Tandem solar cells with integrated intermediate reflectors

Micromorph silicon solar cells are currently used for large-area and building-integrated photovoltaic applications. As shown in Figure 1(a), such cells are composed of a hydrogenated amorphous silicon (a-Si:H) front absorber and a microcrystalline (μc-Si) bottom absorber. However, they are significantly less efficient than high-efficiency copper indium gallium selenide (CIGS) solar cells and more expensive than CdTe solar cells. One important challenge, therefore, is to increase their overall efficiency.

Figure 1. (a) Schematic diagram of the investigated intermediate reflective layer inside the micromorph tandem cell. (b) External quantum efficiencies of a-Si:H and μc-Si junctions measured inside the tandem cell.

Figure 1(b) illustrates the external quantum efficiencies (EQE) of series-connected cells for a typical thickness of the absorbing material. The difference in EQE results in an unbalanced distribution of the absorbed photon flux. Basically, the combined efficiency of micromorph cells is lessened by the difference in absorbed photon flux between a-Si:H and μc-Si under AM1.5 irradiation. The front cell produces a lower electrical current than the bottom cell, which reduces the overall efficiency (since the tandem cell is connected serially). The representative currents of the junctions are j_{sc, a}=10.7mA/cm²and j_{sc, μ}=12.5mA/cm² measured within the tandem.¹ To reduce this current mismatch and to operate the tandem at its maximum power point, intermediate reflective layers (IRL) may be introduced between the a-Si:H and μc-Si to optically alter the photon distribution. Several approaches using IRLs have been investigated successfully in recent years.^2–6 A typical silicon tandem cell with a scattering front of j_sc = 12.1mA/cm² and V_{oc, a} = 0.91V, V_{oc, a} = 0.5V reaches an efficiency of 11.1% (fill factor = 0.65). We have shown that an intermediate reflector has the theoretical potential to increase the current in the cell, resulting in an efficiency of 12.4%, i.e., an absolute increase of 1.3%.⁷

From the working principle of the micromorph tandem cell, three main requirements for the IRL can be deduced. First, it should provide for spectrally limited back reflectance for the a-Si:H top cell, where the EQE of both junctions overlaps. It should also be transparent to red and IR light to avoid a negative effect on the bottom cell. Finally, it must exhibit sufficient electrical sandwich conduction, as the micromorph tandem is series connected. An IRL must reflect photons preferentially in the low-absorption regime of a-Si:H (550nm–700nm) back into the a-Si:H top cell. A common approach is the use of a thin ZnO layer that generates Fabry-Pérot oscillations that enhance the photon flux in the a-Si:H cell.^1,8,9 Three-dimensional thin-film photonic crystals have been suggested^3,7,10 as an integrated reflector to optically match the current by appropriate photon management between the two absorbers (because of their spectral selectivity and diffractive properties).⁸ See Figure 1(a).

In this way, we produced an initial integrated prototype, as shown in Figure 2(a). To verify the effect of the IRL on the top cell, we measured the EQE in the visible spectrum, as shown in Figure 2(b). Note that we measured the EQE at a reversed bias voltage because of the low conductivity of the undoped ZnO. The effect of the bias has been observed to saturate between 2V and 3V, which is interpreted as a large series resistance in the IRL. Compared to a high-efficiency reference cell with randomly textured light-trapping surfaces, the flat cell with a 3D photonic IRL suffers from reduced absorbance over nearly the entire absorption region. Nevertheless, the cell constitutes a proof of concept because the absorption is enhanced in the spectral range of the designed back reflectance of the IRL. In this range, both EQE values are at the same level, as indicated by the black circle in Figure 2(b).

Figure 2. (a) Scanning electron microscope image of the prototype's profile: 1) substrate layer, 2) transparent conducting oxide (TCO) front contact, 3) a-Si:H top cell, 4) photonic IRL, and 5) back-side silicon and metal back contact. (b) Experimentally determined spectral EQE of the prototype and a state-of-the-art reference cell (a-Si:H cell on textured substrate). In the spectral range where the back reflectance of the IRL is maximum, the EQEs are comparable (black circle).

Since the randomly rough surface of HCl-etched ZnO on the front glass is currently the best available light-trapping structure for the tandem cell,¹¹ growing an opaline IRL directly on this conformal roughness would allow direct introduction of a novel interlayer during fabrication of the micromorph tandem cell. We must emphasize that the texture is a key characteristic of an efficient micromorph tandem. However, rough substrates are typically avoided for self-organized growth of photonic crystals. For opals, a rough substrate usually leads to amorphous colloidal films with a highly disordered arrangement of the spheres. Long-range crystal order is impossible for thin films in contact with such a surface. Local crystallization, however, may be possible, depending on the actual size parameters of the roughness. When randomly structured surfaces are combined with highly ordered colloidal photonic crystals, the pivotal question of whether an inverted opal grown on a rough substrate is still crystalline and exhibits diffractive and refractive properties arises. Growth experiments with poly(methyl methacrylate) opal films on the back side of randomly rough HCl-etched ZnO of (superstrate) a-Si:H top cells show that the self-organization process of opals can counteract the distortion caused by the roughness of the substrate.

In conclusion, we have produced a prototype cell with an integrated IRL and demonstrated enhanced absorption in the chosen spectral range. All the experimental steps necessary for a new improved prototype are now possible with the achievement of this milestone. We continue to study the assembly process of inverted opals on textured substrates and the optical properties of these photonic films. We aim to realize experimentally a fully integrated 3D inverted opal IRL within a randomly textured tandem cell and the corresponding EQE measurements.

Financial support from the German Research Foundation under the Pak88 project Nanosun is gratefully acknowledged.