Correlation between micro-roughness,surface chemistry,and performance of crystalline Si/amorphous Si:H:Cl hetero-junction solar cells

The correlation between micro-roughness, surface chemistry, and performance of crystalline Si/amorphous Si:H:Cl hetero-junction solar cells is discussed through a deposition study of amorphous Si:H:Cl (a-Si:H:Cl) films by rf plasma-enhanced chemical vapor deposition using a SiH₂Cl₂–H₂ mixture. The degree of H- and Cl-termination on the growing surface determined the degree of micro-roughness at the p-type a-Si:H:Cl/intrinsic a-Si:H:Cl interface and solar cell performance. A higher degree of Cl-termination compared to H-termination was effective to suppress the micro-roughness at the growing surface and oxygen incorporation into the film, as well as chemical reduction of the intrinsic a-Si:H:Cl layer during the underneath p-layer formation. The study showed that a-Si:H:Cl deposited from SiH₂Cl₂ is a potential material for c-Si hetero-junction solar cells with an intrinsic a-Si:H:Cl thin layer.     

Cross-contamination in single-chamber processes for thin-film silicon solar cells

Boron (B) and phosphorus (P) cross-contamination for single-chamber deposited a-Si:H, μc-Si:H, and a-Si:H/μc-Si:H tandem solar cells has been investigated by studying their impact on the different layers of solar cells. To reduce the B and P cross-contamination into the i-layer and p-layer, respectively, to a tolerable level, for a-Si:H and μc-Si:H cells a 15' evacuation cycle prior to the i-layer deposition is applied. The effect of P cross-contamination into the i-layer is strongly reduced by the p-layer deposition and a 15’ evacuation cycle prior to the i-layer deposition. The p-layer is assumed to cover up or to fix (in form of P-B complexes) most of the P at the chamber walls. This leads to high quality μc-Si:H cells and a-Si:H cells with only slightly reduced performance. Here, a soft-start of the a-Si:H i-layer led to high quality cells, presumably due to reduced P recycling. Further, there is no need to clean the process chamber with, e.g. NF₃, after each p-layer, as applied in many industrial processes. Instead, many cells are deposited without cleaning the process chamber. We established a single-chamber tandem cell process with 15' evacuation cycles prior to the μc-Si:H p-layer and to each i-layer with a cell efficiency of ~ 11.1%.     

Determination of the defect density in thin film amorphous and microcrystalline silicon from ESR measurements: The influence of the sample preparation procedure

Accurate evaluation of the defect density (N_D) is of high relevance for the optimization of thin film silicon. The spin density (N_S) measured in ESR experiments is often used as a measure for the density of deep defects in the material, assuming that all defects are in a paramagnetic charge state. However, exposure to air, water, or acid during ESR sample preparation can potentially change the N_S in a sample and lead to misinterpretation of N_D. We have investigated how the preparation procedures of a Si thin film ESR sample may affect the properties of its ESR spectrum. Samples of different structural composition from highly crystalline μc-Si:H to a-Si:H deposited by PECVD on Mo-foil, Al-foil and ZnO:Al were studied for different states of exposure to ambient conditions and annealing. N_S measured directly after sample preparation and after air exposure was found to be higher than N_S measured in the annealed state. Particularly in highly crystalline material this discrepancy may reach one order of magnitude. On the other hand in a-Si:H and medium crystalline μc-Si:H relevant for applications, the difference in N_S between air-exposed and annealed conditions is smaller. ESR measurements performed at 40 K suggest that atmospheric exposure leads to charging of the defect states, which in turn influences the evaluated spin density.     

The nanostructural analysis of hydrogenated silicon films based on positron annihilation studies

Due to the complex nature of hydrogenated amorphous and microcrystalline silicon (a-Si:H; μc-Si:H) a profound understanding of the Si:H nanostructure and its relation to the Staebler–Wronski effect (SWE) is still lacking. In order to gain more insight into the nanostructure we present a detailed study on a set of Si:H samples with a wide variety of nanostructural properties, including dense up to porous films and amorphous up to highly crystalline films, using Doppler broadening positron annihilation spectroscopy (DB-PAS) and Fourier Transform infrared (FTIR) spectroscopy. The results obtained from these material characterisation techniques show that they are powerful complementary methods in the analysis of the Si:H nanostructure. Both techniques indicate that the dominant type of open volume deficiency in device grade a-Si:H seems to be the divacancy, which is in line with earlier positron annihilation lifetime spectroscopy (PALS), Doppler broadening (DB) PAS and FTIR studies.     

