LETTER pubs.acs.org/NanoLett
Imaging the Bulk Nanoscale Morphology of Organic Solar Cell Blends Using Helium Ion Microscopy Andrew J. Pearson,*,† Stuart A. Boden,‡ Darren M. Bagnall,‡ David G. Lidzey,† and Cornelia Rodenburg§ †
Department of Physics and Astronomy, University of Sheffield, Hicks Building, Hounsfield Road, Sheffield, S3 7RH, U.K. Electronics and Computer Science, University of Southampton, Highfield, Southampton, SO17 1BJ, U.K. § Department of Engineering Materials, University of Sheffield, Mappin Street, Sheffield, S1 3JD, U.K. ‡
bS Supporting Information ABSTRACT: We use helium ion microscopy (HeIM) to image the nanostructure of poly(3-hexylthiophene)/[6,6]-phenyl-C61-butric acid methyl ester (P3HT/PCBM) blend thin-films. Specifically, we study a blend thinfilm subject to a thermal anneal at 140 °C and use a plasma-etching technique to gain access to the bulk of the blend thin-films. We observe a domain structure within the bulk of the film that is not apparent at the film-surface and tentatively identify a network of slightly elongated PCBM domains having a spatial periodicity of (20 ( 4) nm a length of (12 ( 8) nm. KEYWORDS: Polymer solar cells, helium ion microscopy, P3HT, PCBM
O
rganic semiconductors are promising materials for the future production of opto-electronic devices.1 Organic photovoltaic (OPV) devices are the subject of intense research interest, as they offer the prospect of a low cost alternative to existing technologies for solar energy generation.2 To date, the most efficient OPVs use a bulk-heterojunction (BHJ) concept for the semiconducting layer within the device,3 whereby electron donating semiconductors (typically conjugated polymers) are intimately mixed with electron accepting semiconductors (typically functionalized fullerenes).4 The ideal length scale for mixing should facilitate the efficient generation of free charge carriers and provide networks for their transport and extraction from the semiconducting layer.5 This length scale is estimated to be of the order of 5 10 nm to allow for the efficient dissociation of photogenerated excitons.6 To resolve structure in organic semiconductor thin-film blends at a length-scale of ∼10 nm is a challenging goal, however such information may assist the development of bulk heterojunction (BHJ) OPVs having improved efficiency. Transmission electron microscopy (TEM) has been used to determine the nanoscale structures present in thin film blends that have a thickness relevant to OPV operation (∼100 nm),7,8 however such images represent a projection of structures over the sample thickness measured.9 Difficulties with this technique arise in the correct interpretation of contrast differences from compositionally similar materials in bright field data,8,10 since contrast may also arise due to defocus artifacts and thickness variations. Indeed, Kiel et al.11 noted that the bright and dark regions in cross-sectioned blend thin-films of poly(3-hexylthiophene)/ [6,6]-phenyl-C61-butric acid methyl ester (P3HT/PCBM) are r 2011 American Chemical Society
also present in cross-sectioned thin-films of polystyrene. Furthermore, the breakdown of crystalline order in a polymeric material can also take place after a few seconds of exposure to an electron beam.10 Attempts to address projection issues have been made using a focal series12 or tomography reconstructions of brightfield images.9,13 A quantification of crystalline P3HT content in P3HT/PCBM blend films as function of depth has been attempted,9,13 however it was conceded that the technique was insensitive to amorphous P3HT, with prolonged sample exposure to the electron beam causing the specimen to shrink in thickness by 25 30%.9 To reduce the influence of beam damage, energy filtered transmission electron microscopy (EFTEM) has been used image P3HT/PCBM thin films.10 Reconstructions obtained from bright-field images and energy-filtered images suggested the presence of a three-dimensional network of P3HT fibers, however the size and distribution obtained from both image reconstruction techniques were inconsistent.10 Here we describe and evaluate a microscopy technique based on secondary electron (SE) emission resulting from the impact of helium ions. SE emission obtained from argon and nitrogen ions focused onto an area of 0.35 cm2 was previously found to be an effective tool for electron transport studies on carbon based materials,14 however to visualize the nanoscale compositional variations of a BHJ film a finely focused ion beam is necessary. This requirement is fulfilled by the recently developed helium ion microscope (HeIM),15 a technique that is capable of imaging Received: July 5, 2011 Revised: August 30, 2011 Published: September 16, 2011 4275
