Electrical characterization of organic solar cell materials based on scanning force microscopy

Electrical characterization of organic solar cell materials based on scanning force microscopy

European Polymer Journal 49 (2013) 1907–1915 Contents lists available at SciVerse ScienceDirect European Polymer Journal journal homepage: www.elsev...

1MB Sizes 7 Downloads 40 Views

European Polymer Journal 49 (2013) 1907–1915

Contents lists available at SciVerse ScienceDirect

European Polymer Journal journal homepage: www.elsevier.com/locate/europolj

Feature Article

Electrical characterization of organic solar cell materials based on scanning force microscopy Rüdiger Berger a,⇑, Anna L. Domanski a, Stefan A.L. Weber a,b a b

Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Dublin 4, Ireland

a r t i c l e

i n f o

Article history: Received 21 January 2013 Accepted 4 March 2013 Available online 6 April 2013 Keywords: Kelvin probe force microscopy Conductive scanning force microscopy Solar cell Photo-current AFM Time resolved EFM Degradation

a b s t r a c t The application of electrical modes in scanning probe microscopy helps to understand the electrical function of materials that are structured on the nanometer scale. Scanning force microscopes are routinely used for the investigation of surface topography. Here we accentuate the use of electrical modes that are unique for the correlation of structural and electric information on a nanometer scale. This is particularly important for analyzing organic solar cell materials. A special focus is given to experiments aiming at the investigation of light-induced processes which requires the integration of an additional light source into the scanning force microscope setup. Furthermore, we address future challenges for scanning force microscopy investigation of electrical properties of soft matter materials. Ó 2013 Elsevier Ltd. All rights reserved.

Contents 1. 2.

3. 4. 5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrical SFM methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Conductive scanning force microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Kelvin probe force microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Photo-induced effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SFM on solar cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . New SFM based methods for organic solar cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and future directions for the electrical characterization of organic solar cell materials with SFM . . . . . . . . . . . . . 5.1. Material issues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Sample preparation issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Instrumental issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction A constant and ubiquitous supply of energy has become essential for our modern lifestyle. We need energy to heat ⇑ Corresponding author. Tel.: +49 6131 379114. E-mail address: [email protected] (R. Berger). 0014-3057/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.eurpolymj.2013.03.005

1907 1909 1909 1909 1910 1910 1912 1912 1912 1913 1913 1914 1914

our homes, to run our cars or to keep our electronic devices running. Decreasing supply of fossil energy on one side and more and more sophisticated off-grid applications such as mobile telecommunication on the other side have raised the demand for new energy harvesting strategies. A energy source would be sun-light and thus solar cells become

1908

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915

more and more interesting for energy supply. In solar cell materials, the absorption of a photon leads to the spatial separation of a negative charge (electron) and a positive charge. In order to collect the charges, the internal structure of the cell has to guide them to their respective electrodes. Although silicon solar cells are well established, the main drawback of inorganic solar cells is the energy and cost-intense manufacturing process and their weight. Thus, in the last two decades more and more research has been devoted to alternative low-cost materials for solar cells. Organic semiconducting polymers offer a light, cheap and flexible alternative, as they can be processed with well-established coating techniques such as bladecoating or roll-to-roll printing. These so-called plastic solar cells mostly consist of highly conjugated hydrocarbons. For an efficient charge separation, a combination of an electron donor and electron acceptor material is used in the active layer. Widely used examples for donor materials are polythiophenes or polytriphenylamines. The most common organic acceptors are fullerene derivatives with functional side groups that promote solubility in organic solvents, e.g. phenyl-C61-butyric acid methyl ester (PCBM). A detailed overview on the available materials for organic solar cells was given for example in [1]. In organic solar cells, the fundamental process of charge separation is the absorption of a photon in the active layer. Typically, photo-generated excitons in organic semiconductors are more difficult to separate into free charges compared to inorganic materials. The low dielectric constant and the higher degree of localization lead to binding energies in the order of 0.1–1 eV, significantly higher than the thermal energy kT at room temperature [2,3]. Owing to this strong binding energy, the average exciton lifetime is quite short – in the order of some hundreds of picoseconds. Furthermore, excitons in organic materials are strongly localized within the conjugation length of their molecules, which leads to a slow diffusion. Typically, the diffusion length is in the order of some tens of nanometers. In the case of poly(3-hexylthiophene) (P3HT), the diffusion length was measured to be LD  9 nm (exciton lifetime s = 400 ps, diffusion constant D = (1.8 ± 0.3)1013 cm2/s) [4]. If the exciton is close to a donor–acceptor junction, the energetic offset between the lowest occupied molecular orbitals (LUMO levels) will facilitate the spatial separation of the charges: the hole remains in the donor material, whereas the electron is transferred into the acceptor material (Fig. 1a). Owing to the short diffusion length, the exciton has to be created close enough to a donor–acceptor junction before recombination takes place. The latter determines the scale of the morphology needed in donor–acceptor blend materials to be in the nanometer realm. Consequently, such morphologies become a key parameter for organic solar cells. Furthermore, an efficient donor–acceptor junction has a thickness in the order of the absorption length (typically some hundreds of nanometers). Both length scales can be achieved by the concept of the bulk heterojunction, where donor and acceptor are intermixed, forming an interpenetrating network (Fig. 1b). Finally, charges must be transported to the electrodes. In order to achieve this it is important that donor and acceptor form percolating networks.

