Refine
Document Type
- Doctoral thesis (22)
Has Fulltext
- yes (22)
Is part of the Bibliography
- no (22)
Year of publication
Keywords
- Silicium (22) (remove)
Institute
- FG Experimentalphysik und funktionale Materialien (8)
- FG Angewandte Physik und Halbleiterspektroskopie (3)
- FG Anorganische Chemie (1)
- FG Computational Physics (1)
- FG Halbleitermaterialien (1)
- FG Mikro- und Nanosysteme (1)
- FG Mikroelektronik (1)
- FG Schaltkreisentwurf (1)
- FG Statistische Physik und Nichtlineare Dynamik (1)
- FG Systeme (1)
Within this work, the growth of out-of-plane Si and SixGe1-x and in-plane Ge nanowire have been investigated. For this purpose, nanowires have been grown ccording to the vapor-liquid-solid mechanism by means of molecular beam epitaxy with Au as metallic solvent on Si(111) and nano-structured Si(001)/SiO2 substrates. Each stage from surface preparation to final nanowire growth has been examined in order to determine relevant parameters which influence the nucleation process nanowire elongation. Particular attention has been given to describe the variation of Au droplet size distribution on Si(111) to evaluate the subsequent nanowire growth and to obtain optimal growth conditions for an site-selective nucleation on the nano-structured substrates.
Due to increased surface diffusion velocity with raising temperature, the droplet diameter distribution shifts from a positive skewed distribution with a high fraction of smaller droplets to a negative skewed distribution with a high fraction of larger droplets. The temperature dependency of the most probable droplet diameter and the number of Au droplets per area has been determined, which can be applied to predict the resulting nanowire diameters.
The out-of-plane Si and SixGe1-x on Si(111) predominantly grow along the surface normal, exhibiting the characteristic sawtooth-like sidewall faceting. During the nanowire formation, the Au surface diffusion velocity increased and causes an increase of the most frequently observed diameter. Furthermore, the total number of droplets/nanowires decreases by a constant factor due to an increase of the contact angle during Si/Ge deposition. The specific diameter range for an possible nanowire formation is reduced by the incorporation of Ge into SixGe1-x nanowires.
To obtain a regular and uniform nanowire growth, a nano-structured substrate consisting of Si(001) terminated pillars surrounded by a SiO2 matrix has been utilized. The initial growth of Ge nanowire starting from Si-Au droplets with SixGe1-x nucleation from ternary alloy is discussed from a thermodynamic point of view and a model based on the Si-Ge-Au ternary phase diagram has been developed to predict the SixGe1-x concentration gradient in the nanowire base. The fully relaxed in-plane Ge nanowires occur within one of the four distinct in-plane ⟨110⟩ directions and nanowires are mainly bounded by two 55° inclined {111} facets and a less pronounced planar (001) top facet.
High-resolution scanning X-ray diffraction microscopy reveals a slightly tilted growth of individual nanowires with respect to each other, causing an abrupt change in the orientation at junction points of interconnected nanowires.
In dieser Arbeit werden neuartige Mikropumpen auf Siliziumbasis vom Typ oszillierende-Verdrängerpumpe mit einem in der Chipebene operierenden elektrostatischen Antrieb entwickelt, gefertigt und messtechnisch charakterisiert. Der Antriebsmechanismus ist aus einer am Fraunhofer IPMS entwickelten Klasse von Mikroaktoren, bekannt als nanoskopisch elektrostatischer Antrieb (NED) aufgebaut. Zum Einsatz kommen erstmals beidseitig eingespannte laterale NED-Balkenaktoren, primär entwickelt und optimiert für den quasi-statischen Betrieb.
Eingebettet in einem Verbund aus drei Siliziumwafern und in Verbindung mit in der Ebene operierenden passiven Klappenventilen, konnte ein regelbarer Fluss in Luft als gasförmiges sowie in Novec7100 als flüssiges Fördermedium nachgewiesen werden.
