3D photonic crystals for photon management in solar cells

Unbalanced currents in serial-connected tandem solar cells are a major limitation for cost-efficient fabrication of third-generation solar cells.¹ To reduce the current mismatch and to operate a tandem cell at its maximum power point, researchers have suggested introducing intermediate reflective layers between the hydrogenated amorphous silicon (a-Si:H) top cell and the microcrystalline silicon (μc-Si:H) bottom cell to optically alter the photon distribution.² Intermediate reflectors must meet three criteria: selective reflection of photons in the low-absorption spectral region of the top cell, high transmission in the negligible absorption spectral region of the top cell, and high conductivity so that ohmic loss is avoided. Three types of dielectric intermediate reflectors are predominantly discussed in the literature:^{3, 4} transparent, conductive thin films, conductive Bragg reflectors, and conductive 3D photonic crystals (3D intermediate reflection layer, or IRL), as proposed here.

Figure 1. Schematic structure of a micromorph silicon tandem cell with an integrated 3D photonic crystal intermediate reflector. μc-Si:H: Microcrystalline silicon. IRL: Intermediate reflection layer. a-Si:H: Hydrogenated amorphous silicon. TCO: Transparent conducting oxide.

As a model system, we chose the well-established micromorph tandem cell system.⁵ Micromorph silicon solar cells consist of an a-Si:H front absorber, in which the energy equivalent to the band gap energy, E_Tauc, is 1.7eV, and a μc-Si:H bottom absorber with an electronic band gap energy E_gap=1:1eV (see Figure 1). The external quantum efficiency (EQE) values of the series-connected cells for technologically limited thicknesses of the absorbing materials cause an unbalanced distribution of the absorbed photon flux. The top a-Si:H cell typically produces a lower electrical current than the bottom μc-Si:H cell, which limits the overall efficiency.

Figure 2. (a) Scanning electron cross section micrograph of the micromorph tandem cell with 3D IRL. The flat glass substrate (1) is followed by the textured aluminum-doped zinc oxide (ZnO:Al) TCO (2). The a-Si:H cell (3) is deposited conformally, and the inverted opal (4) is grown directly on the back side. Note that the grown opal is crystalline after a few monolayers. On top, the microcrystalline p-i-n solar cell (5) is deposited, followed by the back contact, which acts optically like a back reflector (6). (b) Measured external quantum efficiency (EQE) of the micromorph tandem cell with and without the inverted opal IRL. The 3D IRL enhances the efficiency of the top cell while decreasing that of the bottom cell, as anticipated.

Experimentally, we used a state-of-the-art thin-film a-Si:H solar cell as a top cell. It consists of a glass substrate with a textured transparent conducting oxide (TCO) layer as the front contact and an approximately 280nm-thick a-Si:H p-i-n junction. The solar cell is 10 × 10cm in size, and is extracted from the middle of a large deposition area. The top cell served as a substrate for the inverted opal growth. After atomic layer deposition processing, an aluminum-doped zinc oxide (ZnO:Al) inverted opal is obtained on top of the textured a-Si:H solar cell. The specific resistivity of ZnO:Al has been determined for thin films: it is about 1Ωcm initially and is significantly lower after annealing at 200°C during deposition of the microcrystalline layer.⁶ To complete the micromorph tandem cell, the microcrystalline p-i-n bottom cell (thickness approximately 1.5μm) is deposited via plasma-enhanced chemical vapor deposition. Finally, a ZnO layer and silver grids are deposited on the samples as back reflectors and back contacts, respectively. The resulting solar cells are 5.0 × 2.5mm in size. Figure 2(a) shows a scanning electron microscope cross section of the tandem solar cells, including the 3D IRL.

To verify the impact of the intermediate reflector on the entire cell, we measured the EQE in the spectral range at which the micromorph tandem cell can absorb light, namely, from 300 to 1100nm. Figure 2(b) presents the experimentally measured EQE values for an a-Si:H top cell and a μc-Si:H bottom cell in a tandem solar cell with (solid) and without (dashed) a 3D IRL. The curves show for the first time that an inverted opal can act as a highly efficient intermediate reflector on textured substrates.⁷ For wavelengths between 500 and 800nm, the inverted opal significantly enhanced the EQE of the top cell. Moreover, at smaller wavelengths (below 500nm), no additional losses are observed because all the impinging light is already absorbed in the top cell and the front contact. The short circuit current density j_SC of this top cell without the intermediate reflector is j_SC=9:4mA cm⁻², whereas with the inverted opal it is j_SC=11:7mA cm⁻², which corresponds to an enhancement of η_{jSC; exp}=24:5%. This agrees very well with our numerical simulations,³ which predicted enhancements of up to η_{jSC; theo}=28%. This enhancement of η_{jSC; exp}=24:5% is significantly higher than that achieved with thin-film-based intermediate reflectors. An efficiency increase of η_{jSC; exp}=5%⁸ to η_{jSC; exp}=8%⁹ has been measured experimentally, which is in excellent agreement with our numerical calculations of about η_{jSC; theo}=7%.³

To understand the different contributions to the enhancement inside the tandem solar cell with a 3D IRL, the internal structure of the tandem cell was reconstructed by focused ion beam slicing for further use in a finite-difference time-domain simulation.¹⁰ Because the reference structure (without the 3D IRL) and the structure with the opal were both grown on the textured surface of the top cell, the inverted opal (rather than the textured surface) was found to be the main cause of that cell's enhanced efficiency.

We define a benchmark for the impact of intermediate reflectors as the ratio of the measured EQE with the IRL to that without the IRL: α_EQE(λ)=EQE_w(λ)/EQE_w/o(λ). The enhancement ratios α_EQE(λ) were determined experimentally for a thin-film-based intermediate reflector⁸ and for our 3D IRL.⁷ The benchmark parameter α_EQE for the 3D IRL is enhanced to more than unity across the entire spectral domain, thus demonstrating that the IRL has a purely positive impact. The enhancement factor for the 3D IRL is always greater than that of the thin-film-based IRL and has a maximum at about 720nm, exhibiting an α_{EQE; exp; max} value of 3.6. This is a factor of 2.25 higher than that of the thin-film-based IRL, which explains the significantly improved short circuit enhancement of η_{jSC; exp; 3D}=24:5%, compared to η_{jSC; exp; thinlm}=5%.

In summary, we presented the first 3D photonic crystal-based intermediate reflector fully integrated into a silicon-based tandem solar cell on textured substrates. A ZnO:Al inverted opal enhances the short circuit current of the amorphous silicon top cell by 24.5%, corresponding to a maximum internal path length enhancement of about 3.6. This is about a factor of 2.25 better than current thin-film intermediate reflectors used in micromorph solar cells. As the next step, we plan to investigate optimization of light trapping in the bottom cell.

The authors are grateful for the support of the German Science Foundation DFG under grant PAK88. We also thank the other partners in our consortium (University of Jena and University of Mainz) for their support.