Fig. 1. Energetic diagram of an organic solar cell (a). The charge separation mechanism starts with the absorption of a photon (1), exciting an electron (gray circle) from the highest occupied molecular orbital (HOMO) to the LUMO and leaving a hole (blue circle). The subsequent relaxation leads to the formation of an exciton (2). In order to separate the exciton, it has to find a donor–acceptor junction by diffusion (3), where it can be separated (4). Owing to the electric field generated by the work function offset of the electrodes, the separated charges drift towards their respective electrodes. Here, the absorption takes place in the electron donor material (D); the same mechanism would apply for absorption in the acceptor (A). Adapted from [5]. (b) Schematic view of a bulk heterojunction solar cell which consists of a blend or heterojunction of donor (red color) and acceptor (blue color) material with a phase separation on the order of 10 nm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

For such structures, classic experimental methods, such as current–voltage measurements on macroscopic devices can only provide average information over the microscopic function. Imaging methods such as optical and electron microscopy can provide information on the structure of the material and – with extensions such as Raman, X-ray spectroscopy and electron energy loss spectroscopy – about the composition of the material. However, these methods reach their limits when the energy conversion on the level of individual molecules and nanostructures needs to be studied. Here, scanning force microscopy (SFM) is a method to investigate the above mentioned nanoscopic structures. Uniquely, SFM allows probing electrical properties of organic solar cells such as the contact potential difference, variations in electrical conductivity and local photocurrents. Thus structural information can be directly correlated with electrical information on a nanometer scale. This unique feature is addressed in the

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915

following review. First, we briefly introduce the working principles of conductive scanning force microscopy (CSFM) and Kelvin probe force microscopy (KPFM). Second, we focus on experiments aiming at visualizing photo-induced processes in organic solar cells. For such experiments, an additional light source has to be integrated into the SFM setup. Third, we discuss some challenges and future directions for SFM in the field of organic based electronics.

2. Electrical SFM methods The most common use of SFM (a synonym is atomic force microscopy (AFM)) is the investigation of the topography of surfaces [6]. The basic principle is that the force acting between a nanometer sized tip and the sample surface is kept constant using an electronic feedback circuit upon scanning the tip relative to the sample. Then the electronic feedback signal is used as a measure of the sample topography, h, which has an accuracy <1 nm (Fig. 2a). However for visualizing the function of energy materials, the pure topography imaging is often not sufficient. Here, the use of electrical operation modes extends the range of accessible material properties.

1909

2.1. Conductive scanning force microscopy In contact mode, the tip is in permanent mechanical contact with the sample surface. Thus, on sufficiently conductive samples, an electrical circuit can be established between a conductive tip and the sample. By connecting the SFM tip to a current amplifier and setting the sample on a bias voltage US relative to the tip, the electrical current Itip can be detected (Fig. 2b) [7,8]. Variations in the local conductance G = Itip/Us [X1] of the sample can be studied by scanning the sample at a constant bias voltage. This operation mode is called conductive scanning force microscopy (CSFM [9,10]). The advantage of CSFM is that it uses two independent signals, i.e. the deflection of the cantilever and the electrical current. In contrast, in scanning tunneling microscopy (STM) morphological properties are obtained by adjusting the tip sample separation by keeping a constant tunneling current. In case of electrically insulating domains at the surface, STM would fail. CSFM requires electrically conductive substrates or electrodes (e.g. metal layers, graphite, indium tin oxide) and tips (e.g. metal coated tips, conductive diamond coating [11], PtSi [12]. In addition to topography and current imaging, local current–voltage characteristics can be recorded by positioning the tip at a desired spot on the surface. From the local current voltage dependence, the charge injection mechanism and/or the transport properties of the material under the tip can be studied. This kind of analysis can provide additional information about the material properties, e.g. at surfaces, interfaces and in confined geometries. 2.2. Kelvin probe force microscopy

Fig. 2. Different SFM operation modes: (a) contact mode, (b) conductive SFM and (c) Kelvin probe force microscopy.

