PHYSICAL REVIEW B 86, 235201 (2012) Investigation of trap states and mobility in organic semiconductor devices by dielectric spectroscopy: Oxygen-doped P3HT:PCBM solar cells Oskar Armbruster, 1,* Christoph Lungenschmied, 2,† and Siegfried Bauer 1 1 Department of Soft Matter Physics, Johannes Kepler University, Altenberger Straße 69, A-4040 Linz, Austria 2 Konarka Austria F&E GmbH, Altenberger Straße 69, A-4040 Linz, Austria (Received 26 September 2012; published 3 December 2012) We investigate the dielectric response of solar cell devices based on oxygen-doped poly(3- hexylthiophene):[6,6]-phenyl-C 61 -butyric acid methyl ester (P3HT:PCBM) blends as a function of temperature between 133 K and 303 K. The spectra are analyzed using a recently introduced model [O. Armbruster, C. Lungenschmied, and S. Bauer, Phys. Rev. B 84, 085208 (2011)] which is based on a trapping and reemission mechanism of charge carriers. A dominating trap depth of 130 meV is determined and the broadening of this trap level identified as purely thermal. In addition we estimate the density of charge carriers after doping as well as their mobility. We show that the concentration of mobile holes approximately doubles by heating the device from the lowest to the highest measured temperature. This is indicative of a second, shallow trap level of approximately 14 meV. Dielectric spectroscopy hence proves to be a valuable tool to assess device parameters such as dopant concentration, charge carrier transport characteristics, and mobility which are of crucial interest for understanding degradation in organic semiconductor devices. DOI: 10.1103/PhysRevB.86.235201 PACS number(s): 68.55.Ln, 88.40.jr, 77.22.Gm, 73.50.Gr I. INTRODUCTION Organic solar cells have been shown to degrade upon exposure to atmosphere. The limited lifetime compared to their inorganic competitors is considered a major obstacle for widespread application and commercial success. In the following we demonstrate how dielectric spectroscopy can be used to elucidate phenomena such as doping, charge carrier trapping, and mobility in degraded organic photovoltaic (OPV) devices. We introduce a model which allows for the accurate description of their impedance spectra and the extraction of important device parameters by fitting to experimental data. Despite the sensitivity of organic semiconductors, re- spectable lifetime results have been obtained, owing to the use of flexible barrier materials 2,3 or a favorable de- vice architecture. 4 Alternatively, designing materials with improved intrinsic stability is highly promising. The chal- lenge here is that oxygen, light, or water may not only interact with the absorber materials itself, but also with electrodes or interlayers and thereby negatively influence the device performance. 2,3,5,6 Studying the details of the predominant degradation mechanism in the photoactive layer under certain well-controlled conditions can be a first step to better understand and design more stable materials. Un- fortunately degradation processes are not yet understood in sufficient detail in materials used for OPVs. Even for poly(3- hexylthiophene):[6,6]-phenyl-C 61 -butyric acid methyl ester (P3HT:PCBM), probably the best-studied polymer:fullerene blend in OPVs to date, degradation mechanisms under var- ious aging conditions require more attention. Details of the interaction of P3HT:PCBM thin-film devices with oxygen and light have been published recently. While these reports agree that a fundamental process is the doping of P3HT, 7–9 the consequences for charge transport and photovoltaic properties call for further investigation. In this contribution, we report on dielectric spectroscopy data on P3HT:PCBM-based solar cells at various temperatures when degraded by exposure to oxygen and light. We attribute the features observed in the impedance spectra to charge transport in the presence of traps. Upon exposure to oxygen and light, mobile (positive) charges are introduced in the active layer of organic semiconductor devices. 5 Their ability to respond to an externally applied ac signal is limited by the presence of trap states which can capture the hole, immobilize it for a certain period of time, and eventually release it. 10 The average trapping time depends on the depth and capture cross-section of the trap and on the temperature of the hole. Figure 1 shows an illustration of the levels involved in charge transport and trapping. We identify two types of traps: deep traps 130 meV above the transport level and shallow traps with a depth of approximately 14 meV. The second type is within thermal energy at room temperature (≈ 25 meV) and can hence more easily take part in charge transport compared to the deep-trap level. Applying an ac voltage to the two electrodes causes a periodic displacement of the mobile charges in the polymer. 11–13 The frequency of the ac voltage determines the average distance a hole can move in one half cycle. This movement of charges through the material is determined by the buildup of the electric field inside the device. It thus contributes to the out-of-phase current response (i.e., the capacitance) and is eventually limited by the width of the depletion region. II. MODEL In order to model the measured impedance data we describe our sample as a parallel-plate capacitor filled with a dielectric of a certain relative permittivity (ε ). Doping introduces addi- tional holes in the active layer which causes the low-frequency capacitance to significantly exceed the geometric capacitance. In the presence of traps these charges may be immobilized for a certain time and only contribute to the capacitance when thermally reemitted. In order to account for this phenomenon we have previously introduced a Gaussian density of trap states (DOS) and assumed a Boltzmann term to describe the escape from the traps. In addition to this charge trapping 235201-1 1098-0121/2012/86(23)/235201(7) ©2012 American Physical Society