Linking simulation and synthesis of nickel oxide hole-selective contacts for silicon heterojunction solar cells

In silicon heterojunction (SHJ) solar cells, one factor limiting efficiency is "parasitic absorption" from amorphous silicon (a-Si) at the front surface. Thus, discovery of a transparent hole-selective contact to replace p-type a-Si is a promising area of research, but this has proven challenging due to the additional requirements of surface passivation and band alignment, among others. In this investigation, we select nickel oxide (NiOx) as a test-case hole-selective contact material, since it is p-type and has a tunable valence band (VB) minimum aligned with that of crystalline silicon. We simulate an SHJ front interface with NiOx to investigate possible performance and optimal parameters. From this analysis, we identify a trade-off between VB alignment and hole concentration in the NiOx layer: the VBM has to be very close to its optimum value for low hole concentrations, but for higher hole concentrations there is a wider range of tolerance of VB position. To test these findings, we sputter deposit amorphous NiOx hole selective contacts to fabricate a SHJ device stack, varying oxygen flow rate during growth. NiOx is likely not a great hole-selective contact material, possibly due to VB misalignment, low carrier concentration, and interfacial defects. However, NiOx has effectively served as a test-case to demonstrate the influence of varying VB position and thickness on cell performance and verify our model. By comparing simulations to experiments, we have identified ranges of tolerance for optimal material properties that can inform future materials discovery efforts for alternative transparent hole-selective contacts.

TEM characterization of textured silicon heterojunction solar cells

Christophe Ballif, Aïcha Hessler-Wyser, Christian Monachon

Increasing the efficiency of crystalline silicon (c-Si) solar cells requires the reduction of both bulk and interface recombination. Even if bulk recombination is almost suppressed, the symmetry of the crystal lattice is disturbed at the surface and hence, due in particular to non-saturated bonds (dangling bonds), a large density of defects exists. Thus, in this case, the free-carrier lifetimes are no longer limited by the quality of the bulk c-Si, but by its surface. To keep the recombination losses at the c-Si surface at minimal levels, the surface must be electronically well passivated. An efficient way to obtain passivation is to use low temperature grown (typically 200°C) hydrogenated amorphous silicon (a-Si:H) [1]. In the case of photovoltaic applications, the passivation of both c-Si wafer surfaces (i.e., the emitter and the rear surface) is of crucial importance for good performances. In this work, the 3-7 nm a-Si:H based passivation layers of the device are grown by VHF-PECVD in a single chamber. The solar cells consist of a multilayered structure: Al back contact / DC-sputtered ITO (Indium Tin Oxyde) / n/i a-Si:H back surface field (BSF) / n-type c-Si substrate with a resistivity of 1-3 Ωcm / intrinsic a-Si:H / i/p a-Si:H emitter / and a front contact made out of DC-sputtered ITO via a shadow mask to define the cells (diameter = 4.5 mm). Such devices are called heterojunction (HJ) silicon solar cells. Surface recombination losses are a major concern for all c-Si solar cells. In particular, mastering of HJ emitter and back surface field formation on textured c-Si is crucial for high-performance HJ solar cells. High Resolution Transmission Electron Microscopy (HRTEM) is necessary to identify the key microstructural features of the a-Si:H/c-Si interface, and TEM micrographs of HJ interfaces on flat and textured high-performance devices are shown for the first time and discussed with respect to the resulting solar cell electrical performances. TEM micrographs of our flat high-efficiency HJs show abrupt c-Si/a-Si:H/μc-Si:H interfaces for emitter and BSF formation. Whereas on the pyramidal facets of the textured substrate the growth is identical to the flat substrate interface, we observed unexpected epitaxial growth at the bottom of the pyramid valleys. We have identified these local epitaxial domains as an efficient surface recombination path, and by consequence, as responsible for the observed decreased VOC on textured HJ cells. When minimizing the density of epitaxial domains at the c-Si/a-Si:H interface by adapting the deposition conditions, a solar cell VOC of 660 mV is obtained. An additional modification of the textured c-Si surface morphology leads to VOCs as high as 700 mV.

Springer2008

Textured silicon heterojunction solar cells with over 700mV open-circuit voltage studied by transmission electron microscopy

Christophe Ballif, Stefaan De Wolf, Luc Fesquet, Aïcha Hessler-Wyser, Christian Monachon

In this article, we report on the use of transmission electron microscopy (TEM) for the fabrication of high-performance textured amorphous/crystalline silicon (a-Si:H/c-Si) heterojunction (HJ) solar cells. Whereas classical thin-film characterization techniques allowed us to optimize the a-Si:H layer properties for flat HJ solar cells (open-circuit voltages (VOC) up to 710 mV and energy conversion efficiencies up to 19.1%), these techniques can not always be fully exploited on textured c-Si surfaces. Nevertheless, in this situation, TEM micrographs permit us to identify device performance limiting factors, such as, e.g., local epitaxy in pyramid valleys as the main source of our VOC-loss. Minimizing this local epitaxy by adjusting the amorphous Si based layers growth conditions and improving the c-Si surface morphology yields Si HJ solar cells with VOCs over 700 mV.

2008

Carrier transport and sensitivity issues in heterojunction with intrinsic thin layer solar cells on N-type crystalline silicon: A computer simulation study

Christophe Ballif, Anwitaman Datta, Père Roca i Cabarrocas

Heterojunction with intrinsic thin layer or HIT solar cells are considered favorable for large-scale manufacturing of solar modules, as they combine the high efficiency of crystalline silicon (c-Si) solar cells, with the low cost of amorphous silicon technology. In this article, based on experimental data published by Sanyo, we simulate the performance of a series of HIT cells on N-type crystalline silicon substrates with hydrogenated amorphous silicon (a-Si:H) emitter layers, to gain insight into carrier transport and the general functioning of these devices. Both single and double HIT structures are modeled, beginning with the initial Sanyo cells having low open circuit voltages but high fill factors, right up to double HIT cells exhibiting record values for both parameters. The one-dimensional numerical modeling program "Amorphous Semiconductor Device Modeling Program" has been used for this purpose. We show that the simulations can correctly reproduce the electrical characteristics and temperature dependence for a set of devices with varying I-layer thickness. Under standard AM1.5 illumination, we show that the transport is dominated by the diffusion mechanism, similar to conventional P/N homojunction solar cells, and tunneling is not required to describe the performance of state-of-the art devices. Also modeling has been used to study the sensitivity of N-c-Si HIT solar cell performance to various parameters. We find that the solar cell output is particularly sensitive to the defect states on the surface of the c-Si wafer facing the emitter, to the indium tin oxide/P-a-Si:H front contact barrier height and to the band gap and activation energy of the P-a-Si:H emitter, while the I-a-Si:H layer is necessary to achieve both high Voc and fill factor, as it passivates the defects on the surface of the c-Si wafer. Finally, we describe in detail for most parameters how they affect current transport and cell properties.