1.2 The Potential of Photovoltaics
Photovoltaics is able to convert sunlight directly into electricity and so allows to create an independent power supply. Solar cells can generate power in remote areas, because they do not rely on energy sources other than sunlight. In principle, the average sunshine over an area only as large as Wales (UK) would already be sufficient to supply the worlds electricity needs. Photovoltaics thus has a great potential to address the worlds energy requirement in the long run, in contrast to conventional resources, which are either finite or which generate pollutants. However, in order to compete with conventional energy technologies, solar cells still need to prove higher power conversion efficiencies and lower costs.
If I first consider what limits the conversion efficiency of sunlight into electricity, optical losses are identified as one of the main loss mechanisms, which will be the focus of this thesis. Thermal loss mechanisms, such as Joule heating, are another key factor affecting the power conversion efficiency. The need for good electrical transport properties does not tolerate material impurities or fabrication imperfections, hence electrical loss mechanisms are also important, especially as the cost-pressure in the solar industry makes it often too expensive for industry to use high purity materials.
While my work does not explicitly address the electrical loss mechanisms, I note that the optical light trapping techniques studied in this thesis allow us to use less material, thereby reducing the impact of material defects, as thinner films make it easier for charge carriers to be collected. I also note that a more efficient usage of silicon is not the cost driver any more, because the raw material now is only 20 % of the module cost and less than 10 % of the total cost of a photovoltaic system [].
As performance increases and manufacturing costs decline over time, the ratio of cost to performance, expressed in US Dollars per peak Watt generated by a module (US$/W
), is taken as a standardized figure-of-merit to quantify the improvements in photovoltaics. Figure ]. Grid parity means that the cost of solar electricity is equal to the cost of electricity from the national grid transmission network.
Fig. 1.1
The ratio of cost to generated power in US$/W
is used to quantify the improvement in photovoltaics. The graph illustrates the case of silicon wafer technology, which represents ca. 90 % of the global market share. The trend curve is indicated by a black line and highlights the rapid growth rate of the cumulative capacity worldwide: for each doubling of production volume, the average selling price of solar modules has fallen by 20 % []
If research and development in photovoltaics aims to follow this trend curve, neither high efficiency nor cost reduction is sufficient to displace the current technology. I cite two examples to illustrate this point.
High efficiency. Using a 1
m thick GaAs slab, the US based company Alta Devices set a world record efficiency of 29 %. Nevertheless, high efficiency is not enough, as the GaAs slab is more expensive to produce than comparable silicon cells (in terms of US$/W
). Hence, these world record cells are now not used in solar farms, but in consumer electronics instead, where surface area is premium. Alta Devices was recently sold to the Chinese company Hanergy, who intend to use this technology as a solar gadget [].
Cost reduction. Nanosolar, Abound Solar and Solyndra tried to commercialize low-cost solar modules. The companies were able to increase the throughput of their manufacturing plants by improving the speed of their deposition process, producing high quality CIGS and CdTe material very efficiently. When the modules were ready for the market, the prices for crystalline silicon modules had dropped in the meantime, and the loss of cost advantage ultimately forced the companies to shutdown [].
A new development in photovoltaics therefore needs to move below the trend curve of silicon wafers; in order to be competitive and to justify investments.
One way to achieve this is to reduce the material thickness, thereby reducing material cost, while maintaining cell efficiency. This is the strategy of the US company Solexel [], who are aiming to introduce a 22 % efficient silicon solar cell with only 35
m silicon thickness, aiming for 0.42 US$/W
. This means that their cells would be as efficient or better than State-of-the-Art bulk silicon cells, while costs are lower. Their cell rely on light-trapping, i.e. on increasing the absorption length by efficiently scattering the light inside the cell.
1.3 Light Trapping for Thin Film Silicon Solar Cells
Light trapping aims to overcome the problem of low absorption of red and near-infrared photons (the NIR problem) in silicon. This low absorption is a particular problem for thin silicon absorbers. I will first analyse the problem in terms of the optical depth that defines the onset of photon absorption, i.e. the optical bandgap of a material. Light-trapping can then be understood as a means of lowering the optical bandgap of silicon. This interpretation will allow me to provide a more general discussion about the potential benefit of light-trapping in terms of higher power conversion efficiencies.
1.3.1 The NIR Problem
The n ear- i nfra r ed (NIR) problem stands for the loss of opacity by a thinner silicon slab. If the amount of material is reduced, low-energy photons are less likely to be absorbed, because they require a longer absorption length in a silicon slab compared to more energetic photons.
In order to distinguish the threshold for photon absorption from the electronic bandgap
, authors often quote the optical bandgap
instead. Although both quantities are understood as independent of the absorber thickness, thin-film silicon solar cells suffer from optical losses in the NIR, because the onset
essentially depends on the layer thickness, as illustrated in Fig. for silicon.