Electronic states in a-Si:H/c-Si heterostructures

We have investigated PECVD-deposited ultrathin intrinsic a-Si:H layers on c-Si substrates using UV-excited photoemission spectroscopy (hν = 4–8 eV) and surface photovoltage measurements. For samples deposited at 230 °C, the Urbach energy is minimal, the Fermi level closest to midgap and the interface recombination velocity has a minimum. The a-Si:H/c-Si interface density of states is comparable to that of thermally oxidized silicon interfaces. However, the measured a-Si:H dangling bond densities are generally higher than in thick films and not correlated with the Urbach energy. This is ascribed to additional disorder induced by the proximity of the a-Si:H/c-Si interface and H-rich growth in the film/substrate interface region.     

Influence of the amorphous/crystalline silicon heterostructure properties on planar conductance measurements

We report a quasi-analytical calculation describing the heterojunction between hydrogenated amorphous silicon (a-Si:H) and crystalline silicon (c-Si) at equilibrium. It has been developed and used to determine the carrier sheet density in the strongly inverted layer at the a-Si:H/ c-Si interface. The model assumes an exponential band tail for the defect distribution in a-Si:H. The effects of the different parameters involved in the calculation are investigated in detail, such as the Fermi level position in a-Si:H, the density of states and the band offsets. The calculation was used to interpret temperature dependent planar conductance measurements carried out on (n) a-Si:H/ (p) c-Si and (p) a-Si:H/(n) c-Si structures, which allowed us to confirm a previous evaluation of the conduction band offset, ?E_C = 0.18 ± 0.05 eV, and to evaluate the valence band offset: ?E_V = 0.36 ± 0.05 eV at the a-Si:H/ c-Si heterojunction. The results are placed in the frame of recent publications.     

Structure of the ESR spectra of thin film silicon after electron bombardment

Defect creation by MeV electron bombardment of a-Si:H and μc-Si:H thin films is used to explore hidden features of the electron spin resonance spectra. Different dynamics of creation and annealing for different paramagnetic states is expected and found. In a-Si:H the g-value of the db resonance does not change after irradiation, but a pair of satellites is observed on its wings. In the spectra of μc-Si:H three additional lines can be extracted after irradiation, overlapping with the central resonance. Careful analysis of the spectra shows also modification of the dangling bond resonance in μc-Si:H that is compatible with variations of two components of the spectra and supports the model of two dominant defect states in μc-Si:H.     

Inline deposited thin-film silicon solar cells on imprinted foil using linear PECVD sources

The standard way to improve the light management of thin film solar cells is to introduce a light scattering structure, either on the front window or at the back reflector. Usually, growth conditions of TCO layers are adjusted to get random surface roughness on the front window. In this paper we present an alternative method, which can be applied both on the front window and at the back reflector. It involves imprinting a UV curable coating layer allowing full control on the texture (random or periodic) to fully optimise the light trapping. Light trapping is even more important for microcrystalline Si solar cells. We have fabricated thin film nip Si solar cells with sputtered Ag/ZnO back contacts on embossed barrier layers on steel foil. We show that the UV curable coating is well-suited as imprintable barrier layer between the steel foil and the active layers. For nip a-Si cells we can obtain light trapping, as measured by the short-circuit current, that is almost as good as that of nip a-Si cells made on Asahi U-type glass, covered with a Ag/ZnO back reflector. Furthermore, we show that dynamically processed a-Si nip cells on foil realised efficiencies of over 7%, which are only slightly less than for cells made in a UHV lab-scale cluster tool in static processing. Finally, a-Si/a-Si tandems and μc-Si/a-Si tandems have been fabricated. Initial efficiencies of around 8% on textured barrier layer on steel foil have been achieved.     

