Hybrid thermolectric - photovoltaic generators can be built in two different system design, optically coupled and thermally coupled. For years it was not clear which of these two strategy leads to higher efficiencies. In this project we have shown that thermally coupled approach is preferable. Here is why.
The main motivation for doing research on hybrid thermoelectric photovoltaic devices comes from the fact that in actual solar cells most of the incoming power is lost in the form of heat. This is the reason why solar cells heat up during operation, and this is also why nowadays solar cell conversion efficiencies are limited to 20-30% of the incoming solar power.
It follows that actual photovoltaic systems have a great potential of improvement with heat recovery. This is basically why thermal recovery approaches combined with solar cells (like cogeneration of warm water, thermophotovoltaics, or thermoelectrics) are becoming more and more popular.
Actually as we showed in a recent paper (titled "Experimental Determination of Power Losses and Heat Generation in Solar Cells for Photovoltaic-Thermal Application" - external link) the amount of heat available in a solar cells during operation can be simply evaluated by their efficiency and optical losses (namely reflectance or transmittance).
Thus if we call ηpv , and Rpv, respectively the solar cell efficiency and reflectance (or transmittance), the percentage of heat losses (ξu) can be evaluated as
considering that silicon solar cells have 20% of efficiency with reflectance around 10%, this means that up to 70% of the power coming from the Sun becomes lost heat. Even for multi-junctions solar cells, which are the most efficient photovoltaic devices, heat losses account for up to 50%. This can be seen in the above graph where the heat losses are estimated as a function of the solar cell efficiency for three cases (CIGS, Si and triple-junction solar cells). The histograms reported accounts for every kind of physical mechanism behind heat losses (refer to the paper for an exhaustive explanation of the physics).
Thermoelectric heat recovery is one of the way in which is possible to recover waste heat in solar cells.
There are mainly two approaches to the thermoelectric recovery of waste heat in solar cells.
The first is the spectrum splitting or (as I call it) the optically coupled approach. In this solution the solar light is split by means of a spectrum splitter in two part. The portion with high energy is then directed on the solar cell, while the infrared is deflected on a thermoelectric module covered by an absorber material (called selective absorber - see the fugure on the left).
The second is instead the thermally coupled approach consisting in the combination of a solar cell with a thermoelectric generator simply placing them in thermal contact with each other (see figure on the right).
However from the analysis of the spectral distribution of heat losses within solar cells, it can be seen that they are distributed over the whole range of wavelengths occupied by the solar spectrum. This can be seen by the graph reported here, where the spectral density of heat losses as a function of the wavelength is reported, for three different solar cells (CIGS, Si, and triple-junction) and compared with the solar spectrum.
This means that for the case of spectrum splitting, the part of the heat losses with wavelength smaller than the splitter cutoff wavelength is not accessible to the thermoelectric part.
Normally cutoff wavelengths are set to be around 1000 nm. This means that most of the heat cannot be recovered, as can be seen in the following graph, where the cumulative amount of heat losses is reported as a function of the cutoff wavelength.
Therefore the so-called thermal coupling approach is more convenient in terms of heat recovery, and final efficiency. Namely it is better to sacrifice some of the solar cell efficiency, forcing it to work at high temperature, but making the thermoelectric part able to access all the heat generated within the solar cell.
That’s the reason why thermal coupling is the approach I followed in my project.