Materials research for group IV semiconductors: growth, characterization and technological developments

Resume : Upgraded metallurgical grade silicon (UMG-Si) directly purified via metallurgical routes has attracted much attention in recent years, due to its potential lower price. Compensated doping is the typical characteristic of UMG-Si, since the electrical dopants such as boron (B) and phosphorus (P) are not easily removed via metallurgical routes. Also, in order to keep the resistivity in the right range for the cell fabrication, intentional P doping for electrical compensation is necessary for p-type UMG-Si crystal growth. Here, we will report our recent results on the properties of compensated silicon wafers and compensated silicon solar cells. It is found that with the reduction of the net doping concentrations, the carrier lifetime in compensated silicon could get improved. Due to the greater ionized impurities scattering, dopant compensation could lead to the reduction on both majority and minority carrier mobility. Therefore, the compensated silicon solar cells show weaker spectral responses and lower short-circuit current. However, a higher open-circuit voltage could be obtained from the compensated silicon solar cells due to its larger net doping concentration. As a result, the compensated silicon solar cells could have the same efficiency as the conventional ones. Owing to the larger B concentration in compensated silicon, the compensated silicon solar cells suffer from the light induced degradation (LID) more seriously than the conventional ones. The boron-oxygen complexes in compensated silicon could be permanently deactivated by illumination at elevated temperatures. The compensated silicon solar cells show worse performance than the conventional ones at low light intensities, while they show better performance at high temperatures.

Resume : It is demonstrated that polycrystalline free standing Si wafers, as well as thin Si layers on various substrates can be processed using thermal spray of Si powder. Raman, resistivity, XRD, XPS, EDS and SEM analyses have been used for characterization of such Si based structures. It is shown that Si powder based structures sintered by thermal spray are fully crystalline with a low-fraction of the amorphous phase (Raman, XRD) after appropriate processing conditions and can be considered as a low-cost alternative for Si based wafers and layers for photovoltaic applications.

Resume : The PV industry is putting a lot of effort to reduce costs by reducing solar cell thickness as a way to achieve material savings while taking advantage of expected high conversion efficiencies for suck thicknesses. Besides the material gains, the classic process of ingot sawing cannot be a conceivable solution to produce thin silicon wafers, since approximately 50 % of the silicon material is lost in this step. Unlike other techniques, the Stress induced LIft-off Method (SLIM-cut) does not rely on the creation of weakened regions by ion implantation or porous layers to produce a thin silicon foil. It consists of three steps: (i) dispensing a stress inducing layer on the silicon surface, (ii) a thermal step to activate the stress and detach a thin silicon foil of silicon (iii) a chemical cleaning to obtain a flat thin foil of silicon. In this paper, we present large area (3 x 5 cm2) SLIM-cut foils obtained at room temperature using an epoxy stress inducing layer. Besides numerical simulations we demonstrate the capability to obtain several thin foils from the same substrate. Dislocation density in the foils and the thickness relationship between epoxy layer and foils are also presented. Measured lifetimes in these silicon foils increase after etching suggesting that recombination centers are present close to the foil surface. Effective lifetimes of 50 microseconds were obtained in 120 micron thick foils, giving estimated conversion efficiencies of 19.2 % using PC1D.

Resume : To overcome the interfacial mixing in growth of Ge on Si substrate, filling vacancies and interstitial sites near the interface region is a very promising way. Carbon (C) atoms can diffuse easily through interstitial-site exchange mechanism due to C's small atomic radius. In this study, an effect of sub-monolayer (ML) C mediation on interfacial mixing in Ge growth on C-covered Si and Ge/Si substrates was investigated. Ge/C/Si and Ge/C/Ge/Si structures were prepared by sequential deposition of C and Ge at low temperature of 200-300°C and subsequently annealed at 650°C in the chamber. In Ge/C/Si, peak intensity of Si-Ge vibration mode in Raman scattering decreased with increasing C coverage (CML) for CML ≤ 0.2 ML and diminished for further CML. Ge(220) diffraction peak intensity decreased with increasing CML. This indicates that interdiffusion of Ge and Si atoms were suppressed for CML ≥ 0.3 ML while crystallinity of Ge layer deteriorated due to C incorporation. On the contrary, in Ge/C/wetting Ge/Si, crystallinity of Ge layer for the wetting Ge layer of over 5-nm thick was better than that in Ge/C/Si and kept the suppression of interfacial mixing. Furthermore, crystallinity of Ge layer improved with thickening the wetting Ge layer even for a large amount of CML of 1 ML. This indicates that indirectly-deposited C on Si, that is C on the wetting Ge layer, effectively acts to improve crystallinity of Ge layer without interfacial mixing due to well-organized C mediation effect.