Kelvin probe force microscopy (KPFM) is a SFM method that utilizes the tip as a Kelvin probe [13]. A Kelvin probe is able to quantify contact potential differences (UCPD or simply CPD) between a reference probe and a sample surface. The measured UCPD can have different contributions caused by (i) materials with differing work functions, (ii) surface charges or dipoles or (iii) charging, e.g. by photo-induced charge separation. In KPFM, an AC voltage UAC is applied between the tip and the sample. In the presence of an electrostatic field (caused by CPD) between tip and sample the AC voltage leads to an oscillation of the cantilever. By compensating this electrostatic field by a DC bias voltage UDC applied between the tip and the sample, the vibration amplitude can be nullified. At this condition we find that UDC = UCPD (Fig. 1c). By scanning the tip, the UCPD can be mapped with a resolution down to individual atoms [14]. More detailed descriptions of the working principle of KPFM can be found in numerous reviews [10,15,16]. One implementation of KPFM in the feedback electronic circuit is to track changes in the amplitude of the cantilever vibration by means of a lock-in amplifier. This method is called amplitude modulation (AM) KPFM [17]. An alternative strategy is to track changes in the vibration spectrum of the cantilever. Applying an AC voltage of frequency fm well below the cantilever’s resonance frequency f0 leads to a modulation of the cantilever’s resonance frequency and thus to the appearance of satellite peaks in the canti-

1910

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915

lever spectrum at f0 ± fm. Similar to AM-KPFM, those peaks can be nullified by applying a DC bias voltage UDC. The advantage of FM over AM-KPFM is that it is sensitive to the force gradient signal, which decays more steeply with increasing sample distance than the force signal itself. Thus, contributions in the KPFM signal originating from the cantilever and the tip cone are reduced. Zerweck and co-workers have studied the accuracy and resolution of amplitude and frequency modulation methods in ultrahigh vacuum (UHV). They found that the frequency modulation method is preferable in most applications owing to a higher lateral resolution and reduced influence of tip sample distance [18]. However, for a good signal-to-noise ratio higher AC voltages of typically 1–2 V have to be applied on most systems when using FM mode. Secondary effects such as band bending close to the sample surface may occur. In AM mode, the AC voltage can be reduced down to 50–100 mV, thus minimizing band bending effects [19]. 2.3. Photo-induced effects Photo-induced conductance changes and charging effects can be quantified by comparing G or UCPD maps with and without sample illumination. There are several ways to practically implement an additional light source in a SFM setup. In our lab, we use a combination of fiber-coupled light sources and mirrors to guide the light to the sample (Fig. 3). Two light sources are available: a white light source and a 488 nm diode laser. The positions of the light sources are interchangeable to allow for either bottom illumination for experiments on top of the sample or side illumination, e.g. for experiments at cross cuts where the sample is mounted upright. Other setups were reported which allow KPFM voltage spectroscopy at different illumination wavelengths [20]. Here, a 300 W Xe arc lamp is used as light source which is then passed through a monochromator and into glass fibers. The glass fibers are then guided into a UHV KPFM setup. Light emanating from the glass fiber end illuminates then samples from the side at a shallow angle.

Fig. 3. Setup for SFM experiments with sample illumination. Two light sources are available: (1) a white light source and (2) a 488 nm diode laser which is guided via a mirror (4) to the SFM (3).

3. SFM on solar cells First investigations of photo-electric effects were reported for inorganic solar cells and date back to the early days of scanning probe microscopy. In 1991, Weaver and Wickramasinghe reported for the first time the use of scanning surface photo-voltage microscopy on different types of semiconducting samples in air [21]. An additional He– Ne laser beam was used to create electron hole pairs that diffused to the surface and interfaces. Then UCPD were measured in AM-KPFM mode under illumination and under dark conditions. Using this method, microcracks, dopant profiles and dislocations in semiconductors were investigated. Until today many studies were performed on inorganic solar cells [22–25]. Today’s research in organic solar cell structures creates substantial additional requirements to SFM methods. In addition to the required morphology at a 10 nm scale, organic materials used in organic solar cells are much less stable when exposed to ambient air. In particular in combination with light exposure, this often leads to photo-oxidation. Thus, in order to use electrical modes for the characterization of organic solar cells the SFM should be operated under defined and inert environments, e.g. inert gas or even vacuum. Furthermore, most of the materials are soft and prone to tip induced wear. Hoppe et al. investigated an organic polymer solar cell consisting of blends of MDMO-PPV:PCBM cast from toluene and chlorobenzene [26] (MDMO-PPV stands for poly(2-methoxy-5-(30 -70 -dimethyloctyloxy)-1,4-phenylenevinylene)). They performed KPFM in an UHV SFM system with a sample illumination performed with laser light (Fig. 4). Together with scanning electron microscopy they proved that the films cast from chlorobenzene yielded a much finer phase separation compared to toluene cast films. The comparison of surface photo-voltage images showed that charge trapping in isolated PCBM domains is a major cause for the lower performance of toluene cast films compared to chlorobenzene cast films. Chiesa et al. studied blends of poly(9,90 -dioctylfluorene-alt-benzothiadiazole) (F8BT) and poly-(9,90 -dioctylfluorene-co-bis-N, N0 -(4-butylphenyl)-bis-N,N0 -phenyl-1,4-phenylenediamine) (PFB) [27]. They measured the local surface photo-voltage with KPFM as a function of laser intensity over several orders of magnitude. They were able to demonstrate that the KPFM surface photovoltage correlates well with the open circuit voltage measured on a bulk device. By comparing the CPD patterns of areas with and without illumination they identified different micrometer-sized domains of PFB and F8BT rich phases. By correlating the surface photo-voltage with the topography, a model for the three dimensional structure of the active polymer blend could be deduced. Aiming at the highest possible resolution, Spadafora et al. investigated P3HT:PCBM blends with an optimized bulk heterojunction morphology using an UHV non-contact SFM [28]. By optimizing the experimental parameters of the AM-KPFM setup and scanning with very low scan speeds (image acquisition times 7–10 h), a resolution better than 10 nm was obtained on the CPD images of both