An X-ray scattering study of hydrogenated nanocrystalline silicon with varying crystalline fraction

Using X-ray diffraction (XRD) and small angle X-ray scattering (SAXS), we probed the nanostructural features of several PECVD grown nc-Si:H thin films with varying crystalline volume fraction. XRD results of a mixed phase film, 70% a-Si:H and 30% c-Si:H, show these crystallites have a preferred [220] orientation in the growth direction. Another film with approximately 90% c-Si also shows elongated grains, but with a preferred [111] orientation. The SAXS results also show an increase in scattering intensity when compared to the mixed phase material. In the mixed phase material, models show that the electron density fluctuations between the amorphous and crystalline phases are not enough to explain the measured SAXS scattering. Hydrogen clustered at the crystallite boundaries and in void regions of the a-Si phase must be included as well.     

H NMR in μc-Si:H

NMR, IR, ESR, Raman scattering and X-ray diffraction measurements were performed in μc-Si:H prepared by various methods. Results of H NMR in some films are qualitatively similar to those i n a-Si:H, but the NMR lines exhibit a motional narrowing. Other films which exhibit sharp IR peaks exhibit H NMR signal shape different from that in a-Si:H.     

NMR study of μc-Si:H

The behavior of hydrogen in glow-discharge (GD) μc-Si:H has been characterized by ¹H NMR. The ¹H spectra consist of two components with different linewidths. The linewidth (FWHM) of the narrow component is about 0.5 kHz at ν₀ = 90 MHz, being much narrower than has been observed in GD a-Si:H deposited under a conventional low RF-power condition. It has been demonstrated that the 0.5-kHz FWHM component originates from the hydrogens in motional narrowing state, and such fast-moving hydrogens are incorporated both in μc-Si:H and high-power deposited a-Si:H.     

Universal critical exponents in percolation systems with tunneling

Some specific dependence of micro-crystalline solar cells efficiency on the crystallite size was reported recently [R. Beserman, A. Chak, R. Weil, T. Roschek, in: Proceedings of the 19th European Photovoltaic Solar Energy Conference, Paris, 2004, p. 1520]. The simple explanation of this dependence is suggested in the framework of percolation phenomena for non-equilibrium carriers. Percolation problem for non-equilibrium carriers was solved in continuous space with tunnel coupling between conducting grains. It was shown that local micro-geometry is not essential if the inter-grains transition time is less than the lifetime of the non-equilibrium carriers inside the conducting grains. In this case the main parameter of the problem is the lifetime of the non-equilibrium carriers in the conducting grains. A new bonding criterion was formulated. An amorphous hydrogenated silicon (a-Si:H) matrix with a big number of crystalline silicon (c-Si) inclusions was investigated as an example of the system. It was shown that universal dependence exists for non-equilibrium conductivity on the mean crystalline size. This prediction was verified by experiments with real solar cells produced from a-Si:H with c-Si inclusions.     

New approach to capacitance spectroscopy for interface characterization of a-Si:H/c-Si heterojunctions

A new technique for characterization of interface defects in a-Si:H/c-Si heterostructure solar cells from capacitance spectroscopy measurements under illumination at forward bias close to open-circuit voltage is described. The proposed method allows to significantly increase the sensitivity to interface defects compared to conventional capacitance measurements at zero or small negative bias. Results of numerical modelling as well as experimental data obtained on n-type a-Si:H/p-type c-Si heterojunctions are presented. The sensitivity of the proposed method to interface states and the influence of various parameters like band mismatch, density of interface defects, recombination velocity at the back contact are discussed.     

Variation of the defect density in a-Si:H and μc-Si:H based solar cells with 2 MeV electron bombardment

Improvement of the performance of solar cells based on amorphous (a-Si:H) and microcrystalline (μc-Si:H) silicon requires understanding of the role of the deep defects – dangling bonds – in the bulk of the intrinsic a-Si:H or μc-Si:H absorber layers. A straightforward way to understand how these defects may affect the performance of the cells is to investigate changes in the device performance upon variation in the defect density.In the present work solar cells with a-Si:H and μc-Si:H absorber layers were exposed to 2 MeV electron bombardment. The performance of the cells after various bombardment doses and annealing steps was evaluated in view of the changes in the defect density of intrinsic layers, measured with ESR on nominally identical absorber layers irradiated in parallel with the cells.The defect density was varied over a range of 2 orders of magnitude. In the solar cells a strong degradation of performance is observed upon irradiation with the biggest effect on the short circuit current density J_SC for both types of absorber layers. In most cases both V_OC and J_SC recover after the final annealing step (at 160 °C) for both types of cells.     