Resume : Crystalline group-IV alloys exhibit unique physical properties, which are superior to their elemental counterparts. By implication, these alloys hold large promises for future opto- and micro-electronic devices. However, their metastable nature renders them highly susceptible to compositional/structural changes when exposed to elevated temperature, leading to deterioration of the film properties. We will discuss the effect of post-growth vacuum annealing on the stability of 40 nm-thick strained Ge0.94Sn0.06/Ge(001) films. We observed pronounced differences in the morphological and defect evolution with temperature compared to previous reports [1]. In particular, annealing at 500°C generates faceted pits on the surface, similar to those observed on relaxed SiGe alloys after annealing. Our Raman spectroscopy and reciprocal space mapping measurements clearly indicate that the crystal structure and strain in the layer remain unaltered, and no dislocations could be discerned in electron microscopy images around the pits. With increasing temperature (T>500°C), the length (a few hundred nm) and depth, but not the density of the pits increases. In addition to these surface defects, SIMS and GIXRF data reveal Sn diffusion towards the surface and substrate interface. As a consequence, the “effective” Sn content and the strain in the film are reduced, which will have a direct impact on the optical and electrical characteristics of the film. [1] Deng et al., Phys. Rev. Lett. 80, 1998

Resume : III-V multijunction solar cells on Ge and Si substrates are of the great interest due to their high efficiency (>40% for GaInP/GaAs/Ge structures). However, the electrical properties of III-V/IV interface are not studied enough. The p-n junction in p-doped IV group substrate is usually formed by diffusion of n-dopant (P) from III-V wide-band-gap window-layer (GaInP and GaP for Ge and Si, respectively) during epitaxial growth. Transport of charge carriers over the III-V/IV interfaces was investigated using a combination of I-V and capacitance measurement techniques. It was demonstrated that there is an undesirable potential barrier for majority carriers at the n-GaInP/n-Ge interface. This barrier results in fill factor decrease at low temperatures. The origin of its formation is attributed to diffusion process. The diffusion coefficient of P in Ge is higher than that for Ga, but the solid solubility limit is vice versa. This may result in a barrier formation. In case of GaP/Si structure the diffusion coefficient and solid solubility limit for P in Si are both higher than that for Ga. Study of charge carriers transport over the n-GaP/n-Si interface has demonstrated no potential barrier at this interface.

Resume : We developed a novel, simple method effectively to purify metallurgical grade silicon powder (MG-Si). The method consisted in first time, to prepared porous silicon powder by acid etching Ag-Si alloy powder and in second time, to annealing thermal of porous silicon powder. The morphology and structure of the as-obtained material were investigated using scanning electron microscopy (SEM), X-ray diffraction (XRD), Transmission electron microscope (TEM), Photoluminescence (PL), FTIR, and BET methods. It was found that the porous silicon powder (size about 15µm) had a spongy structure. Inductively coupled plasma-atomic emission spectrometry analysis (ICP-AES) indicates that the concentrations of impurities in silicon.

