Context and background of the ORGAVOLT project The context of the ORGAVOLT project is described in a recent document by Mark E. Casida [1], a member of the ORGAVOLT team at Grenoble. Below are some further explanations, with a minimum of references. The aims of the ORGAVOLT project are twofold, (i) to develop new theoretical/computational methods for treating “excitonic” materials that exhibit long range Coulomb interactions between electrons and holes and (ii) and to apply this theory to organic semiconductors and heterointerfaces that are relevant to organic solar cells. Pilot production, by printing, of organic solar cells (source?)
The surface needed when using solar energy as the only source of energy is that of the black dots
The technological challenge of organic solar cells The first of the above figures taken from [2] shows that the area needed to provide the planet with its energy from sunlight using solar cells is very small. This simple fact is at the root of the strong growth of solar energy, with installed solar peak power doubling roughly every 2 years [2]. Organic bulk heterojunction (BHJ) cells were developed by Alan Heeger and colleagues in 1995 [3] to lower the cost of solar cells. Since then, twenty years of research and development have resulted in BHJ cells of about 10% efficiency as of today, and promising efforts of printing large area devices are under way, see the second figure above.
The challenge of organic solar cells to theory In spite of impressive progress in the efficiency of BHJ cells over the last 20 years, there is still no general consensus on the physical mechanism that generates their photo current. Reviews [4] generally ascribe the photo current as due to excitons that are photo generated in one of the bulk phases and which then diffuse towards the donor/acceptor interface where they are separated into charge carriers that flow to the electrodes. Alan Heeger and others argue instead [5] that exciton diffusion is too slow to account for the fast current response that occurs less than 100 femto seconds after an incident laser flash. The absence of a reliable and predictive theory of organic semiconductors is certainly due to their exceptional complexity. They are indeed far more complex than traditional inorganic semiconductors that are made of
atoms, since the constituents of organic semiconductors are conjugated molecules that are made of many atoms themselves.
Organic semiconductors as an algorithmic challenge Beginning in the early 80’s of the last century, Hedin’s GW self energy was used successfully to compute the gaps of inorganic semiconductors from first principles i.e. without any free parameters. This breakthrough depended on the availability of computers that were powerful enough to solve the equations of Hedin’s GW approximation. While computer hardware kept steadily improving over the last 30 years, we believe that the challenge of extending Hedin's GW theory from inorganic to organic semiconductors may be more easily met by combining (i)the use of ever faster computers of ever increasing memory with (ii) improved computational algorithms. Based on a mathematical/computational trick described in [6] our team developed algorithms of lower complexity scaling or lower growth of CPU load first in the computation of optical spectra in time dependent density functional response theory [7] (ANR NOSSI, 2008-2010), and then in the solution of Hedin’s GW equations for the electronic self energy in finite systems [8]. In both cases the number of computational operations was reduced from O(N4)to O(N3),where N is the number of atoms. Our reduction of complexity scaling was achieved for molecules and finite clusters, but meanwhile we found an analogous algorithm for solving Hedin’s GW equations for periodic systems that should again scale as N3, with N now representing the size of the unit cell and we are presently converting this algorithm into a computer code. This will pave the way for an extension of Hedin’s GW theory from inorganic semiconductors to the more complex organic semiconductors where the number of atoms N in the unit cell N is indeed very large. The computation of (excitonic) optical spectra involves solving the Bethe–Salpeter (BS) equation for the propagation of an electron-hole pair, an operation that is known to scale as N3 when using a basis of atomic orbitals. Therefore the computation of the GW self energy was indeed the bottleneck in electronic structure computations. Computationally cheap GW theory will be useful in a wide variety of electronic structure problems that involve excitonic phenomena, such as molecular electronics, for example.
The practical organization of the ORGAVOLT project This project will deal, over a period of 3 years, with the band structure of bulk phases of organic semiconductors, with their heterojunctions and with the physical mechanism of photo current generation. The last topic will benefit from the expertise in non equilibrium phenomena and Keldysh type techniques of Rémi Avriller, a member of our team at Bordeaux. The photo current generation problem is closely related of the Bethe-Salpeter equation which resembles the celebrated Casida Equations [1] of Mark E. Casida, our team member at Grenoble. Mark E. Casida will indeed supervise, in the framework of the ORGAVOLT project, a thesis on the mechanism of the photocurrent generation at organic heterointerfaces, see [1]. Our methods can be used with any LCAO codes, but we use the SIESTA [9] local orbital code, of which one of our team members, Daniel SanchezPortal at San Sebastian, is a founding author.
Strong interaction between the teams at San Sebastian and Bordeaux is possible thanks to the relative proximity of these two cities that are within only 2 1/2 hours driving of each other. The collaboration of both teams with the team of Mark E. Casida at Grenoble will be greatly helped by the existence of low cost air travel between Bordeaux and Lyon. The University of Bordeaux plays a special role in our project as the laboratory of Georges Hadziioannou (LPCO)http://www.lcpo.fr/wordpress/ at the University of Bordeaux is a leading international center for the synthesis of molecules for organic solar cells. The materials that are synthesized in this lab are then investigated and turned into solar devices in the lab of Lionel Hirsch (IMS), also at Bordeaux. Clearly, the availability of newly synthesized materials and fresh, unpublished spectroscopic data is a tremendous challenge to our project. Conversely, the theoretical prediction, before synthesis and from first principles, of key properties of organic semiconductors and of the photo voltages of their interfaces will help to screen hypothetical organic semiconductors before their actual synthesis. This should help to speed up the optimization of organic solar cells and their use in solar energy.
Conclusion The ORGAVOLT project provides an exceptional opportunity to outstanding and highly motivated candidates to contribute decisively to the challenging theory of excitonic materials, while using and developing an innovative computational algorithm of improved complexity. The candidate will interact strongly with the teams at San Sebastian and at Grenoble and his/her work should help in clarifying the mechanisms of photo current generation in organic BHJ solar cells. In interaction with synthetic chemists and materials scientists at the University of Bordeaux, the candidate will confront theoretical predictions with experimental reality and contribute to the optimization of organic BHJ solar cells, an original and exciting technology of great promise. Some references 1 https://sites.google.com/site/markcasida/home/project-orgavolt and https://sites.google.com/site/markcasida/home/. 2 This graph is from “Wikipedia commons” and it is referred to in wikipedia/Solar_energy. 3 G.Yu, J.Gao, J.C.Hummelen, F.Wudl, A.J.Heeger, Science 270,1789(1995). 4 For example: C.Deibel and V.Dyakonov, Rep.Prog.Phys.73, 96401 (2010). 5 In-Wook Hwang, Daniel Moses, and Alan J. Heeger, Photoinduced Carrier Generation in P3HT/PCBM Bulk Heterojunction Materials, J.Phys.Chem.C 112, 4350(2008). 6 D. Foerster, J. Chem. Phys.128, 034108(2008). 7 Koval, D.Foerster, O.Coulaud, J. Chem. Theory Comput. 6, 2654.
8 D.Foerster, P.Koval, D.Sanchez-Portal, J. Chem. Phys. 135, 74105 (2011).
9 The Siesta method for ab-initio order-N materials simulations, J. Soler, E. Artacho, J.Gale, A. García, J.Junquera, P.Ordejón and D.Sánchez – Portal, J. Phys.: Condensed Matter 14, 2745 (2002)
Below: a choice of electronic spectral densities of molecules and clusters in the GW approximation that were computed on a desktop computer by Peter Koval.