Die Entwicklung der Mikropumpen bildet einen Schwerpunkt dieser Arbeit und ist in zwei Abschnitte unterteilt. Eingangs werden Effekte, welche den Betrieb der NED-Balkenaktoren limitieren untersucht sowie Schlüsselparameter und deren Auswirkung identifziert. Basierend auf der Wahl der Fördermedien und den sich daraus ableitenden elektrotechnischen Randbedingungen wurde die Steuerspannung in Luft auf 400 V festgelegt und per Optimierungsverfahren ein Aktordesignpunkt bestimmt.
Der zweite Entwicklungsschwerpunkt beinhaltet den Entwurf und die Modellbildung der Mikropumpen mit dem bestimmten Aktordesignpunkt. Hier wurde ein modularer Ansatz zum Aufbau der Pumpe eruiert. Mit sogenannten "Basis-Pumpeinheiten" werden verschiedene Konfigurationen realisiert, was zu unterschiedlichen Pumpeigenschaften führt. Diese beinhalten unter anderem serielle Anordnungen von NED-Aktoren und Basiseinheiten sowie die parallele Stapelung der Balkenaktoren. Damit unterscheidet sich der laterale Pumpenansatz in seiner Konfigurierbarkeit grundlegend zu klassischen Mikromembranpumpen.
Die Etablierung einer Herstellungstechnologie mit stabilem Fertigungsprozess führte zur erfolgreichen monolithischen Herstellung von Pumpenchips im Scheibenverbund. Der Prozess ist CMOS-kompatibel und beinhaltet verschiedene Technologiemodule.
In der fluidischen Bauteilcharakterisierung wurden je nach Anzahl der gestapelten NEDAktoren maximale Flussraten von 6.63 µL/min, 22.85 µL/min, 63.75 µL/min, 72.92 µL/min in Novec7100 gemessen, für jeweils 2, 4, 8, 16 parallele NED-Aktoren entsprechend. Die erreichten Gegendrücke sind dabei 2.79 kPa, 6.88 kPa, 20.55 kPa, 29.61 kPa. Typische Antriebsfrequenzen der Mikropumpen liegen im Kilohertz-Bereich, was sich auf den hohen Miniaturisierungsgrad und die damit verbundenen mechanischen Steiffigkeit der NED-Aktoren zurückführen lässt. Die Gegenüberstellung der experimentell bestimmten Kennlinien mit Simulationsdaten aus dem entwickelten Modell der Pumpe, bestätigen das auf dem NED-Effekt beruhende Pumpprinzip.
The Lateral-Photovoltage-Scanning-Method (LPS) operates well for Si, Ge and Si_{1–x}–Ge_x for an analysis in defect regions below one part per million, where Secondary Ion Mass Spectroscopy (SIMS) or X-Ray Fluorescence (XRF) signals fall below its detection limit. Although LPS is well established since 1999, it is still poorly investigated. We used a computational simulation finite volume (FVM) approach, solving the van-Roosbroeck equations in three dimensions using a MUltifrontal Massively Parallel sparse direct Solver MUMPS. The signal transport is simulated by solving the Maxwell equations in two dimension for different sample geometries.
It could be shown that a typical LPS-measurement is distorted due to the samples geometry (except cuboid). This distortion can be simulated, understood and recalculated, as discussed for trapezoidal or cylindrical samples. Also using the signal generation simulation of this measurement technique it can be shown, that the measurement signal is convoluted depending on the inherent minority charge carrier life time reducing the local resolution. An investigation of the local resolution were made using a Gaussian function as the convolution function of this method. A comparison of simulations to real measurements was discussed on silicon samples with boron implantation pattern.
