Integration of a-Ge:H nanocavity solar cells in tandem devices

https://doi.org/10.1016/j.solmat.2015.07.032Get rights and content

Highlights

  • Amorphous germanium (a-Ge:H) replaces µc-Si:H absorber in micromorph devices.

  • A resonant nanocavity is used to enhance the bottom cell absorption.

  • Additional undesired resonances are induced by the top cell.

  • Rough interfaces suppress top cell induced resonances, but keep nanocavity effect.

  • Current density similar to micromorph devices reachable with 20 nm a-Ge:H absorber.

Abstract

By taking advantage of spectrally broad resonances, nanocavity-enhanced a-Ge:H solar cells with an absorber layer thickness below 20 nm can reach current densities similar to micron-thick µc-Si:H devices. However, as nanocavity-enhanced devices are highly reliant on interference effects, further spectrally narrow resonance patterns are generated if an additional top cell is added to form a multijunction solar cell. This may complicate the integration of a-Ge:H nanocavity solar cells in tandem devices. We show that conventionally textured TCO substrates can be employed to suppress the top cell induced interferences, while the required broadband resonance of the a-Ge:H bottom cell nanocavity is maintained. This approach is realized in an a-Si:H/a-Ge:H tandem solar cell with an only 20 nm thick a-Ge:H bottom cell absorber. The spectrally broad quantum efficiency curve of the bottom cell corresponds to a photocurrent density of 12.3 mA cm2, which is comparable to values reached in micromorph devices.

Introduction

Silicon thin-film solar cells feature several advantages over other commercially available solar cell technologies. They are based on environment-friendly and abundant materials, and can be fabricated with comparably low energy consumption. Low film thicknesses and low deposition temperatures allow their fabrication on flexible and lightweight substrates [1], suggesting high potential for cost reduction on the module level as well as for use in special applications such as lightweight construction roofs.

To increase the efficiency of the silicon thin-film solar cell technology, tandem devices have been developed over the last 20 years which consist of an amorphous silicon (a-Si:H) top cell and a microcrystalline silicon (µc-Si:H) bottom cell [1], [2]. This way, the lower thermalization losses of the a-Si:H top cell for short-wavelength light can be combined with the ability of µc-Si:H to absorb light having longer wavelengths. However, to match the current-densities of both sub cells, more than 1 µm thick µc-Si:H bottom cells are typically required, while the top cell has a thickness of only about 300 nm. Hence, while the gain in cell efficiency is moderate compared to an a-Si:H single cell (12.7% vs. 10.2% [3]), prolonged deposition time leads to significantly increased fabrication costs.

To address this issue, microcrystalline silicon–germanium alloys (µc-SiGe:H) have been investigated as alternative low bandgap absorbers [4], [5]. Alloying germanium to silicon increases the absorption coefficient and lowers the bandgap of the material. Unfortunately, the electronic properties strongly degrade with increasing germanium content, implying a rather limited potential for use in electro-optical devices. Although amorphous silicon–germanium alloys (a-SiGe:H) have also been studied as replacement of the µc-Si:H solar cells [6], [7], a germanium content close to 100% is required to take advantage of the same spectral range as in µc-Si:H solar cells. In fact, pristine amorphous germanium (a-Ge:H) of highest electronic quality exhibits an optical bandgap of about 1.1 eV, which is very similar to the mobility bandgap of µc-Si:H. On the other hand, this material suffers from an approximately ten times higher defect density compared to a-Si:H [8], [9]. For this reason, solar cells using 100 nm thick high-quality a-Ge:H absorber layers have reached efficiencies of only about 3.5%, up to now [10]. Hence, it is helpful to reduce the absorber thickness to a few tens of nanometers to enable an efficient charge carrier collection. However, for such thin layers, efficient light absorption can only be achieved by applying a highly efficient light-trapping concept.

As we have recently demonstrated, light-trapping can be addressed by using a Fabry-Pérot resonator to enhance the absorption of photons with energies slightly above the absorber bandgap [11]. Such nanocavities have drawn attention due to their ability to produce strong and spectrally broad interference patterns [12], [13], [14]. Depending on the phase jumps that light undergoes when it is reflected at the resonator surfaces, the required optical cavity length can be much lower than 1/4 of the wavelength for which resonant enhancement is desired. This leads to spectrally broad resonances rather than narrow interference patterns. Together with the exceptionally high absorption coefficient of a-Ge:H, this enables a current density of 20 mA cm2 in an a-Ge:H solar cell with an only 13 nm thick absorber layer [11]. Although the cell efficiency has not been increased significantly compared to the best results reported by Zhu et al. [10], a value of 3.6% has been reached with absorber material of lower quality. Apart from the particular case of a-Ge:H solar cells, this concept is also interesting for devices based on other highly absorbing materials which exhibit low charge carrier mobility-lifetime products.

