Assessment of the dye-sensitized solar cell

Assessment of the dye-sensitized solar cell

(Parte 1 de 3)

Renewable and Sustainable Energy Reviews 6 (2002) 273–295

Assessment of the dye-sensitized solar cell

R.D. McConnell *

Center for Basic Sciences, National Renewable Energy Laboratory, M/S 3211, 1617 Cole Boulevard, Golden, CO 80401, USA

Received 12 September 2001; accepted 12 September 2001


The field of solar electricity, or photovoltaics (PV), is rich in that there are many materials and concepts for converting sunlight into electricity. The technologies accepted as conventional are those well along in the process of commercialization. The dye-sensitized solar cell, developed in the 1990s, is a nonconventional solar electric technology that has attracted much attention, perhaps a result of its record cell efficiency above 10%. This paper reviews the technology, discusses new research results and approaches presented at a recent symposium of many of the world’s important dye solar cell researchers, and presents an assessment of the dye-sensitized solar cell in a comparison with current conventional solar electric technologies. It concludes the dye solar cell has potential for becoming a cost-effective means for producing electricity, capable of competing with available solar electric technologies and, eventually, with today’s conventional power technologies. But it is a relatively new technology and faces many hurdles on the path to commercialization. Because of its potential, this assessment recommends further funding for research and development (R&D) of the dye-sensitized solar cell technology on the basis of the promising technical characteristics of the technology, a strong US and worldwide research base, positive industry interest, and today’s relatively small funding allocation for its R&D. 2002 Elsevier Science Ltd. All rights reserved.

1. Introduction274
2. Assessment perspective275


* Tel.: +1-303-384-6419; fax: +1-303-384-6481. E-mail address: (R.D. McConnell).

274 R.D. McConnell / Renewable and Sustainable Energy Reviews 6 (2002) 273–295

excitonic, nanocrystalline solar electric technology276
4. Recent dye cell research279
4.1. Ecole Polytechnique Federale de Lausanne (EPFL)280
4.2. University of Bath281
4.3. Emory University281
4.4. ChemMotif, Inc282
4.5. National Renewable Energy Laboratory (NREL)282
4.6. California Institute of Technology283
4.7. Imperial College284
4.8. Sustainable Technologies International Pty Ltd (STI)284
4.9. Johns Hopkins University284
4.10. A ngstrom Solar Center286
5. Beyond the Horizon Awards286
5.1. California Institute of Technology286
5.2. Johns Hopkins University287
5.3. DuPont287
6. An assessment framework for the dye cell288
7. Discussion of the dye cell assessment291
8. Government funding supporting dye cell research and development292
9. Conclusions293

3. The dye-sensitized solar cell: a nonconventional, organometallic, biomimetic,

1. Introduction

The dye-sensitized solar cell is a nonconventional solar electric technology that gained the attention of the photovoltaic community with the appearance of a 1991 publication in Nature [1]. The dye cell operates quite differently from conventional solar cells, yet researchers have already demonstrated efficiencies greater than 10%. Its foundations are in photochemistry rather than in solid state physics, the discipline underlying today’s conventional solar cells. Because of the cell’s progress and subsequent interest shown worldwide by solar cell researchers and developers, it is important to conduct an assessment of the dye-sensitized solar cell’s competitiveness with conventional solar electric technologies.

Historically, solar cell operation was first discovered in a photochemical cell. The

French scientist Becquerel, who discovered the solar electric effect in the 1830s, referred to his device as a ‘pile’ or ‘cell’ that produced electric current when exposed to solar radiation [2]. Becquerel’s solar cell was an electrolytic cell made up of two electrodes placed in an electrolyte, in which the current increased when the cell was exposed to light. Becquerel distinguished the behavior of this solar cell from that of a thermocouple (pile thermoelectrique) in that the solar radiation and not solar heating produced the cell’s direct current. Like batteries, solar cells produce direct electric current (DC) that needs converting to alternating electric current (AC). Solar cell is still a good description of the solar electric devices made in the research laboratory, and solar electricity is an easily understandable term for the technology of converting sunlight into electricity.