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915

1911

Fig. 4. Topography and potentials measured on MDMO-PPV:PCBM cast from toluene. (a) shows a schematic view of the distribution of PCBM rich domains (blue) in the MDMO-PPV matrix. (b) The topography revealed that PCBM rich domains partially poke out from the surface. The effect of the sample illumination can be seen in the lower images [26]. The brighter spots were attributed to PCBM rich domains. The potential images indicate strong charge trapping in the electrically isolated PCBM domains as sketched in (a).

Fig. 5. (a) Schematic of the photo-conductive AFM (pcSFM) setup [33]. (b) The local photo-current measured by pcSFM correlates well with the external quantum efficiency measured on bulk devices. (c) Topography and (b) photo-current data collected on toluene cast MDMO-PPV:PCBM, demonstrating that charge carriers are mainly generated close to the interface between MDMO-PPV and PCBM. On the xylene cast film (e and f), the phase separation is much finer and the current generation more homogeneous. Results taken from [33].

1912

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915

illuminated and dark images. Ellison et al. could show that the measured shift in CPD upon illumination corresponds to the HOMO and LUMO levels of the donor and acceptor materials deposited on top of an indium tin oxide (ITO) substrate [29]. Here the measured CPD is a measure of the open circuit potential in organic solar cells. Therefore, KPFM methods are often sufficient to study the photo response in newly developed nanostructures, e.g. for new donor–acceptor systems [30–32] in order to prove novel synthetic routes and to identify promising new material systems. Coffey et al. [33] combined conductive SFM with a laser sample illumination (k ¼ 532 nm, intensity I = 104 W/m2) and recorded the local photo-current via a current amplifier connected to the tip. Photo-currents were recorded under open circuit conditions, that is, no sample bias voltage was applied during the experiments (Fig. 5a). With this socalled photo-conductive SFM (pcSFM), they investigated MDMO-PPV:PCBM blends with different degrees of phase separation. By recording the photo-current maps over several orders of magnitude in illumination intensity, the authors proved that the average photo-current correlates well with the short circuit current measured on a bulk device (Fig. 5b). The photo-current map recorded on a coarsely phase separated toluene cast blend revealed that only the small border region of 20 nm around the MDMO-PPV/PCBM interface generated a measurable photo-current (Fig. 5c and d). In comparison, the previously mentioned KPFM studies revealed [26,27] similar domains of several hundred nanometers in diameter. These domains had a constant CPD when they were illuminated. On samples exhibiting a smaller scale phase separation of some tens of nanometers, the photo-current was more homogeneous on single domains (Fig. 5e and f). In following publications, the method was used to further investigate the interplay between the nanoscale morphology and the overall device performance. In order to achieve the best possible morphology, devices are often annealed after depositing the active layer. The annealing increases the separation between the donor and acceptor phases [34]. Recently, Kamkar et al. have studied fullerene capped P3HT and studied the performance of films made from different solvents using pcSFM [35]. The measurements revealed that samples made from dichlorobenzene organize in a fibrous network and exhibit a high photo-current. Finally an increasing number of pcSFM studies were published, focusing on alternative morphologies such as P3HT nanowires [36–38], or novel material combinations [39–43]. Although pcSFM is a very powerful technique for the visualization of nanoscale photo-current, it also has some drawbacks. The operation of CSFM is based on contact mode that potentially modifies the surface [44]. Thus, pcSFM experiments are limited to sufficiently stable and smooth surfaces. Furthermore, in order to generate a measurable photo-current, the sample has to have a high power conversion efficiency. Otherwise, strong illumination intensities well above the solar irradiation intensity of 103 W/m2 are required to generate a measurable signal. Thus, the application is limited to systems that have already undergone an extensive optimization of material

