Physical processes in organic solar cells

Abstract

For a sustainable future of humankind on this planet a change from fossil fuels as primary energy source to renewable energy sources is needed. Photovoltaics – direct conversion of light into electricity – has a high power per area density compared to other renewable energy sources and can therefore provide a significant contribution to the future energy demand. The growing energy demand requires low cost, fast and large scale power production. Organic solar cells have the potential to meet these requirements, due to the fact that their photoactive layer can consist of cheap organic materials and can be fabricated from solution by fast roll to roll processes. The power conversion efficiency of an organic solar cell strongly relies on the different physical processes involved in the generation of electricity. Each physical process from absorption of light to the generation, transport, and collection of free carriers is crucial for the solar cell’s efficiency. The work described in this thesis focuses on improving our understanding of physical processes that affect the operation and efficiency of organic solar cells. Also a strong aim of the work presented in this thesis is to explore new experimental techniques and improve existing experimental techniques to characterize organic solar cells and study physical processes occurring in organic solar cells. In Chapter 2, electroabsorption (EA) spectroscopy is explored as a non-invasive optical experimental technique to probe electric fields in organic tandem and in organic single junction solar cells. The electric field across the active layer is crucial for efficient charge transport and collection. Especially for organic tandem solar cell it is presently not fully understood how the intermediate recombination contact distributes the electric fields over the individual sub cells. EA-spectra varying with the applied voltage bias have been recorded for single junction and tandem solar cells. The individual contributions of the wide and small band gap photoactive layers can be identified and distinguished in the EA-spectra of the tandem cell. However, the observation of a non-linear voltage bias dependence of the tandem’s EA-signal prevents the determination of the correct electric field distribution in the tandem cell. In the small band gap single junction solar cell, the non-linear dependence of the EA-signal with applied bias voltage is also observed. We were able to attribute this non-linear voltage dependence of the EA-signal to the contribution of two different electro-modulated (EM) signals with different electric field dependencies. The EM-signals are determined to be the EA-signal due to the Stark effect and an EM-signal that results from induced charges in the active layer. Despite the identification of these EM-signals, it was not possible to separate the voltage dependence of the EA-signal from the Stark effect and from the EM-signal from induced charges. Due to the mixing of different EM-signals we concluded that EA-spectroscopy is not a suitable technique to probe routinely electric fields in organic tandem solar cells. Photocurrents are studied experimentally and theoretically in Chapters 3 and 4. The photocurrent in an organic solar cell is defined as the difference between illumination and dark current. Photocurrent - voltage (Jph-V) measurements provide information about the electric field dependence of generation, transport, collection, and recombination of generated carriers. In Chapter 3, photocurrents in organic solar cell are studied experimentally and theoretically with drift-diffusion modeling. In the measured Jph-V curves a reduced photocurrent is observed in forward bias with respect to reverse bias. This reduction is primarily an artifact, caused by the series resistance of the electrodes. Without correcting the measured photocurrent for the series resistance, it is even not qualitatively correct. After correcting for the series resistance, the photocurrent is actually not reduced in forward bias. With drift diffusion simulations the effects of increased recombination rate and different injecting properties of the contacts on the photocurrent are studied. The simulations show that the photocurrent can be reduced in forward bias by bimolecular recombination between photogenerated and injected charges. Furthermore, the simulations show that band-bending or self-selective contacts are not needed to reduce the photocurrent in forward bias. With a simple analytical model we show that the photocurrent under high forward bias can be expressed by: Where ?pre is the Langevin prefactor, ? a positive constant and Jmax is the reverse saturation current. The equation represents the saturation current in forward bias which is related to the reverse bias saturation current and to the strength of bimolecular recombination. For very low ?pre, the forward bias photocurrent is expected to saturate close to Jmax. The reverse saturation current scales with the generation rate in the device and consequently the photocurrent cannot exceed the photon flux. In contrast to this, we show in Chapter 4 that it is possible to observe in forward bias a larger photocurrent than the saturation current in reverse bias for non-annealed and annealed thick P3HT:PCBM solar cells. A larger photocurrent than the number of photons absorbed per unit time is related to a photomultiplication mechanism. Because the photocurrent is defined as the difference between illumination and dark current, photomultiplication occurs in forward bias when the injection current is enhanced under illumination. Photomultiplication can also be observed in drift-diffusion simulation for very low Langevin prefactors or when electron and hole mobilties are different in combination with a low Langevin prefactor. The simulations show that photocurrent multiplication results from a modification of the space charge in the device under illumination. In Chapter 5, we demonstrate experimentally and analytically that the commonly method used to measure spectral responsivity of a solar cell with bias illumination is not correct when the solar cell presents non-linear light intensity dependent losses. A new method is proposed that determines accurately the spectral response under bias illumination. With the new method a much more accurate estimate of the power conversion efficiency can be obtained under bias illumination. The external quantum efficiency (EQE) is defined as the ratio of the number of collected charges at short circuit to the number of photons incident on a solar cell. Typically, the EQE is measured as function of wavelength and is used to determine the short circuit current at the AM 1.5 G solar emission spectrum. It is common practice to illuminate the solar cell at an equivalent intensity of 1 sun to account for intensity dependent losses in the solar cell. In this thesis it is shown that this leads to underestimated EQE values when intensity dependent losses are present in the solar cell. An alternative method to determine EQE values at 1 sun equivalent intensity has been developed. It is shown experimentally and theoretically that the new method to determine EQE at 1 sun equivalent intensity is correct. In Chapter 6, the use of lasers as monochromatic bias illumination on organic solar cell is investigated. Compared to traditional white light bias sources such as halogen or xenon lamps, laser illumination is much easier to handle due to the extremely low beam divergence. Furthermore, monochromatic illumination is required to measure the spectral response of the individual subcells of an organic tandem solar cell. Although, monochromatic laser illumination is more easy to use than white light to bias organic tandem solar cells, it is not clear under which conditions this can be done. The effect of the laser beam intensity profile and the effect of the monochromatic illumination on the J-V characteristics is addressed. We show experimentally and by a 2D analytical model that a non homogenous light intensity profile over the active area of the cell associated with laser light can reduce the performance of the solar cell through a loss in fill factor. This effect can easily be minimized by increasing the beam width. We further show by combined optical and drift diffusion modeling that the wavelength of the monochromatic light does influence the generation profile and can thereby influence the collection of charges via a complex combination of effects. Therefore, a difference in the simulated J-V curves can be observed for different illumination wavelength. Because, the J-V characteristics of an organic solar cell can be dependent on the illumination profile and on the intensity profile of the laser, we advise the use of a solar simulator that approaches the AM 1.5 G reference spectrum and has an uniform intensity for the accurate determination of the organic solar cell’s performance.