Multicrystalline Silicon

Abstract

Abstract People Publications Projects Cooperations

S. Christiansen, M. Becker, and C. Himcinschi

Micro (µ)-Raman spectroscopy is a method that has been used recently in solid state physics to investigate mechanical stresses in semiconductor materials. µ-Raman spectroscopy detects strains and stresses in crystals via an inelastic interaction of laser light with lattice vibrations (phonons). The method probes the material of interest nondestructively without requiring any complex sample preparations. µ-Raman spectroscopy allows for a lateral resolution of the order of ~ 600 nm – 1µm when focusing the incident light beam on the sample surface. Raman signals can be mapped and respective stress mappings can be determined. Meanwhile, the determination of several stress-tensor components in crystals with diamond structure have already been shown. But the stress/strain analysis has been carried out only for monocrystalline samples with known orientation. Similar stress-tensor determination have not been shown for polycrystalline materials with diamond structure e.g. polycrystalline silicon wafers for solar cells or thin film transistors (TFT´s) due to the fact that grain orientation variations make stress measurements increasingly difficult. We developed a method to determine several stress-tensor components with µ-Raman spectroscopy in polycrystalline silicon within grains of arbitrary orientations. A single measurement cycle provides the local determination of orientation and stress/strain tensor components when making use of polarized incident light and detection of the polarization of the scattered light. An examples of practical interest is shown in Fig. 1.

Fig. 1: Light optical micrograph (left) of a surface of a polycrystalline silicon wafer for photovoltaic applications. The micrograph shows a grain boundary which separates a region of twin lamellas from an arbitrarily oriented grain. Close to the grain boundary fan-shaped dislocation clusters are found. A  mapping (right) of the Raman frequency shifts reveals a stress field close to the grain boundary which seems to correlate with the dislocation clusters. A detailed analysis of the Raman frequency shifts, taking into account the crystal orientation, yields the following difference stress tensor (white arrow; red color depicts relative compressive stresses):

Further reading:

M. Becker, H. Scheel, S. Christiansen, H.P. Strunk, J. Appl. Phys. in press (2007)

Silicon Nanowires

The transition to a sustainable global energy system is one of the biggest challenges mankind has ever faced. This transition process involves a huge financial investment and a strong political commitment. In this context, photovoltaic is a key technology.

A variety of photovoltaic technologies and conversion concepts are the subject of intensive research in and outside of Europe. They are all aimed at low cost, high efficiency or the combination of the two. New technologies e.g. thin film semiconductor cells on glass are at various stages of development, most of them however, still require fundamental research as does the nanowire based solar cell, which is based on nanowires for energy conversion. A key factor is reducing the cost of the modules, which is connected to consider the manufacturing processes used. In this context, there is considerable interest in replacing single crystalline and polycrystalline semiconductors by semiconductor thin layers, which are even nano-structured and can be applied cheaply e.g. on glass.

Fig. 2: Schematic cross section of a semiconductor nanowire based thin film solar cell on glass [1,2]. B and C show sequences of the integration process. Each wire is symbolized as shown in A with ‘symbol’ and can either be of type 1 or 2. ‘1’: a pn-junction between substrate and nanowire; the gold-dot serves as contact; ‘2’: a pn-junction is radial and a contact is wrapped around the wire in form of a conformal TCO layer. To get out the current from the solar cell, a screen printed electrode is realized as shown in C.

Table 1 summarizes the state of the art of achievements in terms of efficiencies of photovoltaic materials. Nanowire based solar cells will target the 10% efficiency line. This seems to be doable since nanowires themselves have an inherent light trapping geometry and the absorber material, the nanowire itself, is single crystalline (silicon, GaAs, GaN) and defect free. The nanowire solar cell, consisting of radial pn-junction nanowires at a high density is shown in Fig. 2. A nanowire with a pn-junction in the radial direction enables a decoupling of the requirements for light absorption and carrier extraction into orthogonal spatial directions. Each individual nanowire could be long in the direction of incident light to allow for maximized light absorption, but thin in diameter thereby allowing for effective carrier collection (no long diffusion distances). The beauty of using the nanowires lies in the guided self-assembly bottom-up nanowire synthesis process based on vapor-liquid solid (VLS) growth, which allows to prepare nanowires at high density in a highly parallel fashion and thus very fast and inexpensive, with the potential to scale up to large area processing. For inexpensive processing speaks the fact that integration on glass is possible (see Fig. 2) without having to deal with highly defective low quality material.

To enable this ‘bottom-up’ approach for nano-solar cells on glass there are three key research areas, which are at the heart of materials development and device integration. And these key research areas also apply for the two other applications. At first silicon, gallium arsenide or gallium nitride nanowires need to be grown with adjustable size (diameter, length), morphology, doping and if needed location control. Secondly, it is crucial to explore device integration aspects of nanowires such as radial pn-junction deposition, embedding nanowires in dielectric layers, chemo-mechanical polishing of embedded wires and contacting those wires. The embedding of the nanowires is realized by ALD with Al₂O₃, SiO₂ or HfO₂ as potential candidates. The filling up of the entire space between the nanowires is realized by a gap filling PECVD process or even by a spin on glass. Thirdly, the building blocks for a nanowire based solar cell and the building blocks in the system have to be characterized (electrically, optically  and structurally), i.e. with a metrology that works on atomic scale, and have to be optimized.

First attempts of nanowire growth on multicrystalline silicon on glass were successfully carried out (cf. Fig. 4).

Fig. 5: Silicon nanowires by the VLS mechanism on glass by CVD [3].

[1] B.M. Kayes, H.A. Atwater, N.S. Lewis, J. Appl. Phys. 97, 114302 (2005)
[2] G. Andrä, F. Falk, S.H. Christiansen, patent application pending
[3] Th. Stelzner, G. Andrä, F. Falk, S.H. Christiansen, J. Schneider,  to be published.