synthesis and device preparation parameters. Novel materials and model systems – although highly relevant for testing novel concepts – are thus often not accessible by pcSFM. In conclusion, mapping of photo-induced changes in electrical parameters makes the SFM method particularly interesting for the nanoscale investigation of energy materials. The use of these setups allows comparing SFM data on samples in dark and under illumination at defined light intensities. Furthermore, the wavelenths of the light can be varied in order to determine local electronic band gaps in materials [20] and valuable insights in the structure and working mechanism of organic photovoltaic devices can be obtained. 4. New SFM based methods for organic solar cells In order to identify the structures responsible for the charge generation more precisely, complementary SFM methods are required. With time resolved electrostatic force microscopy (trEFM), Coffey and Ginger introduced an elegant method that combines the advantages of KPFM (non-contact operation) with pcSFM (mapping of local charge carrier generation rate) [45]. In order to achieve the maximum resolution, frequency modulation electrostatic force microscopy (EFM) in interleave mode is used. Here, the cantilever is driven mechanically at its resonance frequency while being lifted 5–100 nm above the sample. Electrostatic fields, for example generated by surface charges, lead to a shift Df in the cantilever’s resonance frequency. This shift is tracked by a phase locked loop. Reid et al. were able to demonstrate that trEFM exhibits a higher sensitivity compared to KPFM when studying degradation patterns in a F8BT:PFB blend material [46]. They used UV light focused onto a sample in the presence of ambient air in order to locally degrade the organic compounds. Photo-oxidation is one of the main reasons for the limited lifetime of present organic solar cells [47]. 5. Challenges and future directions for the electrical characterization of organic solar cell materials with SFM The excerpt given above on the characterization of organic solar cell materials showed that the SFM methods provide valuable information. However, we feel that there are a number of unsolved questions on the instrument, sample preparation and on the materials side which we will point out in the following. 5.1. Material issues Studying degradation processes will most likely gain more and more importance [47]. Apart from the example we presented at the end of the organic solar cell section [46], only a few groups have addressed degradation phenomena on a nanometer scale. Sengupta et al. have used an illumination mask to locally degrade P3HT:PCBM blend films [48]. In this contribution, a method for the quantification of photo-degradation using KPFM and CSFM was described. Since KPFM and CSFM can only measure relative

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915

differences in potentials or conductances, the samples were degraded locally using a micrometer size grid as a shadow mask. This procedure allowed to make a comparison between degraded and non-degraded parts within the same scan area. In addition, the method was used to study the degradation of individual components of the blends, namely P3HT and PCBM. Thus, there is no need for the preparation of fully operational devices, where the degradation processes in the active layer may additionally be influenced by other layers, such as electron transport layers. In a similar study, Lopez-Elvira et al. used focused blue and UV light to locally degrade samples of poly(3-octylthiophene) and investigated the effects with KPFM [49]. They found that blue light had a stronger degradation effect compared to UV. 5.2. Sample preparation issues For quantitative and reproducible SFM experiments it is vital to have defined surfaces, both on the sample and on the tip. As the SFM tip mimics the top electrode, a correct matching of the energy levels is an important issue for electric transport measurements. In the Nguyen group, for example, high work function platinum tips have been used for hole mobility measurements and low work function magnesium coated tips for electron mobility measurements [50]. Measuring the potential distribution inside a working device makes it possible to gain information on the physics at internal interfaces. The preparation of a defined cross cut through an inorganic device, e.g. by fracturing, is usually straight forward as the materials are quite robust against mechanical stress. In an organic solar cell, however, soft materials and poor adhesion between adjacent layers hamper the preparation of cross cuts. Consequently, only few studies have been performed on cross cuts of organic solar cells, so far. Dante et al. have used a focused ion beam (FIB) apparatus to cut a thin slice out of a device and imaged it with CSFM [50]. However, they only imaged the active layer without the adjacent interfaces. Lechmann et al. studied the charge conduction in a hybrid solar cell, that was fractured and subsequently polished by a FIB procedure [51]. By changing the sign of the sample voltage during CSFM imaging, the function of the hole blocking layer was visualized. However, to this point it is unclear, how much the use of high energy ions of the FIB alters the sample structure and function [52]. Hamadani et al. used low angle microtome cutting to prepare a sufficiently flat surface across a P3HT:PCBM solar cell device [53]. The authors found significant structural differences at the surface as compared to the bulk of the active layer, which they attributed to an enrichment of P3HT at the surface. For the interpretation of those experiments, effects on the internal energy levels caused by the additional interface in the sample have to be taken into account. For example, a pinning of the Fermi level at the surface has been observed [54]. In this case, the authors were able to compensate those effects by applying a bias voltage to the sample. Zhang and co-workers and Lee and co-workers simply broke solar cells in air [55] or liquid nitrogen [56],

