A way around a ‘built in’ speed limit in solar power is being worked out, says S.Ananthanarayanan.
As photocells get cheaper and photocell panels come within reach, generating green power from sunlight is being looked as the answer to the problem of meeting growing power needs and yet avoiding the pollution of conventional generation methods. In the context, improvement in the efficiency of the cells themselves amounts to multiplying the benefits that come from the fall in cost and increase in numbers.
The mechanism of the photocells’ working itself creates a limit, no matter how well they are crafted, to how efficiently they can convert sunlight to electricity. While work-arounds have been devised to improve the overall efficiency of conversion, in combination, a team of researchers working in MIT, USA, report in the journal, Nature, an arrangement using carbon nanotubes and special crystals, to more effectively capture sunlight and convert it to a form that the photocell can use without waste.
The photocell is bit of a material from which electrons can be knocked out by sunlight, and where the cell has an arrangement to prevent the electrons from going directly go back – which means the electrons collect, and can drive an electric current. In the atoms of metals, which are good conductors of electricity, the electrons in the outer shell of the atoms are not tightly fixed to the atoms, but can ‘float’ about in the assemblage of atoms that form a crystal. These ‘free’ electrons can then help the metal carry an electric current. This property is the opposite of the case of insulators, where the electrons are securely bound. But in the case of semiconductors, like silicon, the electrons are half-way free and can be nudged loose by a photon of light. This is particularly so if traces of another atom, which has one more electron in the outer shell, are added to the crystal of the semiconductor material. Now, if a crystal treated like this is placed In contact with another crystal where the atoms of the impurity added have not one electron more, but one electron less than the others, then, the ‘extra’ electrons would cross over, but the ‘shortage’ of electrons on the other side could not do the same thing and the boundary would be ‘one way.’
Now, if sunlight were to fall on this junction, a good number of electrons would get freed and cross over, but cannot return, and the result would be that the two sides would get electrically charged. The discharge could then take place through a normal conductor, which could work machinery, boil an egg or charge a battery. We can see that this could go on so long as a source of light is there and provided the photons of the light shining are energetic enough to supply the minimum, or threshold push that the semi-bound electrons need to get free. Solar cell panels consist of thousands of such junctions grown on large sheets, to catch sunlight and generate electricity. There are no moving parts to wear out and there is no cost of fuel to be provided and the arrangement is economical once the photocells and the panels have been built.
The limitations in efficiency arise from the manner of working itself. We have seen that all frequencies of light falling on the photocell cannot knock electrons out. In the case of silicon, the threshold energy is in the range of the energy of photons in the red part of the spectrum. This means red light and light of colours like yellow, green and blue can help produce power, but not infra-red light or radiation in microwaves or radio waves. Now, in the case of sunlight, about 19% of the energy is in this low energy area. This means that this 19% of the energy falling on the photocell is no good for production electricity. On the other hand this energy would go to heat up the photocell, which gives rise to its own inefficiency.
Another effect comes into force when the photons have energy higher than the threshold. Blue light has energy about twice that of red light and all the energy cannot be captured by a single photo-junction. The extra energy goes to excite or heat the crystal. This loss accounts for about 33% of the incident sunlight and the overall loss, on account of mismatch of frequency, is about 48%.
Another loss of energy in the photocell is because the cell warms up. The fact that the photocell is absorbing energy from the sun is really because the sun is a lot hotter than the photocell. The photocell is also a radiator of heat energy, which could go to heat other objects, or environment, that are cooler than the photocell. Hence, on the one hand, energy corresponding to temperature below the temperature of the cell cannot be captured and, at the temperature of the cell itself, there is loss of energy by radiation. In the case of a photocell working at about 20°C, about 7% of incident energy cannot be used. And then, the cell warms up and usually works at about 75°C. There is hence more energy that cannot be used and more energy that is radiated.
Yet another source of energy loss is because an electron knocked out by photon sometimes combines with another atom from which an electron has been removed and releases energy to the crystal. This again represents photon energy that has not been tapped and is a source of heat.
These effect place a limit, called the Shockley-Queisser limit, of 33.7% on the efficiency of a photocell. In the case of a silicon cell, this limit is lower, at about 29%.
The solutions attempted to get over these limitations are generally to capture the solar energy before it strikes the photocells, and convert the energy into radiation that is tuned the band of energy that suits photo cells, using ‘hot absorber-emitters.’ In practice this has proved challenging – to efficiently capture solar energy and then to have spectral control in the emission. The absorber needs to capture energy and hold on to it so that the emitter reaches high temperatures. And then, the distribution of wavelengths of the emission, at the high temperature, needs to be in the band that suits the photocells .
So far, this has been managed with specially designed cavity geometry of the absorber, with very high degree of concentration of the light falling on the absorber. The combined requirement involves a degree of transmission and reflection losses, with efficiency of about 65%. For the emitter, materials such as the metal tungsten have been employed, with no notable success in selecting a narrow band of emission. The result is that the conversion of solar energy to suit photo cells has been poor.
The MIT team, of Andrej Lenert, David M. Bierman, Youngsuk Nam, Walker R. Chan, Ivan Celanovic, Marin Soljacic and Evelyn N. Wang, have worked on the design and employ an array of multi-walled carbon nano-tubes for the absorber,. And for the emitter, they use a silicon/silica crystal. The layout of the arrangement is shown in the picture.
The absorber-emitter ratio was varied from 1 to 10, to achieve best results. As the ratio was increased, the temperature of the emitter was kept high by increasing the increasing the concentration of sunlight and also by tuning the nanotube array. The communication from the absorber to the emitter was via the silicon material in which both were embedded. The emitter consisted of five alternating layers of silicon and silica, chosen to match the frequencies of emission that suited the photovoltaic cell and also for compatibility with the silicon material in which the arrangement was held.
The results of the arrangement are a three-fold improvement over what has been achieved so far and the design is scalable and adaptable. Efficiency of 80%, is considered feasible. “The efficiency improvements demonstrated in this work, as well as the promising predictions using a validated model, suggest the viability of nanophotonic solar thermophotovoltaics for efficient and scalable solar energy conversion,” the authors say in the paper.
Do respond to : [email protected]