In 1955 Tauc already stated that the bulk photovoltaic effect, causative for the LPS measurement set-up, could be used detecting any quantity, which affects the band structure of a semiconductor.As strain is coupled to the conduction and valence band profiles by the deformation potential theory by van-de-Walle, we investigated the possibility to detect strain variations using LPS simulations. For an n-type Si sample with an on-top stressor stripe (silicon-nitride) the strain distribution in Si got calculated by finite elemente simulation (FEM) using solid mechanics module. By directly converting the strain profile to a single conduction and valence band, FVM LPS simulations were performed. It could be shown, that the LPS voltage can be connected to hole traps caused by the conduction and valence band profile. Therefore we can finally conclude, that the LPS measurement set-up is suitable measuring conduction and valence band variations caused by strain.
The subject of this thesis is the development and test of silicon strip detectors for the high luminosity upgrade of the tracking detector of the ATLAS experiment at the Large Hadron Collider. Special emphasis is devoted to the understanding of the impact of mechanical stress on the electrical properties and the particle detection performance of detector modules.
First simulations were done to estimate the maximum expected stress on a sensor when operated at -30 °C within the future silicon strip tracking detector ITk, at ATLAS. The maximum stress in a worst case scenario is expected to be 27 MPa. Tensile strength tests were done to estimate the maximum stress which can be applied to a silicon strip sensor. Silicon shards, with a thickness and dopand concentration corresponding to the ITk sensor specifications break at >23 MPa, wafers at >700 MPa and sensors at ~ 400 MPa. The huge variations lead to the assumption, that the tensile stregth, which is highly dependent on the quality of the crystal lattice, is due to different cutting technology. Wafers, irradiated with a fluence equivalent of a lifetime dose of an ITk sensor, show no stress dependency of the youngs modulus. The tensile strength of irradiated wafers is decreased by ~ 6,6 %. No damage on silicon sensors from mechanical stress is expected for sensor modules installed it the ITk.
The electrical properties of silicon strip sensors were studied for applied mechanical stress on ATLAS07 sensors up to 60 MPa. The specifications of the sensors are similar to the specification of strip sensors in the future silicon strip tracker barrel region of the ATLAS detector. The leakage current changes at 50 MPa by -1.7 %, the bias resistance by +0.8 % and the interstrip resistance by -25 %. The depletion voltage and the implant resistance are not affected by mechanical stress. Except for the interstrip resistance the results can be explained by piezoresistive effects.
Silicon strip modules were build and studied in particle test beams. These modules consists of an ATLAS07 or an ATLAS12 sensor and an analogue readout to study the influence of stress on the module performance. The sensor module noise is independent from the applied stress. An effect of stress on the signal strength was seen. The ATLAS07 sensor module signal strength was decreased and the ATLAS12 sensor module signal strength was increased with a slope of ~0,6 MPa^-1. The average cluster size of the ATLAS07 sensor module was increased by 0,25 % MPa^-1 and the average cluster size of the ATLAS12 sensor module was decreased by 0,06 % MPa^-1 with applied stress.
Over the past seven decades Si microelectronics have developed rapidly. The success of the growing microelectronic industry is also caused by the expansion of materials in addition to Si. Open challenges are the monolithic integration of group IV devices on Si photonics as well as overcoming the size mismatch between electronic parts in the nm range and photonic parts in the µm scale. In this thesis the future application of GeSn NIs on Si as a photodetector is evaluated. The key element required for high performance optoelectronic devices is the formation of high-quality GeSn nano-islands (NIs), i.e. overcoming growth challenges such as introduction of defects due to lattice and thermal mismatch between GeSn and Si substrate as well as suppression of Sn precipitation caused by the limited solid solubility of Sn in Ge. To achieve high-quality nanostructures, the selective growth of GeSn NIs on Si(001) seeds via molecular beam epitaxy is investigated, exploiting the advantages of nanoheteroepitaxy (NHE), i.e. growth on nano-patterned substrates. The best compromise between selective growth of GeSn on Si nano-pillars at significant higher growth temperature than the eutectic temperature of GeSn and the incorporation of Sn into the Ge lattice was achieved at 600°C. X-ray diffraction studies confirmed the substitutional incorporation of 1.4at.% Sn into the NIs avoiding considerable Si interdiffusion from the substrate. Transmission electron microscopy images have shown that dislocations and stacking faults caused by plastic relaxation of the GeSn NIs are located near the NIs/substrate interface and thus, dislocation-free GeSn NIs can be formed, due to gliding out of the threading arms triggered by the NHE approach.