However, photovoltaics, or PV, is the term used for many years by technologists in the field. In either case, PV or solar electricity is one of the renewable energy technologies because the sun is a nondepletable resource. Yet PV is not the best name for the technology, especially when describing it to the general public. There are even some technical inaccuracies with the term. While photo refers to light, volt ignores the fact that amperes are also part of electricity. Strictly speaking, it is the sunlight that produces the amperes of electric current. Also, power in watts is the product of voltage and amperes, so solar power is technically very accurate. Most of this paper will refer to solar electricity, solar electric technologies, solar electric power,o r solar power rather than PV.

2. Assessment perspective

The assessment of a solar electric technology can be made from different perspectives. Homeowners or business owners may consider purchasing backup electric systems in case their electric utility cannot provide power when they need it. In comparing different backup technologies, the users’ assessment may be based on how much electric power they can get for their money, probably measured in $/watt of electric power. Users might then estimate how much fuel will cost in the future for fossilfueled systems and how much maintenance costs will be. Unless the assessment is done carefully by acknowledging the volatility of future fossil fuel prices and other value attributes, solar electric technologies still cost too much when compared with fossil-fired electric generation technologies. All solar electric companies are well aware of these assessments made by potential buyers. But there is an undeniable trend in that over the long term, solar electric costs have declined and conventional energy costs have risen. This trend helps drive the remarkably consistent 20% or greater annual market growth for solar electric technologies [3].

Investors considering expanding their portfolio to include solar electric manufacturing companies will perform a different assessment. They may look at how much production capacity they can get for their money. In other words, they might evaluate their investment in terms of investment dollars needed per production capacity. For example, a solar cell manufacturing plant capable of producing 100 MW per year might cost $50 million, $100 million, or $300 million depending on the technology requirements. The measure here is also $/watt, but it is watts of production capacity. Investors also may want to know how much money they will make in a reasonably short time period and how much risk is involved. There are other factors as well, such as the strength of the management and technical teams the company employs and the clarity of their market strategy. Energy investors typically make their decision in comparison with other potential investments in wind, bioenergy, hydroelectric, or nonrenewable energy companies.

A country or government may conduct yet another type of assessment. This is because it is not concerned with making money and because its citizens typically have long-term interests that they expect their government to accommodate. A government’s assessment may involve some of the issues a purchaser and investor consider, as well as others, such as societal benefits, environmental impact, and the country’s scientific strength and industrial interest in developing solar electric technologies. Governments need to know if investing the taxpayer’s money in the technology’s development will provide benefits to the country and its citizens. As a technology assessment, it differs from a policy assessment, in which a government might consider the benefits of a policy or tax reduction to expand the market growth of a technology that provides strategic benefits to a country, such as a reduction in reliance on foreign energy sources.

The technology assessment presented here is made in response to a milestone in the US Department of Energy’s (DOE) Five-Year Research Plan to assess the dyesensitized solar cell [4]. The assessment relies heavily on portions of a technical assessment conducted by Novem, a Dutch research agency, for several solar cell technologies. The dye-sensitized solar cell, or more simply the dye solar cell,i s a nonconventional solar electric technology not presently part of the world’s PV market. About 280 MW of solar electric modules were sold in 2000 throughout the world, more than 90% based on crystalline silicon solar cell technology, which is a conventional solar electric technology. The technology assessment made in this report is meant to help government decision makers determine whether or not more funding should be allocated to developing the dye solar cell. Some elements of the assessment may be useful to potential investors, but it will be of little use for the individual considering the purchase of a solar electric system for his/her own needs.