1913

respectively. These so prepared surfaces were smooth enough for performing KPFM in order to determine the potential distributions in the stack of layers. Zhang et al. found that P3HT:PCBM type solar cells were almost field free at short circuit conditions [55]. An elegant way to obtain the three dimensional structure in a percolating conductor network has been demonstrated in the Loos group [57]. By a stepwise process of slicing off material with a microtome knife and subsequent CSFM imaging after each cutting step, percolation pathways could be followed and possible potential drops at interfaces be studied. 5.3. Instrumental issues A common issue in CSFM and pc-AFM studies on organic materials is the tip-sample interaction during the experiments. Here, new dynamic operation modes such as jumping mode [58,59] (or the commercially available version called peak-force) or conductive torsion mode microscopy [60–62] can lead to a more defined tip-sample interaction during the experiment. Weber et al. have demonstrated, that in torsion mode a very fragile array of vertical nano-pillars made of an organic semiconductor could be imaged non-destructively while simultaneously measuring the tip-sample current [60]. Thus, conductance variations in between individual nano-pillars could be observed. Furthermore, current voltage characteristics recorded on single nano-pillars revealed a space charge limited conduction behavior, yielding further information about the charge transport mechanism in the nano-pillar. Desbief and co-workers have studied the electrical properties of hybrids of a P3HT and carbon nanotubes using timeresolved current sensing force spectroscopy [63]. This method reduced tip sample forces significantly. In particular, this method does not exhibit lateral wear between the tip and sample surface as it is based on measuring the current during force distance curves recorded at a frequency >1 kHz. Desbief and co-workers demonstrated that this method allows to map the current distribution over 15 nm wide nanofibers. With the development of time resolved EFM experiments dynamic processes in photo-active samples can be studied elegantly. However, the temporal resolution is by far not sufficient to track the fundamental photo-physical processes in solar cell materials that are in the range of micro- down to femto-seconds [64]. The use of faster detection electronics and smaller cantilevers featuring resonance frequencies above 1 MHz could be a first step towards enhanced spatial resolution and faster time resolved measurements. Absolute values of the conductivity of materials could be obtained by four point probe methods. Groups have started to fabricate probes on a nanometer scale [65]. However, applications to soft matter or to organic solar cells were not reported so far. In many SFM studies on solar cell materials, in particular in organic systems, a further increase in lateral resolution is desirable. An alternative strategy to UHV experiments is to apply SFM in liquid environments. The complete immersion of sample and tip into a defined liquid eliminates meniscus forces between tip and sample and protects the surfaces from contamination. In combination

1914

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915

with a low-noise SFM setup [66], atomic and molecular topographic resolution has been demonstrated [67]. Recently, KPFM was performed by Domanski and co-workers in non-polar liquids in order to determine work-function changes upon chemisorption of hexadecanethiols on Au [68]. However, applying a voltage between tip and sample, as required for KPFM, can lead to electrochemical reactions and spurious forces in aqueous environments. Thus, conventional KPFM is not possible. Recently, Fukuma and coworkers have published first results on a new ‘‘open loop’’ KPFM technique in aqueous buffer solution [69]. This type of experiment could be applicable for in situ experiments on energy materials such as liquid electrolyte Grätzel cells or electrochemical battery materials that were not accessible by KPFM so far. Acknowledgements We thank the International Research Training Group 1404 ‘‘Self-organized Materials for Optoelectronics’’ (DFG) and the BMBF (contract #03SF0334) and IMPRS for their financial support. S.W. acknowledges the Alexander von Humboldt Foundation for its support. References [1] Facchetti A. Chem Mater 2011;23:733. [2] Campbell IH, Hagler TW, Smith DL, Ferraris JP. Phys Rev Lett 1900;1996:76. [3] Bredas J-L, Cornil J, Heeger AJ. Adv Mater 1996;8:447. [4] Shaw PE, Ruseckas A, Samuel IDW. Adv Mater 2008;20:3516. [5] Kippelen B, Bredas J-L. Energy Environ Sci 2009;2:251. [6] Binnig G, Quate CF, Gerber C. Phys Rev Lett 1986;56:930. [7] Martin Y, Abraham DW, Wickramasinghe HK. Appl Phys Lett 1988;52:1103. [8] Hosaka S, Koyanagi H, Hasegawa T, Hosoki S, Hiraiwa A. J Appl Phys 1992;72:688. [9] Fumagalli L, Casuso I, Ferrari G, Gomila G. In: Bhushan HFB, Tomitori M, editors. Berlin, Heidelberg: Springer; 2008. p. 421. [10] Berger R, Butt HJ, Retschke MB, Weber SAL. Macromol Rapid Commun 2009;30:1167. [11] Oshea SJ, Atta RM, Welland ME. Rev Sci Instrum 1995;66:2508. [12] Bhaskaran H, Sebastian A, Despont M. IEEE Trans Nanotechnol 2009;8:128. [13] Kelvin V. Philos Mag Series 5 1898;46:82. [14] Gross L, Mohn F, Liljeroth P, Repp J, Giessibl FJ, Meyer G. Science 2009;324:1428. [15] Liscio A, Palermo V, Samori P. Acc Chem Res 2010;43:541. [16] Melitz W, Shen J, Kummel AC, Lee S. Surf Sci Rep 2011;66:1. [17] Nonnenmacher M, O’Boyle MP, Wickramasinghe HK. Appl Phys Lett 1991;58:2921. [18] Zerweck U, Loppacher C, Otto T, Grafstrom S, Eng LM. Phys Rev B: Condens Matter 2005;71:125424. [19] Glatzel T, Sadewasser S, Lux-Steiner MC. Appl Surf Sci 2003;210:84. [20] Streicher F, Sadewasser S, Lux-Steiner MC. Rev Sci Instrum 2009:80. [21] Weaver JMR, Wickramasinghe HK. J Vac Sci Technol B: Microelectron Nanom Struct 1991;9:1562. [22] Cao J, Sun J-Z, Hong J, Yang X-G, Chen H-Z, Wang M. Appl Phys Lett 1896;2003:83. [23] Sadewasser S, Glatzel T, Rusu M, Jager-Waldau A, Lux-Steiner MC. Appl Phys Lett 2002;80:2979. [24] Loppacher C, Zerweck U, Teich S, Beyreuther E, Otto T, Grafström S, et al. Nanotechnology 2005;16:S1. [25] Sadewasser S, Lux-Steiner MC. J Vac Sci Technol B 2010;28:C4D29. [26] Hoppe H, Glatzel T, Niggemann M, Hinsch A, Lux-Steiner MC, Sariciftci NS. Nano Lett 2005;5:269. [27] Chiesa M, Burgi L, Kim J-S, Shikler R, Friend RH, Sirringhaus H. Nano Lett 2005;5:559. [28] Spadafora EJ, Demadrille R, Ratier B, Gre´vin B. Nano Lett 2010;10:3337.

