Showing papers by "Herman Maes published in 1998"
TL;DR: In this paper, a percolation-based model for intrinsic breakdown in thin oxide layers is proposed, which can explain the experimentally observed statistical features of the breakdown distribution, such as the increasing spread of the Q/sub BD/-distribution for ultrathin oxides.
Abstract: In this paper it is demonstrated in a wide stress field range that breakdown in thin oxide layers occurs as soon as a critical density of neutral electron traps in the oxide is reached. It is proven that this corresponds to a critical hole fluence, since a unique relationship between electron trap generation and hole fluence is found independent of stress field and oxide thickness. In this way literature models relating breakdown to hole fluence or to trap generation are linked. A new model for intrinsic breakdown, based on a percolation concept, is proposed. It is shown that this model can explain the experimentally observed statistical features of the breakdown distribution, such as the increasing spread of the Q/sub BD/-distribution for ultrathin oxides. An important consequence of this large spread is the strong area dependence of the Q/sub BD/ for ultrathin oxides.
600 citations
TL;DR: In this article, the authors investigated the field acceleration of intrinsic and extrinsic breakdown and proposed a new analytical expression for fitting competing Weibull distributions for low-field oxide reliability.
Abstract: The field acceleration of intrinsic and extrinsic breakdown is studied. For the intrinsic mode an exp(1/E)-acceleration law is found, while for the extrinsic mode an new exp (E)-acceleration law for Q/sub BD/ is proposed. This field acceleration model is implemented in a maximum likelihood algorithm together with a new analytical expression for fitting competing Weibull distributions. With this algorithm an extensive t/sub BD/-data set measured at different stress conditions can be fitted excellently in one single calculation. From the result, predictions of low-field oxide reliability are made and the screening conditions in order to guarantee a pre-set reliability specification are calculated.
125 citations
TL;DR: In this article, the authors studied soft breakdown in capacitors and nMOSFETs with 4.5-nm oxide thickness and showed that the SBD detection strongly depends on the test structure area.
Abstract: In this work we have studied soft breakdown (SBD) in capacitors and nMOSFET's with 4.5-nm oxide thickness. It is shown that for larger area devices gate current and substrate current as a function of the gate voltage after SBD are stable and unique curves, but for smaller area devices both currents become lower and unstable. This difference can be explained by the different energy available for discharging in the SBD path. It is shown that the SBD detection strongly depends on the test structure area. In nMOSFET's for positive gate polarity, the large increase in the substrate current at the SBD moment is proposed as a sensitive SBD detector. Two level fluctuations in the gate current are investigated at different voltages and are explained by means of a model where electron capture-emission in the traps of the SBD path induces local field fluctuations causing variations in the tunneling rate across the oxide. In the substrate current directly correlated two-level fluctuations are observed.
101 citations
TL;DR: In this paper, a quantitative model for the steady-state component of the stress induced leakage current (SILC) is developed based on the observation of basic degradation monitors on conventional, thermal SiO/sub 2/ gate dielectrics in the thickness range of 6.8-7.1 nm.
Abstract: In this paper a quantitative model for the steady-state component of the stress induced leakage current (SILC) is developed. The established model is based on the observation of basic degradation monitors on conventional, thermal SiO/sub 2/ gate dielectrics in the thickness range of 6.8-7.1 nm. From a systematic, experimental study, it has been found for the first time that the steady-state SILC, observed after a wide range of constant current stress (CCS) conditions (gate injection polarity), can be uniquely described by a simple, semi-empirical relation, which consists of two parts: 1) the dependence on the measurement field is described as Fowler-Nordheim (FN) tunneling through an oxide barrier of reduced but fixed height (i.e., 0.9 eV), and 2) the level of the SILC at a fixed oxide field is given by the density of neutral bulk oxide traps. Except for a calibration, depending on the oxide thickness and processing, no model parameters have to be adjusted in order to describe all our data. Also, based on bake experiments it has been concluded that interface traps are not causally related to the steady-state SILC in spite of the linear relation which exists between both. Furthermore, these bake experiments provide new evidence that bulk oxide traps play a crucial role in the SILC conduction mechanism.
84 citations
TL;DR: In this article, a method is described to calculate the Raman spectrum from a nonuniformly strained sample taking into account the effects that arise due to finite depth of penetration and diameter of the laser beam.