3. The dye-sensitized solar cell: a nonconventional, organometallic, biomimetic, excitonic, nanocrystalline solar electric technology

Researchers had explored the concept of the dye solar cell long before the 1991 breakthrough publication in Nature [1,5]. The significance of this Swiss research was the discovery of a means for making the concept work, in particular attaching the organic dye to the inorganic titanium dioxide, which permitted the efficient transfer of photoexcited electrons from the dye. The conventional solar cell, first discovered in 1954 [6], is a solid wafer-like semiconductor structure in which sunlight is absorbed, creating positive and negative electric charge carriers. The charge carriers are swept from the structure using an internal electric field. Fig. 1 shows two contacts to the solar cell structure, again reminding us of the similarity with battery cells, that conduct the solar-generated electric current to an external electrical load. The dye solar cell has a different structure and operates quite differently.

The dye solar cell, unlike the conventional solar cell, physically separates the absorption and charge-transport processes. The dye cell mimics photosynthesis in

Fig. 1. The conventional solar cell is a solid wafer-like semiconductor structure in which sunlight is absorbed, creating positive and negative electric charge carriers that are swept from the structure by an internal electric field.

that the organometallic dye absorbs sunlight, creating an exciton whose electron is injected extremely rapidly into another medium, a porous matrix of titanium dioxide. In fact, the dye chemistry for this solar cell arose originally from biomimetic considerations based on the model of natural chlorophyll, an organometallic molecule [7]. The other medium consists of a porous structure of titanium dioxide nanocrystals, seen in Fig. 2, that are 10 nm or so in size, and all bonded together, typically through a sintering process. Fig. 3 shows a schematic of the dye solar cell. The dye is attached as a monomolecular layer to the nanoparticle surfaces in the porous matrix. Once the titanium dioxide matrix receives an electron from the dye, the electron diffuses through the matrix to the cell contacts and into an external circuit. Unlike a conven-

Fig. 2. A porous structure of titanium dioxide nanocrystals that are about 10 nm in diameter, which are bonded together typically through a sintering process.

Fig. 3. Schematic of the dye solar cell.

tional solar cell, there is no internal electric field sweeping electric charges to the solar cell contacts. Completing the dye cell circuit requires providing an electron to the dye to replace the one injected into the matrix. The oxidized dye gains an electron by itself oxidizing a mediator — a redox species dissolved in an electrolyte also filling the spaces in the porous matrix. The mediator is, in turn, reduced at the metallic cathode of the solar cell.

In summary, the conventional solar cell is a solid, wafer-like, inorganic semiconductor device in which the minority carriers are critical to the device’s operation. Also critical to the operation of a conventional solar cell is an internal electric field created by a homojunction (in the case of crystalline or amorphous silicon) or heterojunction (in the case of copper indium gallium diselenide, cadmium telluride, or GaAs-based materials). The dye solar cell is a nonconventional solar cell with a porous inorganic matrix incorporating an organometallic dye and a liquid electrolyte within the porous spaces of the matrix. It is quite thin, more than 10 times thinner than a crystalline silicon solar cell. It is a biomimetic device in that the dye mimics essential aspects of photosynthesis. It is a nanoparticle device in that the matrix is made from nominal 10-nm diameter particles. It does not incorporate an internal electric field (although future versions might), and it is a majority-carrier device because the electrons are injected into n-type titania. Finally, positive charge transport occurs through ion transport in the electrolyte, rather than hole conduction. Gregg [8] has written a more extensive discussion of the differences between conventional inorganic and nonconventional organic solar cells, including the dye cell. The dye solar cell is a dramatic example of a nonconventional solar cell; yet, its record efficiency has been confirmed at the National Renewable Energy Laboratory (NREL) at 10.4%, considered to be above the significant threshold efficiency for any new solar electric technology [7]. As we shall see in following paragraphs, several research avenues can increase its efficiency.