Resume : The effect of ultrasonic treatment (UST) on the defect structure of the Si–SiO2 system by means of electron spin resonance (ESR), metallography, MOS capacitance technique and secondary ions mass-spectroscopy (SIMS) is presented. The non-monotonous dependence of the defect densities on the US wave intensity has been observed. The influence of the UST frequency on the ESR signal intensity of the defect centres depended on the defects type and structure and may be caused by vibrational energy dissipation (phonon trapping on the defects site) which are a function of defect centers type. The influence of the UST on the Si–SiO2 interface properties depends on the oxide thickness and crystallographic orientation.The density of point defects and absorbed impurities at the Si–SiO2 interface can be reduced and its electrical parameters improved by an appropriate choice f the UST and oxidation conditions.

Resume : Iron-based Heusler alloy Fe3Si is promising for a spin injector into Si since high quality epitaxial growth of DO3-Fe3Si on Si(111) or Ge(111) has been succeeded by Molecular Beam Epitaxy (MBE) that is carried out near room temperature. The Fe3Si/Si(111) epitaxial interface has a large lattice mismatch by ~4%. This elastic condition may give a considerable effect on epitaxial quality. Under such a concerned situation, we have applied axial ion channeling (AXIC) to investigation of atomic disordering at the Fe3Si/Si(111) interface. The AXICs for both as-grown samples and samples annealed at 100 or 200oC were measured to obtain axial channeling parameters (the minimum yield, km and the critical angle, psi). The km increased a little from 0.18 to 0.19 and to 0.20 and psi decreased much from 0.99 to 0.85 and to 0.83 degree after the annealing at 100 and 200oC for 2h. The small increase of km and the large decrease of psi clearly indicate disordering at the atomic raw along <111> directions. The film composition was completely maintained to be initial one after annealing. This fact suggests that the observed disordering can not be caused by atomic interdiffusion as observed in Fe4Si/Ge, Fe4Si /Si and Fe2MnSi/Ge systems. Using the Debye theory and the Barette-Gemmell model, we will calculate one-dimensional atomic displacements due to thermal vibrations and extrinsic reasons introduced by thermal annealing in order to make quantitative discussion on the disordering at the raw

Resume : Visible range to telecom band spectral translation can be accomplished using a SiC wavelength selector under appropriate near ultraviolet optical light. The selector is realized by using double pi’n/pin a-SiC:H photodetector sandwiched between two ITO contacts. Visible communication channels (400nm-650nm) are transmitted together, each one with a specific bit sequence. The combined optical signal is analyzed by reading out the generated photocurrent, under near-UV background and different intensities. Results show that background intensity works as selectors in the infrared region, shifting the sensor sensitivity. The background intensity balances the electric internal fields across the device leading to diverse charge accumulation at the different layer interfaces, which filter one or more input channels depending on the light penetration depth of each optical signal wavelength. Low intensities select the NIR range while high intensities select the visible part accordingly to its wavelength. Here, the optical gain is very high in the red range, decreases in the green range, and stays near the unity in the blue region and strongly decreases in the near-UV range. The transfer characteristics effects due to changes in steady state light intensity and wavelength backgrounds are presented. The relationship between the optical inputs and the output signal is established. A capacitive optoelectronic model will be presented and tested using the experimental results.

Resume : An extensive multifrequency electron spin resonance (ESR) study has been carried out as a function of oxidation temperature Tox in the range 400-1066 °C on the thermal, higher index, (211)Si/SiO2 interface, with a view to atomically characterize the occurring types of inherent paramagnetic point defects and quantify densities. The (211)Si face is of interest for optoelectronic application and natural boundary faceting of some Si nanowires. The main types of interface defects observed are Pb-type centers, that is, interfacial Si dangling bond (DB) defects (generic entity Si3ºSi•). Two species are observed: A dominant one with the Si DB along the [111] direction at 19.5° with the interface normal. Based on the inferred principal g matrix values and the observed 29Si hyperfine structure, it is typified as Pb0(211). A second, much less intense variant, corresponds to Si DBs along the equivalent [1-11] and [11-1] directions. For low oxidation temperatures, a total Pb0(211) density of ~1  1013 cm-2 is observed which gradually decreases to level out at a 60% reduced value for Tox → 800 °C, exposing interfacial relaxation. In terms of the prevailing Pb-type defect density, the (211)Si/SiO2 interface quality is found comparable to that of (110)Si/SiO2, yet about two times worse than the technologically dominant (100)Si/SiO2 one. The results may be understood within the context of the presence of two non-equivalent Si sites at the higher index (211)Si surface.

