Despite recent price drops, photovoltaic electricity needs to decrease further in cost to become competitive in most markets. The cost of a photovoltaic module is still near $1.5/watt-peak (Wp).1 The largest part of this cost is shared between the silicon wafer and the module (or panel) engineering. The sawed silicon wafer alone, before any processing, costs around $0.4/Wp. In order to tackle both the silicon and module cost issues, we are developing a novel combined approach, using thin crystalline-silicon (c-Si) layers and directly combining cell and module processing at an early stage.2
Thin c-Si layers benefit from all the material characteristics of silicon wafers (stability, abundance, high efficiency potential) but suffer from low absorption over a broad spectral range. A 1μm-thick layer of c-Si coated with an antireflection coating (ARC) layer transmits more than one-half of the incident solar photons on the absorption spectrum of silicon (from 300 to 1200nm, in terrestrial applications). A 40μm-thick layer loses 10% more photons than a 200μm-thick one (5% with a back reflector). Figure 1 shows the transmission spectra of these layers with a silicon nitride (SiNx) ARC.
Figure 1. Numerically calculated transmission spectra of various crystalline-silicon (c-Si) layers with ARC (anti-reflection coating). SiNx: Silicon nitride.
The light-trapping techniques currently used to balance the absorption losses in c-Si are mostly based on pyramidal texturing. But pyramidal textures yield important material losses of 10–30μm of silicon thickness.2–5 In order to adapt to the reduced thickness of the micron-thin c-Si, we propose patterning at the nanoscale. This scale has the combined advantages of consuming less than 1μm of silicon and of offering a very high optical potential. However, this nanopatterning needs to yield an energy-conversion efficiency at least equivalent to the present state-of-the-art pyramidal texturing. For c-Si layers 30–40μm thick, a light-trapping equivalent to pyramidal texturing is still sufficient, as the absorption loss due to the low c-Si thickness is small (only 5% of short-circuit current is lost when thinning down the silicon layer from 200 to 40μm). However, for the 1–30μm thickness range, the current loss increases dramatically (more than 40% loss when thinning down to 5μm, and 60% at 1μm) and an improved light-trapping technique is needed.
Figure 2. Effect of nanopatterning: reduced front-side reflectance due to a graded-index effect and diffraction of light inside the active layer at high angles. I: Incident light. R: Reflectance. T: Transmission.
A randomly patterned surface reflects/transmits light with a Lambertian distribution, i.e., proportionately to the cosine of the reflected/transmitted light angle. Lambertian surfaces are usually considered the upper limit for photovoltaic cell light trapping. This simple statement already shows the potential for improvement: random nanopatterning can help in reaching a Lambertian situation.6 Better than random, controlled nanopatterning could offer the possibility of modifying this angular distribution so as to further increase the light path in the cell. The nanopatterning effect can be divided into two main phenomena (see Figure 2), a decrease in the front side reflection due to a graded-index effect, and an increase in the optical path of the light coupled into the structure thanks to scattering and/or diffraction of light inside the active layer at high angles.
Figure 3. Reflectance spectra and integrated reflectance values for various front-side texturing techniques on wafers, without ARC. Nanopatterning competes with traditional pyramidal texturing for decreasing front-side reflectance, and exceeds it at high wavelengths where diffraction-assisted light trapping takes place.
These two phenomena can be reached in various ways. The front side reflection could, in principle, reach zero if the refractive index transition between the incident medium and the active layer is continuous. This can be approximated with a suitable progressive nanopattern profile (see Figure 3) or a multilayer ARC. The increase in light path can be reached through diffraction thanks to a grating on the front or on the back. Scattering can be enhanced with similar techniques, but using a random structure.
Reflectance of a 1μm-thick cell (1μm-thick monocrystalline layer on a back reflector and coated by an ARC) with and without nanopatterning.4
Upper right: The silicon layer cross section after patterning by nanoimprint lithography and reactive ion etching. Lower right: Cell structure schematics. Ag: Silver. a-Si/ITO: Amorphous silicon/indium tin oxide. Al: Aluminum.
