Photon management structures for solar cells

Photon management structures are becoming increasingly important for solar cells as thinner wafers or even thin absorbing films are used and high external quantum efficiencies over the entire usable spectrum are required for high solar cell efficiencies.¹ In particular, elaborate photonic structures are attracting more interest. Such structures exploit interference, diffraction, or resonance effects to obtain spectral or angular selectivity or electromagnetic near-field enhancement.¹

The question, however, is how to establish an economically feasible process chain to manufacture photonic structures.² Interference lithography can produce finely detailed micro- and nanostructures over large areas (more than one square meter).^3–5 But the technology for originating so-called master structures is demanding. Combining it with nanoimprint processes for replicating structures represents a structuring approach with the requisite upscaling potential. For this reason, we focus on structures that are manufactured by a combination of interference lithography and nanoimprint lithography (NIL).

Interference lithography uses no mask to define the exposure pattern. Instead, a laser beam is split, and the resulting coherent beams are expanded and then superimposed on a photoresist-coated sample. The interference pattern is used to expose a photoresist-coated substrate, producing a surface relief pattern after a development step. Thus, interference lithography offers a large variety of achievable structure types and profile shapes (see Figure 1), structural dimensions ranging from 100nm to 100μm, and seamless structured areas up to 1.2×1.2m² in a single exposure.

Figure 1. Structures formed by interference lithography: (a) linear grating, (b) crossed grating with high aspect ratio, (c) honeycomb structures, (d) 3D photonic crystal, (e) nanotextured scattering surface, and (f) combination structure for enhanced adhesion inspired by the feet of geckos.

The master structures in photoresist can serve as an etching mask for pattern transfer or as a template for infiltration with different materials. Alternatively, they can be replicated via electroplating and subsequent replication processes such as NIL.

NIL describes a process sequence in which a polymer layer is first patterned by hot embossing or a UV-replication process, and this patterned layer is then used as an etching mask to transfer the defined patterns to a substrate underneath.⁶ NIL processes have the potential to allow the fabrication of micro- and nanostructures on large areas in an industrially feasible way.

The master structures fabricated via interference lithography are used to replicate stamps for the nanoimprint process (this can be seen as a preliminary process step). As the stamp material, we use polydimethylsiloxane (PDMS), which allows non-wearing replication by cast molding. The flexibility of these elastomeric stamps enables conformal contact over large areas even on non-planar surfaces.⁷

After the stamps are fabricated, the repetitive process steps for imprinting a resist layer are carried out. The PDMS stamp is pressed onto a resist-coated substrate and exposed to UV radiation while pressure is maintained. After the resist is cured, the stamp must be demolded, and the patterned layer remains on top of the substrate. This patterned resist layer can then be used as the starting point for several processes, such as etching, lift-off, and deposition (see Figure 2, left).

Figure 2. Left: Processing sequence presented in this article. The starting point in this diagram is the structure in photoresist on a glass substrate (top left). From this, a polydimethylsiloxane (PDMS) stamp (center left) is replicated, which can be used in the nanoimprint process. Right: Roller UV nanoimprint lithography (NIL) setup. Red laser radiation is used to visualize diffraction at the structured PDMS stamp.

Figure 3. Structures in silicon after NIL and plasma etching. Left: Honeycomb textured multicrystalline silicon wafer (perspective view of a grain boundary region, with magnified inset). Right: Crossed rear side grating (period 1μm, depth 270nm).

To further increase the feasibility of the imprinting processes on an industrial scale, we are developing a roller NIL tool (see Figure 2, right). It allows the patterning of UV-curing resis layers on brittle, stiff, and opaque substrates (e.g., silicon) in a continuous process flow. We are currently upscaling this tool for patterning 156×156mm²substrates (this is the typical silicon wafer size in solar cell production).⁸

The presented process chain enables the fabrication of photon management structures having various geometries and functionalities, for example, honeycomb front-side textures and diffractive rear-side gratings. Front-side textures are crucial for high-efficiency crystalline silicon solar cells. On monocrystalline wafers, very efficient pyramidal textures can be generated. This, however, is not possible for multicrystalline silicon. On this material, honeycomb textures contributed to the current record efficiency of 20.4%.⁹ However, this result was achieved using a photolithography process that is not suitable for industrial production. To enable record efficiencies in industry, we use NIL in combination with plasma etching to fabricate honeycomb textures (see Figure 3, left). As a result, reflectance values close to those for inverted pyramids¹⁰ and short-circuit current densities exceeding 40mA/cm² on float zone material¹¹ were demonstrated.

Rear-side gratings in solar cells allow the path length of weakly absorbed light to be enhanced by diffractive effects. Thus, this concept is very interesting for silicon, which has a very small absorption coefficient for long wavelengths.¹² We have investigated a process chain based on interference lithography and NIL for the fabrication of linear and crossed gratings for this purpose (see Figure 3, right). As predicted by theory, we demonstrated enhanced absorption in the spectral region close to the band gap of crystalline silicon.¹³

In addition to these examples, various other photon management structures for solar cells can be fabricated. The imprinted resist can be used a a textured substrate for transparent conducting oxide or metal layers in thin-film solar cells.⁵ Furthermore, we have used NIL in combination with lift-off processes to fabricate defined metal nanoparticle arrays for plasmonics.¹⁴

In summary, we have shown that a combination of interference lithography and NIL has potential as an industrially feasible method of producing photon management structures for solar cells. The next step will be to demonstrate efficiency enhancements in solar cells with the described structures. Furthermore, the possibilities of our structuring technologies will be exploited in a variety of new applications not only in photovoltaics, but also in solid-state lighting and display technology.

Parts of this work were funded by the German Federal Ministry of Environment, Nature Conservation, and Nuclear Safety under contract 0325176 (NanoTex).