Resume : High resistivity Si wafers of 8000 Ωcm were irradiated with Bi^{6+} ions of 28 MeV kinetic energy, at fluence of 5x10^{11} ions/cm^2. The irradiation non-uniformity was less than 5%. On their passage through the crystal, they lose their whole energy by interaction mainly with the lattice, generating vacancy-interstitial pairs, which in their turn produce complex point defects, and eventually the ions are stopped. As they are bigger and heavier than Si atoms, Bi ions produce a strain field, which influences the trapping phenomena. This work studies the defects acting as traps by using the experimental method of thermally stimulated currents without applied bias (TSCAB), and modeling the discharge currents. The traps were charged by sample illumination with light of 1000, 800 and 400 nm wavelengths at low temperature, and TSCAB were recorded during quasistatic heating. The detrapped carriers move under internal electric field having two components, one being produced by the still trapped carriers, the other one being a permanent electric field which models the strain due to the stopped Bi ions. The recorded curves are resolved by modeling into seven traps, five of them being assigned to known point defects in silicon. All cross sections are temperature dependent, and some levels are broadened, the broadening being considered Gaussian.

Resume : The low temperature studies of SOI-structures have been carried out in a temperature range of 4.2 – 300 K at magnetic fields up to 14 T. The samples with initial boron concentration of about 2.4·10¹⁸ сm^-3 have been investigated. The treatment of experimental data allowed for determination of the parameters for low dimensional poly-Si layers with p_300K = 2.4·10¹⁸ сm^-3. In particular, a calculated radius of localization, density of states and average distance of hopping on localized states were α₁ = 6 Å, g₁ = 1.3·10²¹ eV·cm^-3, 31 < R < 36 for once occupied centers in the hopping conduction region. The results of the studies of SOI-structure conductance at low temperatures in the range of hopping conductance and possibilities for sensor applications of this material have been analyzed.

Resume : Electrical properties of Si(1-x)Ge(x) alloys, first of all the mobility of electrons and holes, attract close attention of physicists trying to get a deeper insight into a complicated problem of alloy scattering of charge carriers, primarily at x> 0.1. The aim of this work is to furnish information on charge carrier mobilities in moderately doped n- and p-Si(1-x)Ge(x) at x<0.1. Clearly defined changes in electron and hole mobilities over a temperature interval of 20 to 40 K appear in the n- and p-Si(1-x)Ge(x), starting from x= 0.008. Electrical measurements on similar Si materials without Ge doping revealed a pronounced contribution of the alloy scattering, even in diluted Si(1-x)Ge(x). Comparison of experimental data to calculated ones shows that the existing theoretical models strongly overestimate the alloy scattering in diluted n- and p-Si(1-x)Ge(x) at x<0.1.

Resume : Self-interstitial clusters (I_n) in silicon impact the fabrication of electronic devices and have been extensively studied recently. The formation of the silicon di-interstitial (I_2) is the first step in the sequence of the I_n complexes. Theoretical calculations have predicted that I_2 is highly mobile, so it plays an important role in self-interstitial clustering and can interact with other lattice defects. The available information on the interactions of I_2 with impurities in Si is, however, very limited. In this work we argue that a DLTS signal associated with hole emission from an energy level at E_v + 0.095 eV is related to a complex of I_2 with an interstitial oxygen atom (I_2-O_i). This signal has been observed in the DLTS spectra of p-type Si:O samples irradiated with 6 MeV electrons or alpha particles. Isochronal and isothermal annealing studies of the samples have shown that the defect responsible for the DLTS signal from the E_v + 0.095 eV level disappears after anneals in the temperature range 50-100 oC and its annealing behavior is similar to that of a center giving rise to the infrared absorption band at 936 cm^-1 previously assigned to a local vibrational mode (LVM) due to the I_2-O_i complex. Possible configurations of the I_2-O_i complex found by ab-initio modeling have been analyzed. Formation energies, energy levels and LVMs for different configurations have been determined and compared with the experimentally obtained values.