To demonstrate the possibilities offered by such nanopatterning schemes, we use solar cells with layers of 1μm-thick monocrystalline silicon.7 We tried to maximize the short-circuit current density given by the absorption in the structure's active layer. The short-circuit current expresses how well a cell absorbs light, and is one of the key performance indicators of a solar cell. The successful integration of controlled nanopatterning was performed thanks to a combination of nanoimprint lithography and reactive ion etching.4 Figure 4 shows reflectance spectra of a cell stack, with and without nanopatterning. These spectra allow the maximum achievable current to double, potentially yielding ∼29–31mA/cm2 of short-circuit current for 1μm thickness of silicon. The exact value will depend on the contacting scheme and surface passivation. With good (but realistic) cell parameters of 75% diode characteristic filling factor and 600mV open circuit voltage, this would already reach an efficiency over 13% for a 1μm-thick cell. This means we can decrease silicon consumption by a factor of 150 with a corresponding decrease in efficiency of just 30%.
We will investigate different nanopatterning techniques with direct cell-on-module engineering, which enables processing of a thin silicon foil directly at module level.2 Soon we hope to fabricate 1μm-thick solar cells with these patterning techniques. Over the long term, our goal is to integrate these into industrial solar cells. This should allow high efficiencies with ultra-thin c-Si: the best of photonics empowering photovoltaics for affordable renewable energy.
Valerie Depauw, Ounsi El Daif
Valerie Depauw is a researcher. Her work focuses on various fabrication technologies for thin c-Si films for photovoltaics.
Ounsi El Daif is a researcher working on various light management schemes for inorganic solar cells.
Christos Trompoukis is a PhD candidate at imec. The core of his research is nanopatterning for advanced light trapping in thin c-Si solar cells.
1. Solar Photovoltaics Competing in the Energy Sector—On the Road to Competitiveness, tech. rep., European Photovoltaic Industry Association, 2011.
2. F. Dross, K. Baert, T. Bearda, J. Deckers, V. Depauw, O. El Daif, I. Gordon, Crystalline thin-foil silicon solar cells: where crystalline quality meets thin-film processing, Prog. Photovolt. Res. Appl.
, 2012. doi:10.1002/pip.1238
3. V. Depauw, O. El Daif, D. Van Gestel, K. Van Nieuwenhuysen, C. Trompoukis, F. Dross, I. Gordon, J. Poortmans, Decreasing the thickness of crystalline-silicon solar cells below 40 μm and increasing their light absorption with surface nanostructures, Renewable Energy and the Environment, OSA Tech. Dig. (CD), OSA, 2011.
4. C. Trompoukis, A. Herman, O. El Daif, V. Depauw, D. Van Gestel, K. Van Nieuwenhuysen, O. Deparis, J. Poortmans, Enhanced light trapping in crystalline silicon solar cells by nanoimprint lithography, Proc. SPIE
8438, p. 84380R, 2012. doi:10.1117/12.921212
5. V. Depauw, O. El Daif, C. Trompoukis, K. Van Nieuwenhuysen, F. Dross, I. Gordon, J. Poortmans, Nanostructured micron-thin crystalline-silicon solar cells: towards tailored and integrated light-management schemes, Proc. SPIE 8438, 2012. (Invited paper.)
6. I. Gordon, L. Carnel, D. Van Gestel, G. Beaucarne, J. Poortmans, 8% efficient thin-film polycrystalline-silicon solar cells based on aluminum-induced crystallization and thermal CVD, Prog. Photovolt. Res. Appl.
15, p. 575-586, 2007. doi:10.1002/pip.765
7. V. Depauw, Y. Qiu, K. Van Nieuwenhuysen, I. Gordon, J. Poortmans, Epitaxy-free monocrystalline silicon thin film: first steps beyond proof-of-concept solar cells, Prog. Photovolt. Res. Appl.
19, p. 844-850, 2011. doi:10.1002/pip.1048