One fundamental difference between the operation of conventional and dye solar cells can be illustrated using the following equation from Gregg’s discussion of the operational differences between conventional and nonconventional solar cells [8] where the electron current density, Jn, is proportional to the gradient of the conduc- tion band edge, Ecb, and the electron concentration gradient, n, in turn proportional to the chemical potential gradient. The remaining terms are the electron mobility, mn, the electric charge, q, and the diffusion coefficient, Dn. In conventional solar cells, Ecb 0 due to a p–n junction, whereas in the case of no band bending, a solar electric effect can still be achieved if light absorption results in an electron concen- tration gradient. This is the case for current flow in the titanium dioxide matrix of the nonconventional dye-sensitized solar cell. Another important difference is the means of separating electrons and holes. The only way to separate free electrons and holes in a conventional solar cell is with an electric field. But in an organic solar cell, the more effective means of separation of electrons and holes is the result of the dissociation of excitons — electron–hole bound states created by light absorption. Exciton dissociation can occur at an organic/organic or an organic/inorganic interface, such as the dye/TiO2 interface in the dye cell. It is the mechanism for charge separation in the dye cell. Dissociation of the exciton, electron injection into the titania, and subsequent electric current in an electric-field-free titanium dioxide matrix is a fundamentally different mechanism than that of a conventional solar cell.

4. Recent dye cell research

An Electrochemical Society (ECS) symposium, Photovoltaics for the 21st Century,

I, convened in March 2000 in Washington, DC. The symposium highlighted the research from more than a dozen of the world’s eminent researchers and developers of the dye-sensitized solar cell. One aspect of a technology assessment from a government program perspective is identifying research groups that have been contributing to the technology’s progress. Another aspect is the potential for improvement through additional research. The ECS symposium, cosponsored by DOE and NREL, provided a snapshot of the important dye solar cell research under way throughout the world. The researchers invited to the dye-cell session of the ECS symposium were selected from more than 70 dye solar cell presenters at the Thirteenth International Conference on Photochemical Conversion and Storage of Solar Energy held in Snowmass, Colorado, in August 2000. The dye-sensitized solar cell presentations comprised a significant portion of more than 300 presentations at the conference. The following discussion of research results at the ECS symposium is presented by group rather than by technology issues.

280 R.D. McConnell / Renewable and Sustainable Energy Reviews 6 (2002) 273–295 4.1. Ecole Polytechnique Federale de Lausanne (EPFL)

EPFL, located in Lausanne, Switzerland, is the best-known of the research groups at the symposium. It is led by Michael Gratzel and has about 20 researchers. EPFL has a number of funding sources, including the United States Air Force Research Laboratory. As mentioned earlier, Gratzel and O’Regan published the breakthrough dye-sensitized solar cell paper in 1991 [1]. Their early work arose out of an interest in artificial photosynthesis or biomimetics. The prototype of an energy-absorbing dye provided by nature is chlorophyll, a molecule consisting of a central magnesium atom surrounded by a nitrogen-containing porphyrin ring. Emphasizing the significance of the chlorophyll structure in nature is a similar structure in blood — hemoglobin — the oxygen-carrying molecule containing iron. Although nature confines itself to magnesium and iron for its principal pigments, other metallic elements can be incorporated into synthetic porphyrin dyes. Clark and Suttin used a tripyridyl ruthenium complex in 1977 to sensitize titanium dioxide, but the dye was in solution and charge transfer through the solution was very inefficient [7]. In 1980, the idea had emerged of bonding the dye to the metal oxide surface through an acid carboxylate group [7]. This bonding facilitated charge transfer by electron injection. It is now known that an electron passes from the dye molecule through the bridging carboxylate group to the semiconductor substrate within picoseconds, faster than competing recombination processes. Of the hundreds of dye molecules explored over the past decade, one of the most effective is RuL2(NCS)2 (L=bipyridyl), often called ‘N3’, as it was the third dye tried by Nazeeruddin at EPFL.