dx.doi.org/10.1021/nl202269n | Nano Lett. 2011, 11, 4275–4281
Nano Letters samples using a probe having a diameter of less than 1 nm.16 A HeIM is conceptually similar to a scanning electron microscope (SEM), however helium ions generated by a gas field ionization source are used as a primary beam. As image generation results from SEs emitted from the sample surface, there is no need to prepare specimens that are transparent to electrons. Furthermore, as the mean escape depth of SEs is only a few nanometers (3 nm for P3HT17), HeIM is not subject to issues relating to projection, hence tomography techniques are not required to obtain true 2D information. It is clear however that regular SEM can also provide true 2D information regarding a sample surface. Here, the number of SEs emitted from an organic materials is directly related to the electronic structure of the molecule.18 In particular, it has been observed that materials with a high proportion of π-electrons emit a smaller number of SE than those with a smaller fraction of π electrons and therefore appear darker in an SEM image.18 This suggests that both SEM and HeIM should be ideal tools to characterize position-dependent chemical structure across a surface of interest. Despite this, it has been found that SEM imaging can rapidly induce damage when operated at high-magnification when imaging organic-semiconductor thin films. We note that sample damage can still occur when imaging using HeIM, however this is minimized as a much lower primary beam current relative to the SEM is required to form an image. Such lower beam currents can be used as the number of SEs emitted per incident ion is substantially larger than the electron SE yields.19 Although the SE yield in an SEM can be increased by using a relatively low primary beam accelerating voltage (1 kV or less), spatial resolution is compromised because SEM resolution is not limited by the probe size, but by the interaction volume.20 In contrast, the resolution limiting effect of the interaction volume in HeIM is negligible for light elements such as carbon.16 For this reason, HeIM offers the prospect of being a superior microscopy technique compared to SEM for imaging samples relevant to organic optoelectronic devices. In this work, we show that HeIM can be used to map the nanoscale distribution of materials in a thin-film of P3HT/ PCBM with an acquisition time of approximately one minute, thus enabling a greater region of a sample to be probed compared to alternative microscopy techniques where nanoscale resolution has been demonstrated. For example, Kelvin probe force microscopy (KPFM) has been used to achieve a spatial resolution of 2 4 nm, where scan times of 7 10 h were required.21 We also show that buried structure within the film is accessible by using a plasma treatment to selectively remove surface material. We use an image processing procedure based on the HeIM image gray scale distribution to describe the distribution of PCBM domains quantitatively and conclude (in the films studied here) that PCBM domains are typically elongated and have a major axis of approximately 12 nm in length. We believe that this technique will be useful in the analysis of a range of composite organic materials whose electronic properties are determined by their nanoscale structure. We first consider the SE emission from P3HT and PCBM. Figure 1a shows the chemical structure of P3HT and PCBM along with their filled and vacant frontier energy levels. Figure 1b,c shows a 1500 1500 nm2 HeIM scan of pure P3HT and PCBM thin films, respectively. These images were acquired under identical imaging conditions with no plasma treatment or image processing applied. It can be seen that the image corresponding to a P3HT film (Figure 1b) appears relatively brighter
LETTER
Figure 1. The chemical structures and vacant and filled frontier energy levels of the materials studied are presented in panel (a). Panels (b) and (c) present low-magnification HeIM images of as-cast films of P3HT and PCBM, respectively, prior to plasma cleaning. The scale bar represents 500 nm. Normalized gray level distributions of each sample are presented in part (d) alongside the measured background signal from a beam-blanked PCBM thin-film.