The high crystal quality of the GeSn NIs, enables the investigation of the bandgap by μ-photoluminescence (PL) analyses, demonstrating the shrinkage of the direct bandgap with increasing Sn content in the quasi-direct semiconductor.
All NIs however feature a β-Sn droplet on their nano-facets. To suppress the out-diffusion of Sn and hence increase the Sn concentration of the GeSn alloy, the GeSn NIs were overgrown with a thin Ge cap layer. The Ge cap successfully hinders the formation of Sn segregates on top of the NIs. Capping at 600 °C and 650°C results in an enrichment of Sn at the surface, forming a GeSn crust with 8±0.5at.% Sn. This wetting layer both enhances the optoelectronic properties of the NI core and exhibits a relatively strong PL emission attributed to direct radiative recombination.
Finally, a first demonstration of a GeSn NIs based photodetector was successful, due to the utilization of Al nano-antennas exhibiting an enhanced light coupling into the GeSn NIs at a wavelength of 700nm. The responsible mechanisms is the local plasmonic field enhancement of the incoming light. The manipulation of the resonance wavelength into the telecommunication regime, i.e. >1550nm, have to be investigated in future studies.
Um die technischen Limitierungen bezüglich der Geschwindigkeit und Verlustleistung aktueller Mikroprozessoren zu überwinden, sind neue Technologien notwendig. Eine Möglichkeit ist der Austausch klassischer metallischer Verbindungsleiter durch optische Übertragungswege. Als monolithisch integrierbare Lichtquellen kommen Germanium-basierte Leuchtdioden oder Laser auf Silizium-Substrat in Frage. Dafür ist eine intensive direkte Lumineszenz der Ge-Schicht nötig. Aus diesem Grund wurde der Einfluss von Donatorkonzentration, Verspannung, Quanten-Confinement und Versetzungsdichte auf die Lumineszenzintensität von Ge-Strukturen mittels Photo- und Elektrolumineszenz-Spektroskopie (PL bzw. EL) untersucht. Eine Antimon-Konzentration von 3e19/cm³ erhöhte die Intensität um das Vierfache im Vergleich zu einer intrinsischen Ge-Schicht. Eine zusätzliche Zugverspannung durch ein virtuelles Substrat aus GeSn erzielte nur eine geringfügige Intensitätssteigerung. Eine GeSn/Ge Multi-Quantum-Well (MQW) Struktur steigerte die Lumineszenz um den Faktor 16. Jedoch wurde durch transmissionselektronenmikroskopische Untersuchungen der Strukturen eine hohe Dichte von Durchstoßversetzungen (threading dislocations) im Bereich von 1e9/cm² gefunden. Diese bewirken eine Reduzierung der Lumineszenzintensität um zwei Größenordnungen, wie PL-Messungen an pseudomorph gewachsenen Ge/Si-MQW-Schichten zeigen. Eine Analyse der Strom-Spannungs-Kurven ergab einen weitaus weniger schädlichen Einfluss von Durchstoßversetzungen im Germanium als dies für Silizium der Fall ist. Im Hinblick auf eine Steigerung der direkten Ge-Lumineszenz sollten dotierte MQW-Strukturen und eine Reduzierung der Versetzungsdichte Anwendung finden.
The aim of this work is to study the optical properties of crystal defects in multicrystalline solar silicon and poly-/microcrystalline silicon thin films on glass substrate.
First a setup for photoluminescence imaging on multicrystalline silicon solar wafers was developed. This system is suitable for detecting band-to-band luminescence as well as defect-related luminescence at room temperature on large-scale wafers at different stages of their processing.
Spectroscopic photoluminescence investigations of multicrystalline silicon solar wafers indicated a new intense luminescence line at ≈ 0.91 eV at room temperature. The origin of this line is probably found in a specific grain boundary. Furthermore, luminescence in the region of 0.8 eV was investigated in detail, and it was found that probably oxygen is responsible for a peak at 0.77 eV at 80 K.