[29] Ellison DJ, Kim JY, Stevens DM, Frisbie CD. J Am Chem Soc 2011;133:13802. [30] Palermo V, Ridolfi G, Talarico AM, Favaretto L, Barbarella G, Camaioni N, et al. Adv Funct Mater 2007;17:472. [31] Zorn M, Weber SAL, Tahir MN, Tremel W, Butt H-J, Berger R, et al. Nano Lett 2010;11:2812. [32] zur Borg L, Domanski AL, Breivogel A, Burger M, Berger R, Heinze K, et al. J Mater Chem C 2013;1:1223. [33] Coffey DC, Reid OG, Rodovsky DB, Bartholomew GP, Ginger DS. Nano Lett 2007;7:738. [34] Yang X, Loos J. Macromolecules 2007;40:1353. [35] Kamkar DA, Wang M, Wudl F, Thuc-Quyen N. ACS Nano 2012;6:1149. [36] Reid OG, Xin H, Jenekhe SA, Ginger DS. J Appl Phys 2010;108:084320. [37] Rice AH, Giridharagopal R, Zheng SX, Ohuchi FS, Ginger DS, Luscombe CK. ACS Nano 2011;5:3132. [38] Tsoi WC, Nicholson PG, Kim JS, Roy D, Burnett TL, Murphy CE, et al. Energy Environ Sci 2011;4:3646. [39] Bull TA, Pingree LSC, Jenekhe SA, Ginger DS, Luscombe CK. ACS Nano 2009;3:627. [40] Xin H, Reid OG, Ren G, Kim FS, Ginger DS, Jenekhe SA. ACS Nano 1861;2010:4. [41] Dang X-D, Tamayo AB, Seo J, Hoven CV, Walker B, Nguyen T-Q. Adv Funct Mater 2010;20:3314. [42] Guide M, Dang X-D, Nguyen T-Q. Adv Mater 2011;23:2313. [43] Yan H, Li D, Lu K, Zhu X, Zhang Y, Yang Y, et al. Phys Chem Chem Phys 2012;14:15127. [44] Douheret O, Swinnen A, Breselge M, van Severen I, Lutsen L, Vanderzande D, et al. Microelectron Eng 2007;84:431. [45] Coffey DC, Ginger DS. Nat Mater 2006;5:735. [46] Reid OG, Rayermann GE, Coffey DC, Ginger DS. J Phys Chem C 2010;114:20672. [47] Jorgensen M, Norrman K, Krebs FC. Sol Energy Mater Sol C 2008;92:686. [48] Sengupta E, Domanski AL, Weber SAL, Untch MB, Butt H-J, Sauermann T, et al. J Phys Chem C 2011;115:19994. [49] López-Elvira E, Escasaín E, Baró A, Colchero J, Palacios-Lidón E. Polym Degrad Stab 2011;96:1279. [50] Dante M, Peet J, Nguyen T-Q. J Phys Chem C 2008;112:7241. [51] Lechmann MC, Weber SAL, Geserick J, Husing N, Berger R, Gutmann JS. J Mater Chem 2011;21:7765. [52] Kim S, Park MJ, Balsara NP, Liu G, Minor AM. Ultramicroscopy 2011;111:191. [53] Hamadani BH, Jung S, Haney PM, Richter LJ, Zhitenev NB. Nano Lett 2010;10:1611. [54] Moutinho HR, Dhere RG, Jiang CS, Yan Y, Albin DS, Al-Jassim MM. J Appl Phys 2010:108. [55] Zhang Z, Tang X, Lemmer U, Witte W, Kiowski O, Powalla M, et al. Appl Phys Lett 2011:99. [56] Lee J, Kong J, Kim H, Kang S-O, Lee K. Appl Phys Lett 2011;99:243301. [57] Alekseev A, Efimov A, Lu K, Loos J. Adv Mater 2009;21:4915. [58] de Pablo PJ, Colchero J, G’o M-H, Bar’o AM. Appl Phys Lett 1998;73:3300. [59] Gomez-Navarro C, de Pablo PJ, Colchero J, Fan Y, Burghard M, Gomez-Herrero J, et al. Nanotechnology 2003;14:134. [60] Weber SAL, Haberkorn N, Theato P, Berger R. Nano Lett 2010;10:1194. [61] Sun L, Wang J, Bonaccurso E. J Phys Chem C 2010;114:7161. [62] Vetushka A, Itoh T, Nakanishi Y, Fejfar A, Nonomura S, Ledinsky M, et al. J Non-Cryst Solids 2012;358:2545. [63] Desbief S, Hergue N, Douheret O, Surin M, Dubois P, Geerts Y, et al. Nanoscale 2012;4:2705. [64] Giridharagopal R, Rayermann GE, Shao G, Moore DT, Reid OG, Tillack AF, et al. Nano Lett 2012;12:893. [65] Ju BF, Ju Y, Saka M. J Micromech Microeng 2005;15:2277. [66] Fukuma T, Kimura M, Kobayashi K, Matsushige K, Yamada H. Rev Sci Instrum 2005;76:053704. [67] Sheikh KH, Giordani C, Kilpatrick JI, Jarvis SP. Langmuir 2011;27:3749. [68] Domanski AL, Sengupta E, Bley K, Untch MB, Weber SAL, Landfester K, et al. Langmuir 2012;28:13892. [69] Kobayashi N, Asakawa H, Fukuma T. Rev Sci Instrum 2010;81:123705.