Resume : The AlGaN/GaN high electron mobility transistors (HEMTs) is a promising candidate for microwave applications due to its high power and high temperature. owing to their large and direct band gap, as well as favorable transport properties. Moreover, III-V nitrides could be suitable for the emitters and detectors of green and shorter wavelengths, in turn making investments in this class of materials more than worthwhile. The electrical characteristics of metal contactst on-GaN have been examined by numerous research groups. The mechanism of leakage currents through GaN and AlGaN Schottky interfaces is discussed using temperature- dependent current–voltage (I–V–T) measurements, and found that the barrier thinning caused by unintentional surface-defect donors enhances the tunneling transport processes, leading to large leakage currents through GaN and AlGaN Schottky interfaces. Key world: AlGaN/GaN, HEMTs, I–V, Schottky barrier height, Barrier inhomogeneity.

Resume : Several material cost reduction alternative strategies are being searched to increase PV competitiveness in the energy market. These range from i) kerf-free wafering, ii) ribbon technologies, iii) upgraded metallurgical grade Si and iv) thin-films grown over a substrate. This work follows this last line of research, presenting an inline optical continuous CVD process operating at low temperature (<873K) and at atmospheric pressure to grow microcrystalline silicon films on top of crystalline, sintered silicon and compressed silicon powder substrates moving at constant speed (>10mm/min) inside the furnace. In the case of solid substrates, laser patterning and its orientation inside the deposition chamber has a strong influence in the growth rate film deposition inside the deposition chamber, reaching values of 6 µm/min. When a powder silicon substrate is used, microcrystalline silicon films grown on top of this powder substrate can be grown in a continuous mode at 10 mm/min substrate speed with maximum growth rates up to 90 µm/min. Such growth rates result from a unique combination of substrate nature, flow dynamics, gas precursor and carrier gases choices. The grown samples are impurity free since no foreign materials are in contact with deposited silicon. This silicon layer may act as a barrier to substrate bulk impurities, preventing contamination during the solar cell processing of the top layers.

Resume : For the correct measurement of the oxy-hemoglobin in the cerebral blood flow, it is important to adherence with skin of head and the sensors have to fabricate on flexible substrate. In this presentation we would like to report the possibility of Ge NIRS sensors forming on flexible substrate. The 15nm thick Ti and 50nm Cu layers were formed on SiO2 substrate formed by electron beam deposition. The copper substrate was located parallel in a still bath containing 7 vol.% GeCl4 in propylene glycol. The bath was maintained at a temperature of 333K. The deposition rate was about 100nm/h. Results of the XRD measurement indicated that only (211) weak peak was observed but no peak was observed by Raman spectroscopy. The H bonding was measured by a FT-IR reflection spectra of the Ge electrodeposited of 3.4m-thick on the Cu/Ti substrate. Ge-H and Ge-H2 peaks were observed. From XPS depth profile measurement, 10%-oxygen atom was observed in the film, but Cu and Ti were not observed. The oxygen contamination was able to be suppressed by introducing argon gas during film deposition. The concentration of oxygen in the film was almost 0%. Optical band gap was about 0.73eV. It has succeeded also in depositing a Ge film by the electro-deposition method on 100m-Polyimide thin film with Cu/Ti structure. We have concluded that electrodeposited-germanium film have shown the possibility of near infrared sensors on flexible substrate. This work was supported by JSPS KAKENHI Grant (C)24500661.

