The world will need access to 30TW of power without releasing carbon into the atmosphere.
Currently the world consumes an average of 13 terawatts (TW) of power. By the year 2050, as the population increases and the standard of living in developing countries improves, this amount is likely to increase to 30 TW. If this power is provided by burning fossil fuels, the concentration of carbon dioxide in the atmosphere will more than double, causing substantial global warming, along with many other undesirable consequences. Therefore, one of the most important challenges facing engineers is finding a way to provide the world with 30 TW of power without releasing carbon into the atmosphere. Although it is possible that this could be done by using carbon sequestration along with fossil fuels or by greatly expanding nuclear power plants, it is clearly desirable that we develop renewable sources of energy. The sun deposits 120,000 TW of radiation on the surface of the earth, so there is clearly enough power available if an efficient means of harvesting solar energy can be developed.
Only a very small fraction of power today is generated by solar cells, which convert solar energy into electricity, because they are too expensive (Lewis and Crabtree, 2005). More than 95 percent of the solar cells in use today are made of crystalline silicon (c-Si). The efficiency of the most common panels is approximately 10 percent, and the cost is $350/m2. In other words, the cost of the panels is $3.50/W of electricity produced in peak sunlight. When you add in the cost of installation, panel support, wiring, and DC to AC converters, the price rises to approximately $6/W. Over the lifetime of a panel (approximately 30 years), the average cost of the electricity generated is $0.3/kW-hr. By comparison, in most parts of the United States, electricity costs about $0.06/kW-hr. Thus, it costs approximately five times as much for electricity from solar cells. If the cost of producing solar cells could be reduced by a factor of 10, solar energy would be not only environmentally favorable, but also economically favorable.
Although c-Si solar cells will naturally become cheaper as economies of scale are realized, dicing and polishing wafers will always be somewhat expensive. Thus, it is desirable that we find a cheaper way to make solar cells. The ideal method of manufacturing would be depositing patterned electrodes and semiconductors on rolls of plastic or metal in roll-to-roll coating machines, similar to those used to make photographic film or newspapers. Solar cells made this way would not only be cheaper, but could also be directly incorporated into roofing materials, thus reducing installation costs. Organic semiconductors that can be dissolved in common solvents and sprayed or printed onto substrates are very promising candidates for this application.
Because organic semiconductors have different bonding systems from conventional, inorganic semiconductors, they operate in a fundamentally different way. Conventional semiconductors are held together by strong covalent bonds that extend three-dimensionally, resulting in electronic bands that give rise to its semiconducting properties. Organic materials have similar intramolecular covalent bonds but are held together only by weak intermolecular van der Waals interactions. The electronic wave function is thus strongly localized to individual molecules, and the weak intermolecular interactions instigate a narrow electronic bandwidth formed in molecular solids.
The semiconducting nature of organic semiconductors arises from the p electron bonds that exist when molecules are fully conjugated (i.e., have alternating single and double bonds). The weakly held p electrons are responsible for all interesting optical and electronic transitions in organic semiconductors. The p to p* transitions in organic semiconductors are typically in the range of 1.4–2.5 eV, which overlaps well with the solar spectrum and makes them very promising candidates as active light absorbers in solar cells. A few examples of organic semiconductors used in solar cells are shown in Figure 1 (see PDF version).
The main difference between organic semiconductors and inorganic semiconductors as photovoltaic materials is that optical excitations of organic semiconductors create bound electron-hole pairs (called excitons) that are not effectively split by the electric field (Gregg, 2003). To separate the bound electrons and holes, there must be a driving force to overcome the exciton-binding energy, typically 0.1–0.4 eV. Excitons in organic semiconductors that are not split eventually recombine either radiatively or nonradiatively, thereby reducing the quantum efficiency of a solar cell.
In inorganic semiconductors the attraction between an electron-hole pair is less than the thermal energy kT. Therefore, no additional driving force is required to generate separated carriers. Research has shown that excitons in organic semiconductors can be efficiently split at a heterojunction of two materials with dissimilar electron affinities or ionization potentials.
