices, considerable efforts have focused on the development of bulk D/A heterojunctions based on photoactive compounds of electron-donating conjugated polymers and fullerene derivatives.[3–6] In these devices the organic components form, throughout the entire active layer, nanometer-sized D and A domains at whose interfaces photogenerated excitons can dissociate into free charge carriers, which in turn are driven to the collecting electrodes by the built-in electric field of the device.[7,8] Applying this methodology to polythiophene/fullerene blends led to PVdevices with power conversion efficiencies (PCEs) around 5%.[9,10]

Despite this success, polymer/fullerene blends suffer from two major drawbacks: a poorly controlled D/A domain size distribution and inherent morphological instability. The D and A domains generally originate from spinodal decomposition occurring during the film formation from a spin-coated solution and are therefore strongly dependent on the processing conditions and difficult to control.[1] Moreover, macrophase separation of both blend components may occur within the active layer upon extended device operation and considerably modify the as-deposited thin film morphology.[3] The resulting domain size can ultimately become much larger than the exciton diffusion length (about 10nm in semiconducting polymers[12,13]) and diminish the device performances.

The use of rod–coil block copolymers as photoactive material in bulk heterojunction devices is a possible way to overcome these drawbacks. Rod–coil block copolymers are indeed well known to self-assemble through microphase separation into highly ordered nanostructures that are thermodynamically stable and exhibit spatial periodicities on the 1–10nm length scale.[14–19] Block copolymers composed of an electron-donating and an electronaccepting block are therefore particularly interesting for PV applications and are presently studied worldwide by several research groups.[20–37] Particularly, rod–coil block copolymers using poly[(2,5-di(2-ethyl)hexyloxy)-1,4-phenylenevinylene] (DEH-PPV) as electron donor and various coil blocks (such as polystyrene or polybutylacrylate) with covalently linked fullerene moieties as electron acceptor have been investigated intensively.[23–28] Although these studies have given considerable insight into the physics of copolymer self-assembly, their efficient utilization as the active layer in PV devices has not yet been fully demonstrated.

In the present work, we report on the thin film nanostructure of blends of regio-regular poly(3-hexylthiophene)-block-poly(4- vinylpyridine) (P3HT-b-P4VP) rod–coil block copolymers with [6,6]-phenyl-C61-butyric acid methyl ester (PCBM) and on the optoelectronic properties of preliminary PVdevices made thereof.

We anticipate that a conjugated polymer with strong p–p stacking interactions, such as regio-regular P3HT (r-P3HT), used as the rod block should stabilize the copolymer nanostructure in the presence of fullerene derivatives and allow good hole transport. Furthermore, we show that the utilization of a P4VP coil block demonstrates a new way to form electronacceptor domains within a block copolymer self-assembled nanostructure. Indeed, polyvinylpyridines are known to experience weak supramolecular interactions with electron-deficient chemical species.[38–40] These interactions would make free C60 molecules preferentially accumulate within the coil domains and, w.MaterialsViews.com w.advmat.de

Institut d’Electronique du Solide et des Systemes Centre National de la Recherche Scientifique, Universite de Strasbourg 23, Rue du Loess, 67037 Strasbourg (France) E-mail: thomas.heiser@iness.c-strasbourg.fr

Prof. G. Hadziioannou,[+] Dr. F. Richard, Dr. C. Brochon[+] Dr. N. Leclerc Laboratoire d’Ingenierie des Polymeres pour les Hautes Technologies Universite de Strasbourg, Ecole Europeenne de Chimie Polymeres et Materiaux 25, Rue Becquerel, 67087 Strasbourg (France) E-mail: hadzii@ecpm.u-strasbg.fr

Prof. R. Mezzenga, Dr. N. Sary Department of Physics and FRIMAT Center for Nanomaterials, University of Fribourg Ch. Musee 3, CH-1700 Fribourg (Switzerland) E-mail: raffaele.mezzenga@unifr.ch

Prof. R. Mezzenga Nestle Research Center, Vers-Chez-Les-Blanc 1000 Lausanne 26 (Switzerland)

Dr. J.-N. Audinot Science and Analysis of Materials Department Public Research Centre Gabriel Lippmann 41 rue du Brill, L-42 Belvaux (Luxembourg)