Abstract: A method is described to calculate the Raman spectrum from a nonuniformly strained sample taking into account the effects that arise due to finite depth of penetration and diameter of the laser beam. Both the parallel and the focused beams are considered. The case of stress in a Si substrate decaying monotonically with depth z (rapidly near the interface and slowly at larger depths) is considered in detail. The predicted Raman shifts are found to be sensitive to both the distribution of stress and to the absorption coefficient α for the laser light wavelength used. It is found that light scattered from distances much larger than 1/α still contribute significantly to the observed Raman spectrum. The observed shift in the peak of the spectrum does not correspond to the stress close to the interface. If the stress decays more rapidly than the light intensity, the Raman line that originates from the unstrained lower part of the substrate dominates. For transparent material (α=0) and unfocused beam the Raman spectrum consists of only the unstrained Si line; the contribution to Raman line from the strained interface region is completely masked. For measurements of stresses near the interface short wavelength light with an absorption depth of 5–10 nm is recommended. The calculated and observed Raman shifts in a local oxidation of silicon (a processing technique for isolation) with polysilicon buffer between the nitride stripe and the Si substrate are compared. The agreement between the calculated and the observed Raman shifts is very good. The salient points of our approach which enabled us to obtain this agreement are: We took into account the effects of laser beam width, penetration depth, and focusing; we included the stresses in the polysilicon layer and near the polysilicon/silicon interface, and we included contributions from large depths.
34 citations
TL;DR: In this article, the authors proposed an interesting alternative route for further voltage scaling by optimizing the composition and stoichiometry of PZT thin-films in order to achieve less than 1.5V switching.
Abstract: For the possible application of PZT thin films in ferroelectric non-volatile memories, the potential of low-voltage switching is a major requirement. By straightforward thickness scaling down to 75nm, operation voltage can be reduced to 1.5V without degradation of the hysteresis properties. However, interface effects increase the coercive field for thinner films, so that operation voltage does not scale linearly with thickness. In view of the large difference in coercive field between bulk and thin-films, on the other hand, material (and interface) optimization may constitute an interesting alternative route for further voltage scaling. By optimizing the composition and stoichiometry, less than 1.5V switching has indeed been obtained even for 150nm films. It is argued that an important material parameter affecting low-voltage switching behavior is the ferroelectric domain configuration and presence/absence of mobile domain walls in the PZT film.
33 citations
01 Jan 1998-Microsystem Technologies-micro-and Nanosystems-information Storage and Processing Systems
TL;DR: The application of micro-Raman spectroscopy for measurements of mechanical stress in silicon microelectronics devices is discussed in this paper, where the advantages and disadvantages of the technique are shown through different examples such as Si3N4 and metal lines, isolation structures and solder bumps.
Abstract: The application of micro-Raman spectroscopy for measurements of mechanical stress in silicon microelectronics devices is discussed. The advantages and disadvantages of the technique are shown through different examples such as Si3N4 and metal lines, isolation structures and solder bumps.
29 citations
TL;DR: In this paper, a review of the principles and various applications of charge pumping in submicron MOSFETs is made, and the role and detection of so-called geometric components is illustrated.
20 citations
TL;DR: In this paper, the authors reported results of the finite element (FE) calculations of strains and stresses in stripe-substrate (S-S) heterostructures for several values of Young's moduli (for the stripe and for the substrate), stripe halfwidth l and its thickness h.
Abstract: We report results of the finite element (FE) calculations of strains and stresses in stripe-substrate (S-S) heterostructures for several values of Young's moduli ( for the stripe and for the substrate), stripe halfwidth l and its thickness h. The normalized stress ( is the stress in the large area layer where effect of edge induced relaxation is absent) in both the stripe and the substrate depends only on the ratio and not on the individual values of and . Numerical values of are calculated by the FE method for nine values of and five values of . It is found that has only a weak influence on the stress distribution. in both the stripes and the substrates decreases monotonically and approximately linearly as increases. The values reported in this paper can be used for other values of , l and h by interpolation without the need of making fresh FE calculations for each case. Experimental values of the stresses determined from the luminescence and the micro-Raman data support strongly the conclusion that the dependence of stress on is weak.
12 citations
TL;DR: In this article, the parasitic bipolar transistor inherent to grounded gate nMOS transistors is modelled, accounting for the specific conditions applied by CDM ESD stress, and the optimal gate length for CDM protection in advanced submicron technologies is discussed.
9 citations
TL;DR: In this paper, the authors reviewed the techniques and methodologies to be applied in R&D laboratories for the assessment of thin gate dielectrics reliability and hot carrier degradation and how the application of these techniques allows to obtain a better insight in the physics of the degradation process.