The narrow electronic bandwidth in organic semiconductors has a few consequences. First, the absorption-spectrum bandwidth is narrower than in conventional inorganic semiconductors. Consequently, a single organic material can be potentially photoactive only in a narrow optical-wavelength range of the solar spectrum (Figure 2). Although this is a disadvantage in terms of harvesting solar flux, multiple absorbers in stacks of solar cells connected in series can be engineered to expand the absorption range. Because the valence band and conduction band are concentrated in narrower energy regions, the absorption coefficient resulting from the excitation of electrons from the valence band to the conduction band is very strong, typically >10–5 cm–1 at peak absorption. This high absorption coefficient means that only a thin (100–200 nm) film is required to absorb most incident light, an attractive characteristic for solar cells because less material is required to make them.
Second, the charge carriers in organic semiconductors do not exhibit band-like transport as they do in inorganic semiconductors. Instead, they move around by a hopping mechanism between localized states. The charge-carrier mobilites in organic semiconductors are, therefore, inherently low, with typical values of <10–2 cm2/Vs. This low charge-carrier mobility puts a constraint on the thickness of organic materials that can be used in a solar cell because recombinative loss increases with increasing thickness. Fortunately, this drawback is offset because only a very thin layer of organic materials is necessary because organic semiconductors are highly absorptive. Organic solar cells may potentially perform better than conventional solar cells at higher temperature, because hopping is a thermally activated process. The performance of inorganic solar cells typically decreases as operating temperature increases.
A third key difference between organic and inorganic semiconductors is that organic materials do not have dangling bonds at surfaces. Therefore, organic-organic junctions or organic-metal junctions in organic solar cells (interface states) do not act as potential charge-carrier recombination sites.
Production of Heterojunction Devices
The simplest organic solar cells can be made by sandwiching thin films of organic semiconductors between two electrodes with different work functions. The work function is the amount of energy necessary to pull an electron from a material. When such a diode is made, electrons from the low-work-function metal flow to the high-work-function metal until the Fermi levels are equalized throughout the structure. This sets up a built-in electric field in the semiconductor. When the organic semiconductor absorbs light, electrons are created in the conduction band, and holes (positive-charge carriers) are created in the valence band. Thus, in principle, the built-in electric field can pull the photogenerated electrons to the low-work-function electrode and holes to the high-work-function electrode, thereby generating a current and voltage (Figure 3a). In practice, however, these cells have very low power-conversion efficiency (< 0.1 percent) because the electric field is not strong enough to separate the bound excitons (i.e., the excited-state species formed in organic semiconductors described above).
A significant improvement in the performance of organic solar cells was achieved by Tang (1986). His device consisted of a heterojunction between donor and acceptor semiconductors, resembling a p-n junction in conventional solar cells (Figure 3b). The benefit of this device derived from the use of two organic materials with offset electron affinities (lowest unoccupied molecular orbital, LUMO) or ionization potentials (highest occupied molecular orbital, HOMO). Excitons that diffuse to the interface undergo efficient charge transfer, as this offset in the energy levels provides a sufficient chemical potential energy to overcome the intrinsic exciton-binding energy. Upon charge transfer, the electrons are transported in the acceptor material and the holes in the donor material to their respective electrodes.
The efficiency of this type of planar heterojunction device is limited, however, by the exciton diffusion length, which is the distance over which excitons travel before undergoing recombination, approximately 5–10 nm in most organic semiconductors. Excitons formed at a location further than 5–10 nm from the heterojunction cannot be harvested. The active area of this type of solar cell is thus limited to a very thin region close to the interface, which is not enough to adsorb most of the solar radiation flux.
In the mid 1990s, Yu and colleagues (1995) showed that excitons can be rapidly split by electron transfer before the electron and hole recombine if carbon-60 (C60) derivatives are blended into the polymer (Figure 3c). Blend solar cells were made simply by blending the C60 derivative, which acts as an electron acceptor, into the polymer at concentrations in the range of 18–80 wt. percent. At these concentrations, the polymer and the C60 derivatives form a connected network to each electrode. The key to making efficient blend solar cells is to ensure that the two materials are intermixed very closely at a length scale less than the exciton diffusion length so that every exciton formed in the polymer can reach an interface with C60 to undergo charge transfer.
At the same time, the film morphology has to enable charge-carrier transport in the two different phases to minimize recombination. The film morphology (i.e., phase separation between the two materials) and, ultimately, the efficiency of the device, are determined by the concentration of materials, film-casting solvent, annealing time, temperature, and other parameters. Solar cells made by this method have continuously improved to better than 2 percent power efficiency under solar AM 1.5 conditions over the last few years (Padinger et al., 2003; Shaheen et al., 2001), and recently, an efficiency of 5 percent was reported (Ma et al., 2005).