Dr. S. Berson Laboratoire des Composants Solaires, Institut de l’Energie Solaire Commissariat a l’energie atomique BP 332 50 Avenue Du Lac Leman, 73377 Le Bourget Du Lac (France)

[+] Present address: Laboratoire de Chimie des Polymeres Organiques

(LCPO) CNRS/UNIV. Bordeaux1/ENSCBP-16 av. Pey-Berland-33607 PESSACCedex(France);E-mail(fromJanuary2010):hadzii@enscbp.fr

DOI: 10.1002/adma.200902645

w.advmat.de w.MaterialsViews.com when combined with an electron-donating rod block, lead to the endeavored D/A interpenetrated networks. Also, in the case of block copolymer/fullerene blends, the copolymer microphase separation has been found to be less affected by the C60 crystallization than for fullerene-grafted block copolymers.[34,39]

In particular, the pristine DEH-PPV-b-P4VP nanostructure was shown to be preserved when blended with 10% C60 and to be thermally stable up to at least 16h at 1808C.[39] Finally the PCBM fullerene derivative was chosen for its high solubility in common solvents.

In this paper, special emphasis is put on the copolymer nanostructure, the morphology thermal stability, and the device properties, for different PCBM contents. We find that the P3HT ordering as well as the copolymer nanostructure can be maintained even at relatively large fullerene contents (36 vol%) and that the thermal stability is dramatically improved in comparison to P3HT:PCBM blends. Finally, we show that a high photon-to-current conversion efficiency (above 40%) and an overall PCE of 1.2% can be reached even with non-optimized PV devices, which positions the present solar cells among the best-performing PVdevices having block copolymersas the major constituent in the active layers.

Figure 1a sketches the blend of the P3HT-P4VP block copolymer and PCBM studied in the present work. P3HT-P4VP has been obtained by anionic polymerization of 4-vinylpyridine and quenching with an aldehyde endfunctionalized P3HT. This route has been adapted from the synthesis of a PPV-based block copolymer previously described[17,18] and will be discussed in detail in a separate manuscript. The synthesized diblock copolymer (Fig. 1a) has a total molecular weight of 1.6kg mol 1 and a P3HT52-P4VP28 architecture. Three series of P3HT-P4VP:PCBM blends, based on either 8, 17, and 36% volume fractions of PCBM were investigated. These blends and related PV devices are hereafter referred to as C8, C17, and C36. Standard P3HT:PCBM (1:1 weight ratio) blends were used as reference material. The details about film formation and device elaboration procedures are described in the Experimental Section.

The thin film nanostructure was investigated by transmission electron microscopy (TEM) and UV–vis absorption spectroscopy. PVdevices using the C8, C17, and C36 blends as photoactive layer were elaborated according to both, standard (Fig. 1b) and inverted (Fig. 1c) configurations, for reasons which will become evident in what follows. The devices current–voltage characteristics were measured under darkness and under air mass 1.5 (AM1.5) illumination. The incident photon-to-current conversion efficiency (or IPCE) of the device was measured with a standard experimental set-up.

According to Ikkala and co-workers,[38] each PCBM molecule can form noncovalent bonds with up to six 4-vinylpyridine (4VP) monomer units. Considering the number of PCBM molecules per 4VP monomer unit actually present in the active layers, this is only possible in the case of C8, whereas at higher PCBM content only partial binding to 4VP units can be achieved.

Figure 2 compares the thin film morphology of the 1:1

P3HT:PCBM blend to that of the C36 copolymer blend, after annealing at 1508C for either 30min (Fig. 2a,c) or 24h (Fig. 2b,d). The morphologies of the corresponding pristine P3HT-P4VP block copolymer are shown in the insets. After 30min of annealing, both the reference and block copolymer active layers exhibit comparable nanostructures, with a high level of mixing of the blend components. Major differences arise however after longer annealing times. Micrometer-sized dark domains, which most likely correspond to PCBM crystallites, are present in the PCBM:P3HT blend and point out significant macrophase separation. On the other hand, the C36 nanostructure shows an increased structural order and no formation of microdomains, maintaining a morphology similar to that of the pristine block copolymer (see Fig. 2 insets). These results therefore suggest that i) the high PCBM loading does not perturb the copolymer self-assembly and i) the block copolymer:PCBM system provides a significantly improved structural stability. Most importantly, these findings show that supramolecular bonding between the fullerene and P4VP coil block avoids the formation of macroscopic fullerene crystals without hampering the copolymer self-assembly. This behavior contrasts with that of previously reported fullerene-grafted block copolymer self-assembly.[34] Furthermore, from the pronounced thermal stability of the

Figure 1. a) P3HT-P4VP block copolymer/PCBM compound used for the polymer bulkheterojunction activelayer: b) standard andc)invertedPVdevice structure used in this study. (PEDOT:PSS¼poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate), ITO¼indium tin oxide, HTL¼hole transport layer).