Er-doped hydrogenated amorphous silicon: structural and optical properties

The effect of erbium-doping on the structural and optical properties of hydrogenated amorphous silicon (a-Si:H) is investigated. Optical absorption and Raman spectra indicate that erbium doping introduces defect states, and that above a concentration of 0.27 at.%, induces strong structural disorder. The photoluminescence measurements show that erbium doping introduces non-radiative decay paths for carriers in a-Si:H, leading to decrease in both the Er³⁺ and intrinsic a-Si:H luminescence intensity when the Er concentration is increased to more than 0.04 at.%. The results are compared to that of Er-doped crystalline Si, and the possible excitation mechanisms of Er in a-Si:H are discussed.     

Network structure and dynamics of hydrogenated amorphous silicon

In this paper we discuss the application of current ab initio computer simulation techniques to hydrogenated amorphous silicon (a-Si:H). We begin by discussing thermal fluctuation in the number of coordination defects in the material, and its temperature dependence. We connect this to the ‘fluctuating bond-center detachment’ mechanism for liberating H bonded to Si atoms. Next, from extended thermal MD simulation, we illustrate various mechanisms of H motion. The dynamics of the lattice is then linked to the electrons, and we point out that the squared electron-lattice coupling (and the thermally-induced mean square variation in electron energy eigenvalues) is robustly proportional to the localization of the conjugate state, if localization is measured with inverse participation ratio. Finally we discuss the Staebler–Wronski effect using these methods, and argue that a sophisticated local heating picture (based upon reasonable calculations of the electron-lattice coupling and molecular dynamic simulation) explains significant aspects of the phenomenon.     

Order and disorder in edge-supported pure amorphous Si and pure amorphous Si on Si(001)

We report results from an investigation into hidden anisotropy in pure fully-dense amorphous silicon. For amorphous silicon in intimate contact with a crystalline Si(001) substrate, one can reasonably expect that the interface with the substrate may impose anisotropy in the form of distorted ordering within the film. Indeed, we found four-fold periodic intensity variations, with bimodal intensity centered along the substrate c-Si < 110> directions, in the X-ray scattering from a-Si on Si(001). These well-defined intensity variations disappeared entirely in X-ray scattering from edge-supported a-Si films, where there was no detectable anisotropy.     

X-ray emission spectra and electronic structure of amorphous silicon

X-ray K and L emission bands of c-Si and a-Si are reported and compared with available XPS and UPS measurements. The experimental results for a-Si are found to be consistent and in excellent agreement. From comparison of the experimental results with available electron density-of-states calculations based on different structure models of the amorphous state, we conclude that only ST-12 structure and the Polk-Boudreaux model provide results that are compatible with experiment.Furthermore we have studied the high energy L satellite band of c-Si and a-Si.     

Investigation of gap states in phosphorous-doped ultra-thin a-Si:H by near-UV photoelectron spectroscopy

A variant of photoelectron spectroscopy with near-UV light excitation was established and applied to an n-type doping series of ultra-thin a-Si:H layers (layer thickness ～10 nm). Using this technique, the position of the surface Fermi level E_Fs is obtained and the density of recombination active defect states in the a-Si:H band gap down to ～10¹⁵ states/cm³ can be detected. Defect densities are generally about one order of magnitude higher than in the bulk of thicker (several 100 nm) layers, and the minimum achievable distance of E_Fs from the conduction band is ～360 mV for doping with 10⁴ ppm PH₃. The optimum doping for the fabrication of solar cells is almost one order of magnitude lower. This discrepancy may be explained by enhanced recombination at the a-Si:H/c-Si interface at high doping levels, and in addition by an efficient recombination pathway where charge carriers tunnel from c-Si via a-Si:H band tail states into the a-Si:H and subsequently recombine at dangling bond states.