The work on polymer/C60-derivative blend cells has created a new paradigm in the field of organic-based solar cells, which is the notion of bulk heterojunction devices, wherein two semiconductors with offset energy levels are interpenetrated at a very small length scale to create a high interface area for achieving high-efficiency devices. Since then, similar bulk heterojunction devices using electron acceptors other than the C60 derivative, such as CdSe nanorods (Huynh et al., 2002), a second semiconducting polymer (Granstrom et al., 1998), and titania nanocrystals (Arango et al., 1999), have been demonstrated, albeit with slightly lower efficiencies.
Limits on Performance
To understand the limits on the performance of bulk heterojunction devices and find ways to improve them, it is important to consider all of the processes that must occur inside the cells for electricity to be generated. These processes, shown in Figure 4a, are: (1) light absorption; (2) exciton transport to the interface between the two semiconductors; (3) forward electron transfer; and (4) charge transport. One must also consider undesirable recombination processes that can limit the performance of the cell, such as geminate recombination of electrons and holes in the polymer and back electron transfer from the electron acceptor to the polymer (Figure 4b).
The necessity of absorbing most of the solar spectrum (process 1) creates two requirements. First, the band gap must be small enough to enable the polymer to absorb most of the light in the solar spectrum. Calculations to determine the band gap that optimizes the amount of light that can be absorbed and the voltage that can be generated show that the ideal band gap is approximately 1.5 eV, depending on the combination of semiconducting polymers and electron acceptors (Coakley and McGehee, 2004). Second, the film must be thick enough to absorb most of the light. For most organic semiconductors, this means that films must be 150–300 nm thick, depending on how much of the film consists of a nonabsorbing electron acceptor. The optimum film thickness will absorb much incident light without significant recombination losses.
Once an exciton is created in the polymer, it must diffuse or travel by resonance energy transfer (process 2) to the interface with the other semiconductor and be split by electron transfer before it recombines (process 5). Experiments have shown that an exciton can diffuse approximately 5–10 nm in most semiconducting polymers before recombination. Therefore, no regions in the polymer can be more than 5–10 nm from an interface. Templating or nanostructuring of the donor and acceptor phases to fabricate ordered bulk heterojunction with controlled dimensions is an attractive approach to achieving full exciton harvesting (Figure 5) (Coakley and McGehee, 2003). Some small-molecule semiconductors have been shown to have larger exciton diffusion lengths (Peumans et al., 2003).
Research is under way to improve exciton transport in organic semiconductors, for example by using resonance energy transfer to funnel excitons directly to an absorber located at the charge-splitting interface or by incorporating phosphorescent semiconductors, which exhibit longer excited-state lifetimes (Liu et al., 2005; Shao and Yang, 2005).
Forward Electron Transfer
The actual process of charge transfer (process 3) requires that the offset in LUMO levels of the donor and acceptor be sufficient to overcome the exciton-binding energy. However, this drop in energy must not be excessive, because the maximum voltage attainable from this type of bulk heterojunction solar cell is determined by the gap between the HOMO of the electron donor and the LUMO of the acceptor. The gap becomes smaller as the LUMO of electron acceptors is moved farther away from the LUMO of the polymer, which corresponds to a larger driving force for charge transfer.
As can be seen from processes 1, 2, and 3, the design of an efficient organic solar cell involves optimizing the various energy levels to achieve the optimum level of extracted current with respectable voltage, as the power supplied by a solar cell is the product of current and voltage. Fortunately, the wealth of chemical synthetic knowledge and the dependence of electronic properties of organic molecules on their molecular structures allow for flexible tuning of the band gap and energy levels of organic semiconductors by chemical synthesis. Significant research on band engineering of this type should yield very promising results in the near future.
After forward electron transfer, the holes in the polymer and the electrons in the electron acceptor must reach the electrodes (process 4) before the electrons in the acceptor undergo back electron transfer to the polymer (process 6). Even in the best bulk heterojunction cells, this competition limits the efficiency of the cells.
