Engineering

Publication Search Results

Now showing 1 - 10 of 107
  • (1998) Bradley, Peter; Rozenfeld, Anatoly; Lee, Kevin; Jamieson, Dana; Heiser, Gernot; Satoh, S
    Journal Article
    The first results obtained using a SOI device for microdosimetry applications are presented. Microbeam and broadbeam spectroscopy methods are used for determining minority carrier lifetime and radiation damage constants. A spectroscopy model is presented which includes the majority of effects that impact spectral resolution. Charge collection statistics were found to substantially affect spectral resolution. Lateral diffusion effects significantly complicate charge collection

  • (1998) Heiser, Gernot; Elphinstone, Kevin; Vochteloo, J; Russell, Stephen; Liedtke, Jochen
    Journal Article
    Single-address-space operating systems (SASOS) are an attractive model for making the best use of the wide address space provided by the latest generations of microprocessors. SASOS remove the address space boundaries which make data sharing between processes difficult and expensive in traditional operating systems. They offer the potential of significant performance advantages for applications where sharing is important, such as object-oriented databases or persistent programming systems. We have built the Mungi system to demonstrate that a SASOS can offer these performance advantages without resorting to special hardware. Mungi is a very `pure` SASOS, featuring an unintrusive protection model based on sparse capabilities, a fast protected procedure call mechanism, and uses shared memory as the exclusive inter-process communication mechanism, as well as for I/O. The simplicity of our model makes it easy to implement it efficiently on conventional architectures. Our implementation of Mungi for the MIPS R4600 64-bit microprocessor is presented, which is based on our port of the L4 microkernel. Mungi is shown to outperform, in some instances by more than an order of magnitude, two UNIX operating systems, Irix and Linux, in several important operations, such as task creation and inter-process communications, and on the OO1 object-oriented database benchmark. As well, we describe how our approach to key issues in SASOS design provides better performance than other systems, such as Opal. Our experience shows that the SASOS concept is viable, and that a well-designed microkernel is an excellent base on which to build high-performance operating systems.

  • (1997) Altermatt, Pietro; Schmidt, Jan; Heiser, Gernot; Aberle, Armin
    Journal Article
    In traditional band-to-band Auger recombination theory, the low-injection carrier lifetime is an inverse quadratic function of the doping density. However, for doping densities below about 3E18cm^-3, the low-injection Auger lifetimes measured in the past on silicon were significantly smaller than predicted by this theory. Recently, a new theory has been developed [A. Hangleiter and R. H¿cker, Phys. Rev. Lett. 65, 215 (1990)] which attributes these deviations to Coulombic interactions between mobile charge carriers. This theory has been supported experimentally to a high degree of accuracy in n-type silicon, however, no satisfactory support has been found in p-type silicon for doping densities below 3E17cm^-3. In this work, we investigate the most recent lifetime measurements of crystalline silicon and support experimentally the Coulomb-enhanced Auger theory in p-type silicon in the doping range down to 1E16cm^-3. Based on the experimental data, we present an empirical parameterisation of the low-injection Auger lifetime. This parameterisation is valid in n- and p-type silicon with arbitrary doping concentrations and for temperatures between 70 and 400K. We implement this parameterisation into a numerical device simulator to demonstrate how the new Auger limit influences the open-circuit voltage capability of silicon solar cells. Furthermore, we briefly discuss why the Auger recombination rates are less enhanced under high-injection conditions than under low-injection conditions.

  • (1997) Schenk, Andreas; Heiser, Gernot
    Journal Article
    Direct and Fowler-Nordheim tunneling through ultra-thin gate dielectrics is modeled based on a new approach for the transmission coefficient (TC) of a potential barrier which is modified by the image force. Under the constraint of equal actions the true barrier is mapped to a trapezoidal pseudobarrier resulting in a TC very close to the numerical solution of the Schrodinger equation for all insulator thicknesses and for all energies of the tunneling electron. The barrier height of the pseudopotential is used as a free parameter and becomes a function of energy in balancing the actions. This function can be approximated by a parabolic relation which makes the TC of arbitrary barriers fully analytical with little loss of accuracy. The model was implemented into a multidimensional device simulator and applied to the self-consistent simulation of gate currents in metal-oxide-semiconductor (MOS) capacitors with gate oxides in the thickness range 15A-42A. Excellent agreement with experimental data was obtained using a thickness-independent tunnel mass mox=0.42m0. Thanks to the CPU-time efficiency of the method the simulation of a complete MOS-field-effect-transistor with dominating gate current becomes possible and shows the potential for further applications.

  • (1996) Wenham, Stuart; Green, Martin; Edminston, Sean; Campbell, Patrick; Koschier, L; Thorpe, D; Honsberg, Christiana; Shi, Z; Heiser, Gernot; Sproul, Alistair
    Journal Article
    Thin film crystalline silicon solar cells can only achieve high efficiencies if light-trapping can be used to give a long optical path lengtrh, while simulatneously achieving near unity collection probabilities for all generated carriers. This necessitates a supporting substrate of a foreign material, with refractive index compatible with light trapping schemes for silicon. The resulting inability to nucleate growth of crystalline silicon films of good crystallographic quality on such foreign substrates, at present prevents the achievement of high efficiecny devices using conventional single junction solar cell structures. The parallel multijunction solar cell preovides a new approach for achieving high efficiencies from very poor quality material, with near unity collection probabilities for all generated carriers achieved through appropriae junction spacing. Heavy doping is used to minimise the dark saturation current contribution from the layers, therefore allowing respectable voltages. The design strategy, corresponding advantages, theoretical predictions and experimental results are presented.