TL;DR: In this article, the backward adaptive conservative methodology is step by step replaced by a methodology considering the ESD issue already during process development and involving more predictive ESD-TCAD into the development cycle.
TL;DR: In this paper, a fast identification of I ddq failures using spectroscopic photon emission microscopy (SPEMMI) is proposed, where the spectra obtained from failure sites on the failed chips were compared with the ones of known defective components.
TL;DR: In this paper, a modified sol-gel process was developed for the preparation of precursor solutions of undoped, La and Ta doped PZT (25/75), and a rectangular hysteresis loop, featuring high Pr, was obtained for one specific precursor chemistry.
Abstract: Modified sol-gel processes have been developed for the preparation of precursor solutions of undoped, La and Ta doped PZT (25/75). These processes use different solvents (methoxyethanol, butoxyethanol) and different lead sources (lead acetate, lead oxide). Due to variations in the structure and in the composition of the solutions, significantly different thermal decomposition behaviors were found. These inevitably affected texture and microstructure of the sol-gel derived thin films, demonstrating the important role of precursor chemistry in the improvement of film properties. In the case of tantalum doped Pt/PZT/Pt ferroelectric capacitors a rectangular hysteresis loop, featuring high Pr, was obtained for one specific precursor chemistry.
TL;DR: In this article, the use of thin (∼5 nm) high-Ti PZT seed layer was found to be effective for improving fatigue performance of sol-gel PZTs based ferroelectric capacitors with RuO2 electrodes.
Abstract: New approaches for improving fatigue performance of sol-gel PZT based ferroelectric capacitors with RuO2 electrodes are discussed. The use of thin (∼5 nm) high-Ti PZT seed layer was found to be effective for improving fatigue performance. For an optimized seedlayer thickness, FECAPs with excellent fatigue characteristics (less than 10% decrease in Pr after 1011 cycles) were obtained.
Patent•
24 Nov 1998TL;DR: In this paper, a contactless 5 V-only Flash EEPROM array configuration that relies on shared program lines is presented, in order to minimize the area overhead that is caused by this program gate.
Abstract: The present invention relates to the field of electrically erasable and programmable nonvolatile semiconductor memories (EEPROM) and, in particular, to contactless array configurations that are used for the practical and efficient implementation of a particular type of memory transistor. Such a memory transistor allows fast 5 V-only programming by the use of an enhanced source-side injection mechanism. This concept requires a program gate in the field oxide region which serves to capacitively couple a high voltage to the floating gates. Thus, a very high injection current is established during programming. This additional program gate, however, increases the cell area considerably. The present disclosure shows a contactless 5 V-only Flash EEPROM array configuration that relies on shared program lines in order to minimize the area overhead that is caused by this program gate. Furthermore, a memory array with shared wordlines is presented which further enhances the density achievable.
TL;DR: In this paper, the switching quality of PZT thin-film ferroelectric capacitors was studied, using hysteresis and pulse switching measurements, and the results proved the very high switching quality required for their application as nonvolatile memory elements.
Abstract: The switching quality of PZT thin-film ferroelectric capacitors was studied, using hysteresis and pulse switching measurements. The ferroelectric properties of epitaxial Pt/PZT/LSCO/MgO capacitors are strongly affected by the LSCO material. A strong deviation between hysteresis and pulse measurements has been observed for one type of LSCO, corresponding to a strong relaxation of (oxygen-vacancy) space charge polarization. On the other hand, polycrystalline Pt/PZT/Pt show very rectangular hysteresis loops with high P, and pulse measurements reveal high speed switching and absence of rapid depolarization effects. These results prove the very high switching quality of Pt/PZT/Pt capacitors, required for their application as non-volatile memory elements.
TL;DR: In this paper, a model for thin-film PZT ferroelectric capacitors is presented and an expression for the switching current as a function of instantaneous values of polarization and voltage is proposed.
Abstract: A model for thin film PZT ferroelectric capacitors is presented. An expression for the switching current as a function of instantaneous values of polarization and voltage, that is not explicitly a function of time is proposed. We call this a time independent model. Hysteresis, non-switching and switching current responses of these capacitors have been simulated, showing a good agreement with the experimental data.
01 Jan 1998
TL;DR: In this paper, the authors studied the relationship between the statistics and the mechanisms of intrinsic oxide breakdown for ultra thin oxides between 12 nm and 2 nm and showed that the statistical spread of the intrinsic TDDB-distribution, which is quantified by the slope of the Weibull distribution, depends on the oxide thickness.