In view of the goal to replace the µc-Si:H absorber in a-Si:H/µc-Si:H solar cells, the question arises of how such nanocavity-enhanced solar cells can be integrated into multijunction concepts. We use 1-dimensional optical simulations of an a-Si:H/a-Ge:H tandem device to show that additional top cell induced resonances strongly modify the infrared absorption of the bottom device compared to the single cell result. An approach to solve this problem is discussed and implemented in first a-Si:H/a-Ge:H tandem solar cells.

Section snippets

Simulation

One-dimensional optical simulations have been performed in order to calculate quantum efficiency (QE) spectra of single and tandem solar cells. Calculations were carried out using the software package Scout/Code (W. Theiss Hard- and Software). Complex refractive index data were taken from literature (a-Ge:H [8] and silver [15]) or were obtained by ellipsometric spectroscopy and subsequent modeling of the measured spectra (a-Si:H, µc-Si:H), respectively. For ZnO, data were taken from the

a-Ge:H nanocavity single cells

While a detailed explanation of the general concept of a-Ge:H nanocavity solar cells has been given before [11], [12], Fig. 1 briefly illustrates the effect: Fig. 1a displays a simplified version of the a-Ge:H solar cell stack used throughout this work. The a-Ge:H absorber layer as well as the silicon layers exhibit a high refractive index of around 4, while the TCO layers typically show a refractive index of about 2. This means that the interface between silicon and the TCO exhibits a

Conclusion and outlook

We have presented quantum efficiency measurements of a-Si:H/a-Ge:H tandem devices employing a nanocavity to increase the absorption inside the a-Ge:H bottom cell absorber. Our results demonstrate that textured substrates, which are commonly used in silicon-based thin-film solar cells, can be employed to suppress the additional interference patterns induced by the top cell. At the same time, the nanocavity resonances required for efficient absorption inside the bottom cell are maintained. As

Acknowledgments

We would like to thank Ulrike Kochan for UV–vis measurements, Martin Kellermann for top cell fabrication, and Antje Schweitzer and Ortwin Siepmann for laser ablation of the front contacts. We would also like to thank the BMBF for supporting this work in the framework of the Project SiSoFlex (FKZ 03SF0418B).

References (18)

There are more references available in the full text version of this article.

Cited by (14)

  • Diffusion and reaction kinetics governing surface blistering in radio frequency sputtered hydrogenated a-Si <inf>x</inf> Ge <inf>1-x</inf> (0 ≤ x ≤ 1) thin films

    2019, Thin Solid Films
    Citation Excerpt :

    The structural and surface quality of the material employed is one of the most critical issues as regards the large scale application of electronic devices based on hydrogenated amorphous silicon (a-Si:H), germanium (a-Ge:H) and a-SiGe:H. Atomic hydrogen migration occurs in the amorphous network. The high temperatures applied during growth of those materials, e. g. by chemical vapor deposition [1–3], or reached during device operation [2,4] enhance the diffusion of H atoms, in particular of those liberated from their bonds to the host atoms as a consequence of annealing. Such enhanced diffusion favors the migration of H atoms towards nanovoids where they very likely form molecular H2 since the reaction 2MeH → H2 + Me-Me is an exothermic one [5] (Me indicates the host atom: Si or Ge).

  • Vegard's-law-like dependence of the activation energy of blistering on the x composition in hydrogenated a-Si<inf>x</inf>Ge<inf>1-x</inf>

    2018, Journal of Alloys and Compounds
    Citation Excerpt :

    Much less work has been done for not implanted semiconductors. Among them are hydrogenated a-Si (amorphous Si), a-Ge and their alloy a-SixGe1-x that find applications in solar cells [16,17,22–28]. As regards a-Ge and a-SixGe1-x they are also very suitable for IR radiation sensors [29], like un-cooled microbolometers [30], thin films transistors [31], detectors for X- or γ-ray imaging [16] and fiber-optic systems [32].

  • Nanostructuring solar cells using metallic nanoparticles

    2018, Metal Nanostructures for Photonics
  • Resonant-cavity-enhanced a-Ge:H nanoabsorber solar cells for application in multijunction devices

    2016, Nano Energy
    Citation Excerpt :

    While this is still far from the 11.8% currently achieved with µc-Si:H solar cells [23,24], single cell efficiencies should not be directly used as a figure of merit, as they do not reflect the fact that the RCE a-Ge:H solar cell presented here can show its full potential in multijunction configurations only. As demonstrated recently [15], it is impossible to integrate an RCE a-Ge:H solar cell into a tandem device by merely stacking a second device on top of it. To illustrate this, Fig. 3 depicts the simulated QE of an RCE a-Ge:H single solar cell as well as of a corresponding a-Si:H/a-Ge:H tandem cell stack.

View all citing articles on Scopus
View full text