Electroluminescence investigations at room temperature at both materials exhibit extended defect structures such as grain boundaries. Furthermore, it can be concluded that electroluminescence imaging in reverse bias mode indicate on serious breakdown points in solar cells, which can lead to destruction of solar cells and modules. By comparing defect-related and reverse bias electroluminescence images, a difference in the spatial distribution of defects emitting D1 radiation and defects emitting light under reverse bias beyond -12 V is detectable.
In addition, there seems to be a correlation in the distribution of non-doping impurities and photoluminescence. Concerning this, vertical slabs of two silicon blocks were examined by means of Fourier-transform infrared spectroscopy and photoluminescence. A correlation of the distributions of interstitial oxygen and the band-to-band luminescence profiles could be found. Additionally, a correlation between D3/D4 luminescence profile and nitrogen distribution in the blocks was observed.
Finally, the growth process, particularly the transition from amorphous to microcrystalline silicon by PECVD, was studied by combined photoluminescence and Raman investigations. Formation of silicon nano-grains was detected by means of photoluminescence and Raman spectroscopy.
The aim of this work is to describe and explain the properties of defects in multicrystalline (mc) and thin-film solar silicon (Si). For this reason, investigations with scanning electron microscope methods were performed, namely cathodoluminescence (CL), electron beam induced current (EBIC), electron backscatter diffraction (EBSD) and transmission electron microscopy (TEM). Additionally, photoluminescence (PL) and reverse-biased electro luminescence (ReBEL) measurements were also conducted. Through correlation of PL, ReBEL and EBIC, it was possible to localize breakdown sites at mc-Si solar cells. Problems that occurred during the thin-film EBIC investigations could be demonstrated and explained. For the first time cross sectional EBIC investigations could be performed on thin-film silicon tandem cells. At mc-Si, it was possible to observe the oxygen related P-line next to the common D1-line luminescence at 10 K clearly distinguishable from each other at once. Furthermore, a hitherto not comprehensively discussed intense luminescence line at 0.93 eV could be described in detail. Through correlation of PL, CL, EBIC, EBSD, and TEM measurements, the origin of the now named Di luminescence at 0.93 eV is postulated to be in connection with Frank partial dislocations, with two energetic levels inside the band gap, one at 112±9 meV below the conduction band and the other at 93±10 meV above the valence band. Finally, it was attempted to explain the behavior of twin boundaries at temperatures below 30 K, where these show an enhanced collection efficiency in comparison to the surrounding grains. An alteration of the local “freeze out” temperature, possibly by a local band gap narrowing, is suggested as a reason. Another conceivable explanation is a breakdown of the diode potential at the grains.
Die vorliegende Arbeit befasst sich mit der Struktur und der Entwicklung von Versetzungen in multikristallinen Siliciumblöcken aus der gerichteten Blockkristallisation. Versetzungen können für die Rekombination von Ladungsträgern verantwortlich sein und damit den Wirkungsgrad von Solarzellen und -modulen mindern. Die experimentelle Arbeit gliedert sich bezüglich ihrer Fragenstellung und der Methodenwahl in drei Teile und wird in den Kapiteln 2 bis 4 behandelt.
In Kapitel 2 wird die Defektstruktur vieler Siliciumblöcke mittels automatisierten Verfahren untersucht. Die Analysemethoden umfassen dabei die Auswertung der Wafertextur und der Infrarot-Durchlichtbilder sowie Ätzgrubendichtezählung (EPD) und Photolumineszenz (PL). Der Betrachtungsabstand ist makroskopisch und die Ergebnisse geben das Verhalten von vielen Versetzungen wieder.