Resume : This work is focused on physical properties of silicon oxycarbide (SixCyO1-x-y) thin films. The SixCyO1-x-y films are deposited onto <100> p-type silicon substrate by radio-frequency reactive magnetron sputtering using SiC target. During the growth, RF gun power is controlled to 100 W, argon flux is fixed to 10 sccm, and oxygen flux is fixed to 0.5 sccm. The chemical content of the as-grown films is extracted by X-ray photoelectron spectroscopy (XPS). The Si content is ~31 at.%, the C content is ~43.0 at.% and the O content is ~26 at.%. Si 2p core level spectrum is centered at ~101.4 eV. Since the binding energy of Si for pure Si is about 99 eV and that in pure SiO2 is ~103.4 eV, our result suggest that most Si atoms are bounded to at least one O atom. C 1s core level spectrum is peaked at ~284 eV. The binding energy of C is ~283.1eV in SiC and is about 284.6 eV for pure C. The result indicates that most C atoms are bonded to either C or Si atom. We have also performed X-ray diffraction (XRD), Raman, and TEM measurements. SiC crystal seems exist within the as-grown Si0.31C0.43O0.26 film. More details will be demonstrated. Photoluminescence (PL) measurements are performed at room temperature under excitation at 254 nm. The PL spectrum is centered at 395 nm and the FWHM is around 168 nm. The Si0.31C0.43O0.26 film emits light in the UV to visible range. This result infers that the as-grown SixCyO1-x-y films are capable for emission applications.

Resume : We investigated the deposition of polycrystalline Ge (poly-Ge) films for the application to the channel layers in 3-dimensional NAND flash memories. In-situ B-doped poly-Ge layer was deposited by ultra high vacuum chemical vapor deposition process. The effects of in-situ B doping on the poly-Ge layer were examined by varying the flow rate of the B2H6 gas up to 200 sccm with the fixed flow rate of GeH4 at 600 °C. We found that the growth rate of poly-Ge increased with the B2H6 flow rate. Poly-Ge layers were formed with an island shape at B2H6 flow rates of up to 50 sccm, which is changed to continuous layer with the increased deposition rate above the flow rate of 100 sccm. The resistivity of the poly-Ge layers was calculated to be about 0.001 Ω•cm. From secondary ion mass spectroscopy data, the B concentration was found to be about 1.5E20 atoms/cm3 for B2H6 flow rates above 100 sccm. Because Ge/SiO2 interfacial energy decreased with B concentration, the growth rate of poly-Ge layers increased with the B2H6 gas flow rates.

Resume : Germanium has become an important semiconductor in advanced electronic devices with implantation being a central technology for doping devices. Under certain implantation conditions, Ge becomes porous via the generation and clustering of vacancies in the amorphous Ge (a-Ge) phase. This process may result in deleterious device behaviour. Such nano-structured materials also tend to possess properties that their macroscopic counterparts do not and may be utilised in novel device applications. This work seeks to investigate both the effect that a porous-germanium surface layer may have on the annealing kinetics within the underlying a-Ge substrate and the structural properties of the a-Ge within the porous-germanium layer as revealed by Raman scattering. This structure is found to be very structurally relaxed with further relaxation occurring on annealing suggesting that it contains very few lattice defects. At temperatures around 500C the a-Ge recrystallises leaving a porous structure composed of crystal Ge. Given the high surface area the porous structure may be used as a sensing material. We also investigate surface enhanced Raman scattering for such applications.

Resume : Different types of nanocomposite thin films based on silicon (Si) and silicon carbide (SiC) nanocrystals were investigated by means of the optical spectroscopy of photoluminescence (PL), Raman scattering (RS) and second harmonic generation (SHG). Samples of the first type were prepared by thermal annealing of hydrogenated amorphous silicon (a-Si:H) and silicon carbide (a-SiC:H) films of 0.5-1 µm thickness deposited on substrates of crystalline Si (c-Si) and quartz. Samples of the second type were prepared by direct deposition of carbon and silicon ions in vacuum on c-Si substrates by using arc ion source. The RS measurements revealed that Si nanocrystals (nc-Si) with sizes from 3 to 10 nm could be formed. The SHG measurements allowed us to monitor an appearance of nanocrystals of SiC (nc-SiC) in the prepared nanocomposite films. The thermally annealed nc-Si films are characterized by PL emission in the region of 600-900 nm similar to the PL of electrochemically prepared porous silicon. The nanocomposite films exhibit PL bands at 400 – 500 nm and at 600 – 1000 nm, which are attributed to the radiative recombination related to electronic states in nc-SiC and nc-Si, respectively. The annealed films were subjected to stain etching in hydrofluoric acid solutions in order to enhance the PL intensity due to the passivation of the nonradiative defects. The obtained results indicate new possibilities to prepare SiC-based luminescent films, which can be used in light-emitting devices.