  • (1996) Edminston, Sean; Heiser, Gernot; Sproul, Alistair; Green, Martin
    Journal Article
    This paper provides a theoretical investigation of recombination at grain boundaries in both bulk and p-n junction regions of silicon solar cells. Previous models of grain boundaries and grain boundary properties are reviewed. A two dimensional numerical model of grain boundary recombination is presented in this paper. This numerical model is compared to existing analytical models of grain boundary recombination within both bulk and p-n junction regions of silicon solar cells. This analysis shows that, under some conditions, existing models poorly predict the recombination current at grain boundaries. Within bulk regions of a device, the effective surface recombination velocity at grain boundaries is overestimated in cases where the region around the grain boundary is not fully depleted of majority carriers. For vertical grain boundaries (columnar grains), existing models are shown to underestimate the recombination current within p-n junction depletion regions. This current has an ideality factor of about 1.8. An improved analytical model for grain boundary recombination within the p-n junction depletion region is presented. This model considers the effect of the grain boundary charge on the electric field within the p-n junction depletion region. The grain boundary charge reduces the p-n junction electric field, at the grain boundary, enhancing recombination in this region. This model is in agreement with the numerical results over a wide range of grain boundary recombination rates. In extreme cases, however, the region of enhanced, high ideality factor recombination can extend well outside the p-n junction depletion region. This leads to a breakdown of analytical models for both bulk and p-n junction recombination, necessitating the use of the numerical model.

  • (1996) Altermatt, Peter; Heiser, Gernot; Green, Martin
    Journal Article
    This paper presents a quantitative analysis of perimeter losses in high-efficiency silicon solar cells. A new method of numerical modelling is used, which provides the means to simulate a full-sized solar cell, including its perimeter region. We analyse the reduction in efficiency due to perimeter losses as a function of the distance between the active cell area and the cut edge. It is shown how the optimum distance depends on whether the cells in the panel are shingled or not. The simulations also indicate that passivating the cut-face with a thermal oxide does not increase cell efficiency substantially. Therefore, doping schemes for the perimeter domain are suggested in order to increase efficiency levels above present standards. Finally, perimeter effects in cells that remain embedded in the wafer during the efficiency measurement are outlined.

  • (1996) Altermatt, Peter; Heiser, Gernot; Dai, Ximing; Jurgens, J; Aberle, Armin; Robinson, Steven J.; Young, Timothy; Wenham, Stuart; Green, Martin
    Journal Article
    The passivated emitter, rear locally diffused (PERL) cells, fabricated in our laboratory, reach an efficiency of 24.0%, the highest value for any silicon-based solar cell under terrestrial illumination. In an attempt to improve the rear surface passivation, which is usually obtained by a thermally grown oxide, we add a floating (i.e., noncontacted) p-n junction at the rear surface, resulting in the passivated emitter, rear floating p-n junction (PERF) cell design. Although these cells exhibit record 1-sun open-circuit voltages of up to 720 mV, their efficiency is degraded by nonlinearities ("shoulders") in the logarithmic I-V curves. In order to understand and manipulate such nonlinearities, this paper presents a detailed investigation of the internal operation of PERF cells by means of numerical modelling based on experimentally determined device parameters. From the model, we derive design rules for optimum cell performance and develop a generalized argumentation that is suitable to compare the passivation properties of different surface structures. For example, the oxidized rear surface of the PERL cell is treated as an electrostatically induced floating junction in this approach and analogies to the diffused floating p-n junction are drawn. Our simulations indicate that optimum rear surface passivation can be obtained in three different ways. (i) The floating junction of the PERF cell should be very lightly doped, resulting in a sheet resistivity of 5000 Omega/[D'Alembertian], and losses due to shunt leaking paths between the p-n junction and the rear metal contacts must be avoided. (ii) The rear surface of the PERL cell should be passivated by chemical vapor deposition of a silicon nitride film containing a larger positive interface charge density than exists in thermally grown oxides. (iii) An external gate can be added at the rear with low leakage currents and gate voltages of around 15 V.

  • (1996) Altermatt, Peter; Heiser, Gernot; Aberle, Armin; Wang, Alan; Zhao, Jun; Robinson, J; Bowden, Simon; Green, Martin
    Journal Article
    This paper presents an improved method for measuring the total lumped series resistance (Rs) of high-efficiency solar cells. Since this method greatly minimizes the influence of non-linear recombination processes on the measured Rs values, it is possible to determine Rs as a function of external current density over a wide range of illumination levels with a significantly improved level of accuracy. This paper furthermore explains how resistive losses in the emitter, the base, the metal/silicon contacts and the front metal grid can be separately determined by combining measurements and multidimensional numerical simulations. A novel combination of device simulation and circuit simulation is introduced in order to simulate complete 2 × 2 cm2 PERL (passivated emitter and rear locally-diffused) silicon solar cells. These computer simulations provide improved insight into the dynamics of resistive losses, and thus allow new strategies for the optimization of resistive losses to be developed. The predictions have been experimentally verified with PERL cells, whose resistive losses were reduced to approximately half of their previous values, contributing to a new efficiency world record (24.0%) for silicon solar cells under terrestrial illumination. The measurement techniques and optimization strategies presented here can be applied to most other types of solar cells, and to materials other than silicon.

  • (1995) Heiser, Gernot; Aberle, Armin; Wenham, Stuart; Green, Martin
    Journal Article
    This paper reports on the first use of two-dimensional (2D) device simulation for optimising the front-finger spacing of one-sun high-efficiency silicon solar cells of emph{practical} dimensions. We examine the 2D current flow patterns in these devices under various illumination conditions, resulting in improved insight into the operating conditions of the cells. Results for the optimal spacing of the front metal fingers are presented and compared to predictions obtained from 1D simulations. We also address difficulties facing the numerical modelling of high-efficiency silicon solar cells.