Die Versetzungsstruktur in multikristallinem Silicium ist geprägt durch scharf abgegrenzte Bereiche mit sehr hoher Versetzungsdichte. Diese sog. Versetzungscluster bestehen aus einem Netzwerk aus Versetzungs-Pile-ups (Kleinwinkelkorngrenzen) und können in sogenannte leichte und dichte Cluster unterschieden werden. Die Versetzungscluster haben im Siliciumblock einen Ausgangspunkt, von dem aus sie sich ausbreiten. Dieser befindet sich hauptsächlich an Korngrenzen und generiert sich spontan während der Erstarrung in der Nähe der Phasengrenze. Das Auftreten von leichten und dichten Clustern ist abhängig von der Kornorientierung parallel zur Wachstumsrichtung. Körner mit Orientierungen nahe <111>, <211> und <311> neigen dazu leichte Cluster auszubilden, während Körner mit Orientierungen um <110>, <331> und <531> eher dichte Cluster ausbilden. Kornorientierungen um <100> und <511> sind unauffällig bzgl. der Ausbildung von Versetzungsclustern. Der Zusammenhang liegt in der Orientierung der Gleitebenen begründet.
Kapitel 3 behandelt die strukturelle Untersuchung der Versetzungen mit der Transmissionselektronenmikroskopie (TEM) sowie der Synchrotron- Röntgentopographie (WB-XRT). Die Versetzungen formen streng geordnete Pile-ups, welche letztendlich Kleinwinkelkorngrenzen entlang der Wachstumsrichtung ausbilden. Der Versetzungsabstand beträgt etwa 30 bis 800 nm, was mit einer Verkippung in der Kristallorientierung von 0,3 bis 0,07 ° korrespondiert. Die Rotation der Kristallorientierung verläuft hauptsächlich um eine Achse parallel zur Wachstumsrichtung. Anhand dieser Beobachtungen wurde ein Modell zur Abschätzung der Versetzungsdichte aufgestellt. Für die leichten Cluster beträgt diese ca. 2*10^5 cm^2 und für die dichten Cluster etwa 3*10^7 cm^2.
Die Auswirkungen der Versetzungsstruktur auf die elektrischen Eigenschaften werden in Kapitel 4 behandelt. Mit Electron Beam Induced Current (EBIC) und Dunkel-Lock-In-Thermographie (DLIT) wurde herausgestellt, dass vor allem solche Versetzungen elektrisch aktiv sind, welche sich zu Kleinwinkelkorngrenzen angeordnet haben. Ein niedriger Versetzungsabstand innerhalb der Subkorngrenzen scheint nur eine Bedingung für eine elektrische Aktivierung zu sein. Ein Zusammenhang zwischen dem Betrag der Verkippung einer Subkorngrenze und dem EBIC-Kontrast konnte nicht hergestellt werden.
In der abschließenden Diskussion (Kap. 5) wird ein Modell zur Entstehung und Evolution von Versetzungsclustern vorgeschlagen. Die Versetzungscluster generieren sich hauptsächlich an Korngrenzen. Dabei werden die Kleinwinkelkorngrenzen durch die Restrukturierung von Versetzungen sekundär gebildet. Die Ergebnisse legen nahe, dass dies während der Erstarrung unmittelbar nach der Phasengrenze geschieht. Die treibenden Kräfte sind demnach thermische Spannungen an der Erstarrungsfront. Die Versetzungen und damit die Subkorn-Strukturen erreichen die Phasengrenze und bleiben bei der weiteren Kristallisation erhalten, sodass die kontinuierlich auftretenden thermischen Spannungen mit neuen Versetzungen und damit mit erneuter Bildung von Versetzungen und Subkorngrenzen abgebaut werden müssen. Durch diesen Vererbungseffekt erhöht sich die Versetzungsdichte stetig mit steigender Blockhöhe.