R. Berger et al. / European Polymer Journal 49 (2013) 1907–1915 Rüdiger Berger studied physics at the Friedrich-Alexander University of Erlangen-Nürnberg. He got his diploma in Physics in 1994 in the field of high Tc-superconductors where he worked on imaging and manipulation methods using scanning probe microscopy. Then he moved to Switzerland where he made his PhD thesis and a PostDoc in the field of micromechanical sensors at the IBM Zurich Research Laboratory. In 1998, he joined the Analysis Laboratory at the IBM Deutschland Speichersysteme GmbH (Mainz). Here, he worked on the automation of test systems, in situ flight height control and the investigation of ion beam structuring of magnetic materials. Since October 2002, Rüdiger Berger is working at the MPI for Polymer Research where he is investigating surfaces using scanning force microscope methods.

Anna Domanski studied chemistry at the Friedrich-Alexander University ErlangenNürnberg and at the Johannes-Gutenberg University Mainz and received her diploma in 2010. For her diploma thesis, she worked in the group of Prof. Hans-Jürgen Butt at the Max Planck Institute for Polymer Research and studied the degradation processes of organic solar cell materials by means of electrical SPM modes. Since September 2010 she is doing her PhD in the group of Prof. Hans-Jürgen Butt as a stipend holder of the Korean-German

1915

International Research Training Group (IRTG) ‘‘Self-Organized Materials for Optoelectronics’’. Her main focus is on the application of Kelvin Probe Force Microscopy in liquid environments.

Stefan Weber studied physics at the University of Konstanz and received his diploma in 2007. For his diploma thesis, he worked in the group of Prof. Johannes Boneberg and Prof. Paul Leiderer and studied gold nanoparticles under pulsed laser irradiation. For his PhD he moved to the MPI for Polymer Research to the group of Dr. Rüdiger Berger and Prof. HansJürgen Butt. In a joint German-Korean Graduate School he worked on electrical SPM modes and their application to organic optoelectronic materials, such as polymer solar cells, and spent six months at Seoul National University, Korea. After receiving his PhD in 2010 he went to University College Dublin, where he worked in the group of Prof. Suzi Jarvis and Dr. Brian Rodriguez on lownoise AFM in liquid environments. Since 2012 Stefan Weber is working at the MPI for Polymer Research on combining the high-resolution of lownoise AFM with electrical SPM modes in liquid.