EPFL holds the record for the highest dye cell efficiency (10.4%), confirmed at NREL under air-mass (AM) 1.5 sunlight [7]. It is also responsible for the develop- ment of the state-of-the-art black dye in which L2, the bis-bi-pyridyl ligand system, is reduced to ter-pyridyl and a thiocyanate group is added. The black dye’s spectral sensitivity extends throughout the visible and into the infrared ranges, approaching the ideal absorption edge position (1.4 eV) for optimal solar energy conversion. The bonding of the Ti ion to the dye’s carboxylate group has also been optimized by adding a suitable salt, permitting higher temperature thermal treatments to 300°C during manufacturing for subsequent bonding, sealing, and curing processes.

EPFL’s recent work explores an innovative configuration in which the n-type nanocrystalline titanium dioxide is contacting an organic p-type semiconductor in what they have called a sensitized nanostructured heterojunction [7]. This innovative structure might enhance electron transport, and perhaps injection, because of the heterojunction’s electric field. And the organic semiconductor replaces the liquid electrolyte’s function as a positive charge conductor.

The group has licensed their technology to at least eight companies and thus has acquired an awareness of its commercialization potential. One important commercialization issue is the device’s lifetime. The EPFL group has investigated and contributed to the long-term stability of the device. With adequate cell sealing and the addition of suitable solvents in the electrolyte formulation, the system has been able to pass standard stability qualification tests for outdoor applications, including ther- mal stress for 1000 h at 85°C. The expected lifetime is presently beyond 10 years, again quite respectable for a relatively new solar electric technology.

4.2. University of Bath

The University of Bath research effort is directed by Lawrence Peter. Their presentation included a comprehensive summary of the unusual transport properties that distinguish dye cells from conventional solid-state devices [9]. For example, charge transport in the electrolyte phase takes place by ion transport, a relatively slow process leading to ionic diffusion transit times of the order of 0.1 s. This study, however, focuses on understanding electron charge transport, trapping, and recombination so that more efficient dye cells may be realized. The University of Bath group found that because of remarkably small electron diffusion coefficients, the diffusional transit time for electrons in a typical 10-µm thick titanium dioxide film is of the order of milliseconds at AM 1.5 illumination, but as slow as 100 s at light intensities some seven orders of magnitude less. As noted in their paper, it is surprising that an efficient solar cell can be based on a system that involves such slow charge transport. Also, there is an increase in electron lifetime with decreasing light intensity, which partially explains a reduction in cell efficiency that is much smaller than might be expected at low light intensities. From another point of view, the dye cell performs well in low light levels, which presents opportunities to develop indoor applications.

This study of photovoltage decay followed by short-circuit extraction of remaining electron charge permits the derivation of information about trapped electron charge and its corresponding trap distribution. Electrons are evidently trapped at levels in the bandgap, and the assumption is that these levels are associated with surface states present at the high internal surface area of the nanocrystalline oxide ( 100 m2/g). These levels are responsible for the retardation of electron transport that is characteristic of dye cells in that diffusion takes place through electron trapping and detrapping processes.

4.3. Emory University

Emory University’s group, headed by Professor Tianquan Lian, reported on their experimental studies of interfacial charge separation and charge recombination in two classes of sensitized nanocrystalline films [10]. Their work was funded by DOE’s Office of Science and the National Science Foundation (NSF). One class includes the dye-sensitized, wide-bandgap semiconductor, nanocrystalline films used in Gratzel solar cells. The second contains conjugated polymer/nanocrystalline thin-film composites for completely solid-state cells (the MEH-PPV is also the hole transporter). They made measurements on their own samples of MEH-PPV/SnO2 and compared them with measurements on their dye-sensitized nanocrystalline thin films of TiO2, SnO2, and ZnO. Their experimental results show that for both systems, charge separation can be very rapid and efficient (some 300 ps), whereas charge recombination is much slower (microseconds). For example, for MEH-PPV/TiO2, the electron transfer is believed to be substantially faster than the rate of polymer exciton decay because transfer has been demonstrated and the polymer photoluminescence is quenched. The Emory technique, using femtosecond mid-infrared (IR) spectroscopy, leads to direct monitoring of injected electrons and substantiates this belief. The comparison is intriguing because they have shown that subpicosecond electron injection from dye to semiconductor is possible with a favorable electronic coupling of the electron-donating and -accepting orbitals and/or a large density of accepting states. But for the MEH-PPV nanocrystalline films, the electron injection is just as fast, yet there are no covalent linkages between MEH-PPV and the SnO2. The fast charge separation and slow recombination results, like those in dye-sensit- ized nanoporous films, show that conjugated polymer/inorganic semiconductor nanoporous film composites have promise for solar electricity production.

