Charge generation mechanism in organic solar cells

The level of interest in organic photovoltaics is extraordinary in many ways. A very large number of scientists from different areas of the physical sciences have been attracted to the field that promised a revolution in solar cell technology and the results of this effort have been remarkable. With the current top power conversion efficiency close or exceeding 10%, up from the ∼1% of the original prototypes from almost 20 years ago, it seems that the coordinated work of physicists, material scientists and chemists is paying off. However, possibly not less extraordinary is the realization that few crucial aspects of the operation of an organic photovoltaic cell are still not well understood. Possibly the most controversial of all is the mechanism for the generation of free charges from the exciton generated by solar radiation. In particular the microscopic mechanism of the charge generation and its detailed steps have been broadly discussed. A plethora of spectroscopic, device physics and theory studies have discussed the role of charge transfer states in the generation of free charges and in the final device performances. This special issue of Physical Chemistry Chemical Physics proposes a snapshot of the current thinking on these topics. Its contributions represent very well how a satisfactory understanding of this problem can only emerge from a combination of many different types of experiments and models.

The perspective by Gao and Inganäs (DOI: 10.1039/C4CP01814A) helps in framing the problem by clearly separating some well-established facts (e.g. the existence of bound charge transfer (CT) states) from contrasting theories put forward to describe the observations (e.g. ultrafast generation of free charges vs. dissociation of the the CT state into a free carrier). The paper stresses how the mechanism based on the dissociation from the CT state is at the same time more consistent with a few recent experiments but also more difficult to turn into a quantitative model that explains efficient charge separation. The authors discuss how further measurements exploring the effect of the electric field, the temperature, the charge mobility, and the disorder may be necessary for a quantitative model of charge separation.

The overview of the techniques used to extract information on the charge generation process highlights a mismatch between the very fast (≲ ps) timescale investigated by pump–probe measurements of exciton dissociation and the (very slow) steady state or quasi steady state condition employed to measure the wave-length dependent internal quantum efficiency. A methodology to connect the different timescales is proposed by Jones, Dyer, Clarke and Groves (DOI: 10.1039/C4CP01626B) who adopted a Kinetic Monte Carlo method to model the formation of hot charge transfer states. Surprisingly, they find that these states, that may very well be the ones observed by spectroscopists, have little beneficial effect to the device performance as they have the tendency to form bound CT states for the most plausible range of model parameters.

Clearly, to improve the connection between spectroscopic observation and kinetic models one needs a microscopic theory of the dynamical processes that captures the quantum mechanical nature of the charge generation process. Smith and Chin (DOI: 10.1039/C4CP01791A) present a model Hamiltonian study of charge separation that incorporates the effect of vibrational relaxation and therefore energy dissipation. In the short time of charge separation the nuclear modes are essentially immobile, while, if one considers the thermalized CT state, this is stabilized by electron phonon coupling. The authors show that this term also promotes coupling with other electronic states allowing the dissociation of the hole–electron pair.

Alongside Kinetic Monte Carlo and quantum dynamics a third branch of theoretical modelling, computational chemistry, helps in describing the microscopic detail of the interface and suggests plausible scenarios encountered in realistic materials. The contribution by Caster, D'Avino, Muccioli, Cornil and Belijonne (DOI: 10.1039/C4CP01872A) provides an overview of the power of computational methods to describe the energetics of organic heterojunctions. Molecular dynamics can be used to model the structure of the interface, quantum chemistry to describe the local density of states and microelectrostatics to capture long-range electrostatic effects. It is interesting to note that a complete theoretical description of the device may stem from the combination of the approaches exemplified by the three papers above.

It has been speculated for many years that local morphology plays an important role in the process of charge separation on the basis of different performances recorded by blends subject to different processing protocols. It is now widely believed that many high efficiency polymers possess amorphous and ordered regions, even when the ordered regions are not directly evident from crystallographic measures, and that the description of charge/exciton dynamics needs to account for the presence of these separate phases. Two contributions in this issue explore the role of the local ordering of the polymer in bulk heterojunctions. Mangold, Bakulin, Howard, Kästner, Egbe, Hoppe and Laquai, (DOI: 10.1039/C4CP01883D) studied a blend where the acceptor is mixed with two donor polymers, extremely similar in chemical structure, one of which having the tendency to form ordered domains and the other being fully amorphous. The device characterization and the determination of the charge generation and recombination parameters (via transient absorption pump-probe and pump-push photocurrent spectroscopy) lead to the interesting observation that the best cells are obtained by using the semicrystalline polymer blended with a small amount of amorphous polymer.

The role of crystallinity was also explored by Tamai, Tsuda, Ohkita, Benten and Ito (DOI: 10.1039/C4CP01820F) who were able to identify different spectroscopic signatures for the holes generated in the ordered and in the disordered phases. The geminate recombination is suppressed in the most ordered phase suggesting that the delocalization of the charge carrier is extremely beneficial for efficient free charge generation, an opinion that seems to be shared by several authors in this issue.

Disorder does not only influence the delocalization of charge carrier but plays an essential role in many charge transfer processes at the interface. Bittner, Lankevich, Gélinas, Rao, Ginger, and Friend (DOI: 10.1039/C4CP01776E) explored a model system of varying dimensionality to study the suppression of triplet charge recombination in bulk heterojunctions with relatively ordered fullerene phases. The increased disorder has the qualitative effect of increasing the interface localization of the CT state, therefore increasing the transition rate between triplet CT states and the triplet state in the fullerene.

Reading these contributions reinforces the idea of organic photovoltaics as one of the most dynamic and diverse fields of physical science. We hope that the readers my find this special issue a valuable starting point to contribute to an improved understanding of these devices.

Maria Antonietta Loi

Photophysics and OptoElectronics, Zernike Institute for Advanced Materials, University of Groningen

Alessandro Troisi

Department of Chemistry, University of Warwick

---