Abstract: In this paper new insights in the relationship between the statistics and the mechanisms of intrinsic oxide breakdown for ultra thin oxides between 12 nm and 2 nm are discussed. It is shown that the statistical spread of the intrinsic TDDB-distribution, which is quantified by the slope of the Weibull distribution /spl beta/, depends on the oxide thickness. A large increase in the spread on the distribution is found for decreasing oxide thicknesses. This increased spread is not determined by larger statistical variations of process parameters, such as e.g. oxide thickness uniformity, but by the nature of the breakdown mechanism itself. A statistical percolation model is presented, and it is shown that this model can explain the experimental statistical features of the breakdown distributions, which were not yet understood before, Next, it is demonstrated that for ultra-thin oxides the conventional interpretation of constant current Q/sub BD/ to evaluate tile influence of process variations on the reliability of MOS-structures can lead to erroneous conclusions. For a fixed thickness, the comparison of Q/sub BD/. Distributions of all processes that affect the breakdown statistics becomes even meaningless. For ultra-thin oxides, the impact of different processing conditions therefore requires constant gate voltage instead of constant current density Q/sub BD/ tests.
01 Jan 1998
TL;DR: In this article, micro-Raman spectroscopy and emission microscopy are used to study the crystallographic phase of 0.25 /spl mu/m wide TiSi/sub 2/ lines.
Abstract: Micro-Raman spectroscopy and emission microscopy are used to study the crystallographic phase of 0.25 /spl mu/m wide TiSi/sub 2/ lines. It is shown that these techniques allow nondestructive mapping of the local phase of TiSi/sub 2/. The results show that there is a direct correlation between the resistance variation of these lines and the local occurrence of the high resistivity C49 phase of TiSi/sub 2/ in the lines.
TL;DR: In this paper, PZT and PLZT precursor solutions were prepared by a solgel route, varying the Zr/Ti ratio and the La doping level, and they were characterized by TGA coupled with FT-IR.
Abstract: PZT and PLZT precursor solutions were prepared by a sol-gel route, varying the Zr/Ti ratio and the La doping level. They were characterized by TGA coupled with FT-IR. Reactions taking place during heat treatments are discussed. Composition of PZT and PLZT thin films was measured by RBS and their orientation determined by XRD. Ferroelectric capacitors with Pt electrodes were characterized by hysteresis and fatigue measurements.
TL;DR: In this article, critical thickness (hc) is calculated for capped and uncapped lattice mismatched II-VIsemiconductor epilayers and the observed values of hc are larger than the calculated values.
Abstract: Critical thickness (hc) is calculated for capped and uncapped lattice mismatched II-VIsemiconductor epilayers. Both the old equilibrium theory and the improved theory have been used.The calculated values are compared with the experimental data on epilayers of several II-VIsemiconductors and alloys. The observed values of hc are larger than the calculated values. Howeverthe discrepancy is much smaller than that found in InGaAs/GaAs and GeSilSi layers. Moreover ascompared to InGaAs/GaA.s:a nd GeSilSi layers, the experimental data show a much smaller scatter andcan be fitted with one curve. Strain relaxation in layers with thickness h > hc is also calculated. Strainrelaxation in ZnSe layers grown on (100) GaAs shows good agreement with the equilibrium theory. Inother cases the observed relaxation is sluggish, the residual strain is larger than its calculated value.Thick highly mismatched layers behave differently. The residual strain agrees with theory anddislocations are distributed periodically, A model to interpret these observations is suggested.Implications of this study on the stability of 11V- I strained layers are discussed.
01 Nov 1998
TL;DR: In this article, it was shown that two dominant defect levels with activation energy E v + 0.19 eV and E v+ 0.36 eV (C s C i or C i O i complex) are created in the lower part of the bandgap after low energy electron irradiation of boron doped float zone or Czochralski silicon.
Abstract: Two dominant defect levels with activation energy E v + 0.19 eV and E v + 0.36 eV (C s C i or C i O i complex) are created in the lower part of the bandgap after low energy electron irradiation of boron doped float zone or Czochralski silicon. The E v + 0.19 eV level is associated to the donor state of the divacancy (V 0/+ 2 ), since the annealing behaviour of this level in the temperature range between 250 and 400 C is in good agreement with the behaviour reported for the divacancy in silicon. However, there still exists some controversy in the literature about the exact position of the donor level V 0/+ 2 for which values are reported at around E v + 0.20 eV or E v + 0.24 eV. This paper also reports on the transformation of the E v + 0.19 eV level appearing immediately after electron irradiation in p type silicon to the E 1 + 0.24 eV level after annealing at low temperatures for several days.