than that of PCBM (Figure 1c). This is confirmed quantitatively by the histogram distribution of gray levels for each image, presented in Figure 1d. Included alongside is the gray level distribution for a beam-blanked image. These raw data show that both films appear brighter than the measured background signal and that P3HT appears brighter (average gray level of 116) than PCBM (average gray level of 70) when identical microscope brightness and contrast settings are used. These observations are consistent with the work of Kishimoto et al.18 who observed that the efficiency of SE emission is linked to the ratio of π-bonds to the sum of π- and σ-bonds in a molecule. 4276
dx.doi.org/10.1021/nl202269n |Nano Lett. 2011, 11, 4275–4281
Nano Letters
Figure 2. HeIM images of an annealed P3HT/PCBM blend thin-film subject to plasma cleaning for a period of (a) 3, (b) 6, and (c) 14 min. The scale bar represents 150 nm. A thin-film of as-cast (unannealed) P3HT/PCBM, subject to 14 min plasma cleaning is presented in panel (d). Thin-films of PCBM and P3HT imaged at the same magnification as the blends after subject to plasma etching for 14 min are presented in panels (e) and (f), respectively.
Figure 2a d shows HeIM images of a series of (60:40 by wt%) P3HT/PCBM blend thin films having a thickness of 74 ( 2 nm. Panels (e,f) show control images of PCBM and P3HT thin films. Prior to imaging, films shown in panels (a c) were annealed at 140C for 10 min in nitrogen. The annealing protocol used in this work is within the range commonly used to fabricate efficient P3HT/PCBM OPVs.22 This step promotes phase separation between P3HT and PCBM, leading to the formation of nanoscale domains of these materials.23 All films were also subjected to a plasma etching process that was applied for a range of different times, with images shown in Figure 2a,b being exposed to the plasma etch for 3 and 6 min respectively, while films presented in panels (c f) being plasma etched for 14 min. As shown in Figure 2a, the annealed P3HT/PCBM blend does not initially show any evidence of a phase-separated domain structure. We believe that this may reflect the existence of a 1 2 nm thick P3HT wetting layer located at the film surface24 that is believed to exist even after a thermal annealing step that encourages PCBM migration to the surface.25 We also note that prior to imaging, all samples were briefly exposed to air and that HeIM imaging is not conducted under ultrahigh vacuum.
LETTER
Figure 3. (a) Line profiles of the HeIM image (top), AFM height (middle), and AFM phase (bottom) images, illustrating the differences in spatial periodicity measured by the three techniques. AFM height (b) and phase (c) images of the annealed P3HT/PCBM blend thin-film after 14 min of plasma etching. The color scale in (b) is 10 nm.
The absence of any prominent structure on untreated films may therefore arise from a combination of vertical stratification of P3HT and the presence of surface contaminants. We note that previous EFTEM imaging of a 20 nm thick P3HT/PCBM film was also unable to provide images of nanoscale phase separation8 with structure only being observable in films whose thickness exceeded 100 nm. To determine whether a phase-separated domain structure exists within the bulk of the films, a plasma-etching treatment was used to selectively remove material from the film surface. The effect of the plasma treatment is immediately apparent as shown in Figure 2b,c, where a distribution of bright and dark regions is observed. This is in direct contrast to the image presented in Figure 2a of an otherwise identical film that was only subjected to a short (3 min) plasma etch. We have characterized the etching process using atomic force microscopy and find that the 14 min etch process removed the top 26 ( 4 nm of the film. We find that the plasma etch increases the rms surface roughness of the P3HT/PCBM blend film as measured by AFM from 0.7 to 1.5 nm. It is known that surface structure can influence SE 4277
dx.doi.org/10.1021/nl202269n |Nano Lett. 2011, 11, 4275–4281
Nano Letters
LETTER
Figure 5. Panel (a) shows gray level distributions of the annealed and as-cast blend samples with panel (b) showing gray-level distributions for the pure P3HT and PCBM films. Panel (c) shows a fit of the annealed sample data shown in panel (a) to two Gaussian functions. Here, a threshold level defined using an arrow is used to create a threshold image as shown in panel (d). Black areas in the threshold image are attributed to PCBM rich regions. The scale bar represents 150 nm. Figure 4. Power spectral density (a) and autocorrelation functions (b) of the annealed and untreated sample images presented in Figure 2c,d, respectively. The high power components at spatial frequencies