The problem can usually be mitigated by making cells that are only 100-nm thick so that the carriers do not have to travel very far. Unfortunately, most of the light is not absorbed by films this thin. If the films are thick enough to absorb most of the light, then only a small fraction of the carriers escape the device. Many researchers are now trying to optimize the interface between the two semiconductors and improve charge transport in the films so that the charge can be extracted from 300-nm-thick films before recombination occurs.
The outlook for organic solar cells is very bright. Efficiency greater than 5 percent has been achieved (Ma et al., 2005; Xue et al., 2004), and many are optimistic that 20 percent can be achieved by optimizing the processes described above. Once this goal is achieved, a primary research challenge will be making cells that are stable in sunlight and that can handle wide temperature swings. The survival of many organic pigments in car paints in sunlight and the production of organic light-emitting diodes with operational lifetimes greater than 50,000 hours are encouraging signs that the required stability can be achieved. The final challenge will be scaling up the process and manufacturing the cells at a cost of approximately $30/m2. Shaheen and colleagues (2005) have described several approaches to making organic cells.
Arango, A.C., S.A. Carter, and P.J. Brock. 1999. Charge transfer in photovoltaics consisting of polymer and TiO2 nanoparticles. Applied Physics Letters 74(12): 1698–1700.
Coakley, K.M., and M.D. McGehee. 2003. Photovoltaic cells made from conjugated polymers infiltrated into mesoporous titania. Applied Physics Letters 83(16): 3380–3382.
Coakley, K.M., and M.D. McGehee. 2004. Conjugated polymer photovoltaic cells. Chemistry of Materials 16(23): 4533–4542.
Granstrom, M., K. Petritsch, A. Arias, A. Lux, M. Andersson, and R. Friend. 1998. Laminated fabrication of polymeric photovoltaic diodes. Nature 395(6699): 257–260.
Gregg, B. 2003. Excitonic solar cells. Journal of Physical Chemistry B 107(20): 4688–4698.
Huynh, W.U., J.J. Dittmer, and A.P. Alivisatos. 2002. Hybrid nanorod-polymer solar cells. Science 295(5564): 2425–2427.
Lewis, N.S., and G. Crabtree. 2005. Basic research needs for solar energy utilization. Available online at: http://www.sc.doe.gov/bes/reports/files/SEU_rpt.pdf.
Liu, Y., M.A. Summers, C. Edder, J.M.J. Fr?chet, and M.D. McGehee. 2005. Using resonance energy transfer to improve exciton harvesting in organic-inorganic hybrid photovoltaic cells. Advanced Materials (in press).
Ma, W., C. Yang, X. Gong, K. Lee, and A.J. Heeger. 2005. Thermally stable, efficient polymer solar cells with nanoscale control of the interpenetrating network morphology. Advanced Functional Materials 15(10): 1617–1622.
Padinger, F., R.S. Rittberger, and N.S. Sariciftci. 2003. Effects of postproduction treatment on plastic solar cells. Advanced Functional Materials 13(1): 85–88.
Peumans, P., A. Yakimov, and S.R. Forrest. 2003. Small molecular weight organic thin-film photodetectors and solar cells. Journal of Applied Physics 93(7): 3693–3723.
Shaheen, S.E., C.J. Brabec, N.S. Sariciftci, F. Padinger, T. Fromherz, and J.C. Hummelen. 2001. 2.5% efficient organic plastic solar cells. Applied Physics Letters 78(6): 841–843.
Shaheen, S.E., D.S. Ginley, and G.E. Jabbour, editors. 2005. Organic Photovoltaics. MRS Bulletin 30(1). Special Edition.
Shao, Y., and Y. Yang. 2005. Efficient organic heterojunction photovoltaic cells based on triplet materials. Advanced Materials (in press).
Tang, C.W. 1986. Two-layer organic photovoltaic cell. Applied Physics Letters 48(2): 183–185.
Xue, J.G., S. Uchida, B.P. Rand, and S.R. Forrest. 2004. Asymmetric tandem organic photovoltaic cells with hybrid planar-mixed molecular heterojunctions. Applied Physics Letters 85(23): 5757–5759.
Yu, G., J. Gao, J.C. Hummelen, F. Wudl, and A.J. Heeger. 1995. Polymer photovoltaic cells: enhanced efficiencies via a network of internal donor-acceptor heterojunctions. Science 270(5243): 1789–1791.