Im Abschluss werden Maßnahmen zur Reduktion der Versetzungsdichte im Kristallisationsprozess diskutiert. Beim Ankeimen am Tiegelboden oder während der Erstarrung sollten geeignete Kornorientierungen bevorzugt werden. Eine weitere Maßnahme ist die Reduktion der radialen thermischen Spannungen während der Kristallisation. Unbekannt bleibt die genaue, atomare Ursache der erhöhten Versetzungsgeneration an Korngrenzen und die damit verbundene spontane Bildung von Versetzungsclustern. Weiterhin bleibt offen, ob und in welchem Umfang Lomer-Cottrell-Versetzungen innerhalb der Subkorngrenzen gebildet werden und welchen Einfluss diese auf die elektrische Aktivität haben.
Internal gettering based on oxygen precipitates is a technique which is used to remove occasional metal contaminations from the active region of microelectronic devices. In order to obtain efficient internal gettering, the precipitation of oxygen in silicon must be precisely controlled. This can be done by rapid thermal annealing (RTA). RTA offers the possibility to establish well defined vacancy concentrations in silicon wafers. Since vacancies are well known to enhance the precipitation of oxygen in Czochralski silicon, RTA pre-treatments can be used to control the generation of oxygen precipitates. This work provides information about the influence of vacancies, introduced by RTA, in silicon wafers on the nucleation of oxygen precipitates during a subsequent annealing in the temperature range between 400 °C and 1000 °C. Moreover, detailed investigations show morphologies and sizes of oxygen precipitates observed in vacancy supersaturated samples subjected to an annealing in the temperature range between 700 °C and 1000 °C for different annealing times. The morphology of the oxygen precipitates was investigated by scanning transmission electron microscopy (STEM) and Fourier transform infrared (FTIR) spectroscopy. In case of FTIR spectroscopy, the absorption bands were assigned to morphologies of the oxygen precipitates observed by STEM. The final part of investigations presented in this work is dedicated to the gettering efficiency of metal impurities. Special focus is devoted towards the gettering efficiency of Cu in vacancy supersaturated samples. In order to understand the gettering of Cu in samples contaminated with high and low concentrations of Cu, two getter tests were carried out. A haze getter test was used to investigate the getter efficiency of Cu in highly contaminated samples. In this particular case, the concentration of Cu equals the concentration of Cu at its solubility limit at 900 °C. A “7 day storage getter test”, developed in this work, was used for the investigation of the getter efficiency of Cu in samples contaminated with low concentration of Cu amounting to 1×1013 cm-2. It was found that the density of oxygen precipitates increases with increasing concentration of vacancies. The nucleation curves of oxygen precipitates in the vacancy supersaturated samples consist of three maxima wherein the maximum observed at 800 °C can be found only in the vacancy rich samples. These maxima can be explained assuming the nucleation of coherent plate-like nuclei consisting of oxygen mono-layers ((Oi)2-p1) and oxygen double-layers ((Oi)2-p2) for the peaks at 450 °C and at 650 °C, respectively, and VO2 mono layers for the peak at 800 °C. The STEM investigations have shown the change of the morphology of oxygen precipitates in samples subjected to nucleation annealing at various temperatures. It was observed, that different temperatures of the RTA pre-treatment and thus different supersaturations of vacancies did not influence the morphology of oxygen precipitates in samples annealed at 800 °C. After annealing at a temperature of 800 °C three and two dimensional dendritic precipitates were found. This kind of precipitates gave rise to an absorption band at 1040 cm-1 as shown by FTIR investigations. From the results of the getter test it was deduced that secondary defects like dislocations have a strong influence on the getter efficiency in samples contaminated with high concentrations of Cu and Ni. In case of the samples contaminated with low concentration of Cu, gettering at dislocations is less important and oxygen precipitates become the main getter sink for Cu. It was also observed, that Cu aggregates at the edge of plate-like precipitates at the site of tensile strain of the silicon lattice. Moreover, the size and density of oxygen precipitates can strongly influence the getter efficiency of metal impurities. The results and observations presented in this work can be very useful for designing and fabrication of high performance silicon wafers. The results can be used for the development of a gettering simulator based on oxygen precipitation. The results of the analysis of the FTIR spectra can be helpful for the fast characterization of the morphologies of oxygen precipitates by means of FTIR.