2 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2009, 21, 1–6 Final page numbers not assigned w.MaterialsViews.com w.advmat.de

copolymer morphology at 1508C, it may be anticipated that devices made thereof will experience a significantly enhanced lifetime, in comparison to P3HT:PCBM reference devices, when used in typical solar cell operating conditions (temperature 608C). Because the substrate–blend interface may have a nonnegligible influence on the thin film morphology in a PV device, we also investigated the TEM morphology of the copolymer blend in a real device configuration. Figure 3 shows the top view and cross-section of a C17 solar cell active layer in a standard device structure (Fig. 1b) annealed for 15min at 1408C. The PEDOT:PSS layer was dissolved in water, and the active layer was collected on a copper grid. The cross-section was performed below the aluminum top electrode. A nanostructure similar to the morphology given in Figure 2d can be observed in both the in-plane (Fig. 3a) and out-of-plane (Fig. 3b) projections. This observation highlights the existence of an interpenetrated network of donor and acceptor nanodomains throughout the film thickness. However, close inspection of the active layer near the bottom interface, initially in contact with the PEDOT:PSS layer, reveals the existence of a rather uniform bright layer. The latter suggests the occurrence of preferential wetting of one of the copolymer blocks at the PEDOT:PSS interface during the film formation. In fact, complementary investigations using secondary ion mass spectroscopy (SIMS) provide evidence supporting the accumulation of P4VP at the PEDOT:PSS layer; these are discussed in the Supporting Information. This behavior corroborates the already reported tendency of polyvinylpyridine to preferentially wet oxide or charged surfaces.[41] The impact of this layer on the device performances will be discussed below.

The UV-light absorption spectra of C17 and C36 thin films, before and after annealing at 1508C for 24h, are given in Figure 4. Three absorption peaks located at 519, 5, and 603nm can be seen and are similar to the spectral features frequently observed in r-P3HT thin films.[42,43] The P3HTabsorption peaks are known to be related to the vibronic splitting of the p–p* transition, and their amplitude ratio has been correlated before to the degree of polymer ordering. Their presence in the absorption spectra reflects a significant degree of structural ordering, which increases upon annealing, even for relatively high fullerene volume fractions (up to 36%). This property is essential for the charge carrier mobility within block-copolymer films.

Both, TEM and spectroscopic results indicate the ability of

P3HT-P4VP block copolymers to self-assemble into thermally stable nanostructured thin films consisting of ordered P3HT domains and PCBM-enriched P4VP domains. The preferential positioning of PCBM molecules within the coil domains will be further assessed through the PV device performances, since the latter are strongly dependent on the existence of percolating transport paths for both electrons and holes.

The IPCE for solar cells obtained with the standard device structure (Fig. 1b) were found to exceed 40% (see Supporting Information). Such high values indicate efficient exciton dissociation into free carriers and hence corroborate the conclusions from our structural investigations. Nonetheless, the PV performances of these devices turned out to be extremely

Figure 2. TEM images of P3HT:PCBM (1:1) and P3HT-P4VP:PCBM (C36) thin films after iodine staining and various annealing times at 1508C: a) the P3HT:PCBM reference film and b) C36 film after 30min annealing; b) the reference and c) C36 films after 24h annealing. Macrophase separation is observed only in the reference film (b), whereas the block copolymer nanostructure improves upon long-term annealing (d). Inset of c,d: the nanostructures of the pure block copolymer P3HT-P4VP after 30min (c, inset) and 24h (d, inset) annealing over a 200nm 200nm surface area.

Figure 3. TEM images (iodine-stained) of the active layer of a C17 standard device. a) Top view of an Al-free region of the film, and b) device crosssection. The active layer exhibits the same percolating structure in lateral and perpendicular projections.

(Parte 1 de 2)