4.4. ChemMotif, Inc.

ChemMotif, a small business headed by Mark Spitler in Concord, Massachusetts, has been funded by the DOE’s Small Business Innovative Research program for research and development of dyes for the dye-sensitized solar cell. The company explored the use of different cyanine dyes and made devices with higher short-circuit photocurrents than those of cells using the well-known N3 dye [1]. The trade-off between broad-band absorption and extinction coefficient for a single dye like N3 leads to slightly thicker titanium dioxide films (10 µm) than could be achieved with aggregates of dyes needing only a 4-µm film thickness. High extinction coefficients for monomer dyes, such as rhodamine, oxazine, thiazine, and cyanine, are associated with very narrow absorption bandwidths. ChemMotif tried to circumvent this obstacle by using aggregates of organic dyes with high absorption over a greater portion of the spectrum. Their work focused on the use of different carboxylated cyanine dyes. They measured absorbance versus wavelength for 14 dyes and their combinations by exploring the structure of the tether of the dye to the titanium dioxide, the terminal heterocycle on the chromophore, the length of the methine bridge, and substituents on the molecule. Some dye pairs did not mix and appeared to collect in different regions on the nanocrystalline surface. They observed the degree of aggregation of the cyanine dyes on the surface to be highly dependent on the procedure for attaching the dyes to the titanium dioxide. An enlightening observation in their work is that some of their heterogeneous collections of dye molecules are used in photography in panchromatic films. This presentation hinted at the large number of dye molecules possibly beneficial to the dye solar cell concept and suggested that dye aggregates may be even more important.

4.5. National Renewable Energy Laboratory (NREL)

The NREL group, under the leadership of both Arthur Nozik and Arthur Frank, is probably the strongest in the US. NREL presented three papers at the ECS symposium, two on electron transport and one on a new family of dyes [12–14]. The NREL team recently measured a preliminary 9.18% efficiency for a dye cell using a titanium dioxide film deposited by screen printing, possibly the highest efficiency by any group for films prepared by screen printing. This group has been jointly funded by DOE’sO ffice of Science and Office of Energy Efficiency and Renewable Energy.

The NREL group has modeled electron transport using a random walk analysis [12]. In their introduction, they state that the reason transport of photoinjected elec- trons in the TiO2 occurs by diffusion is because of the absence of any electric field across the film due to screening by the electrolyte. The collection time for electrons is on the order of milliseconds, much slower than collection times in single-crystal titanium dioxide. The slow collection time is attributed to electrons undergoing many trapping–detrapping events in trap states believed to be on the surface, although the trap states could just as well be inside the nanoparticles. Another attribute of the dye cell system is the coupled motion of electrons and ions moving in opposite directions — ambipolar diffusion. A consequence of ambipolar diffusion is that a current can only be measured in the external circuit upon the arrival of charge carriers at the electrode; this is called an arrival current. Their modeling and corresponding experimental results support the concept of an arrival current, as well as multiple trap states with an exponential distribution [13]. This experimental agreement suggests that even shallow traps, owing to their relative abundance, can contribute significantly to the overall transport kinetics.

Another NREL study described a new generation of perylene-based dyes that may make excellent photosensitizers of nanocrystalline titanium dioxide [14]. Although the perylene dye cell efficiencies are low — about 1% — they are reported for the first time in this paper. Perylenes are inexpensive and durable dyes, which is important because the widely used N3 dye is presently quite expensive, more expensive per square meter than glass. Its high price is a consequence of its limited use only in the dye solar cell.

(Parte 1 de 3)