Resume : Scanning x-ray diffraction microscopy (SXDM) has become a routine technique giving access to structural properties of materials at the nanoscale. In this context, an optimized technique and a new software package have been implemented at the ID01 beamline (ESRF, Grenoble). The technique consists in a two-dimensional (2D) quicK continuous Mapping (K-Map) with sub micrometer resolution of a sample at a given reciprocal space position. These real space maps are recorded in continuous motion of the sample for every point across a rocking curve while recording 2D detector images using a sufficiently fast pixel detector. Five dimensional data sets are then produced consisting of millions of detector images. The images are processed by a user friendly X-ray Strain and Orientation Calculation Software (XSOCS) which has been developed at ID01 for automatic and fast analysis. It separates tilt and strain, and generates 2D maps of these parameters. At spatial resolutions of typically 200 to 800 nm, this quick imaging technique obtains strain sensitivity below ∆a/a=10-5 and a resolution of tilt variations down to 10-3 ° over a field of view of 200 x200 µm2. The K-Map technique, the XSOCS software and several examples will be presented: (i) The ESRF logo, printed in a SiGe thin layer, for testing the K-Map technique as well as XSOCS. (ii) The investigation of strain distribution of Ge microstripes, a candidate as active material for integrated light emitter applications.

Resume : We demonstrated that low-concentration interstitial carbon impurities affect properties such as V(on-switch) and the carrier lifetime of silicon power devices such as IGBTs using photoluminescence (PL) spectroscopy after luminescence activation by carbon ion implantation and electron irradiation. Control of light element impurities is important for improving device performance. We assumed that interstitial carbon is important factor for device performance, because they are one of origins of deep level.　In order to verify the effects of carbon impurities separately, we quantitatively investigated carbon impurities in oxygen-free silicon epitaxial wafers. 　　We implanted carbon ions into the epitaxial layers at a controlled concentration and annealed them at 1173 K to recover the crystalline quality. Then, we irradiated the layers with electrons to activate the carbon-related luminescence centers that are complexes of interstitial carbon and oxygen [Ci-Oi at 0.79 eV (C-line)] and interstitial and substitutional carbons [Ci-Cs at 0.97 eV (G-line)].　Carbon impurities are difficult to detect, but we successfully evaluated them using PL spectroscopy after these treatments. 　　A calibration curve relating to the G-line (Ci-Cs) intensity to the carbon concentration was obtained on the basis of the quantitative relationship between the carbon ion-implantation levels and the G-line intensities. Carbon impurity concentrations, even on the order of 1014, resulted in an increased V(on-switch) and decreased carrier lifetime. To decrease the carbon concentration is essential for increasing the performance of high-voltage silicon power devices.

Resume : The Ribbon technologies make excellent use of silicon, as wafers are crystallized directly from the melt at the desired thickness and no kerf losses occur. Therefore, they offer a high potential for significantly reducing photovoltaic electricity costs as compared to technology based on wafers cut from ingots. However, the defect structure present in the ribbon silicon wafers can limit material quality and cell efficiency. In this work, we report on the electronic properties on ultrathin p-type silicon ribbon (< 100 μm) grown by the RST method. We investigated the electronic properties of majority (conductivity and mobility) and minority (lifetime) charge carriers of p-type Si RST materials before and after plasma hydrogenation. The hydrogenation was carried out on bare Si ribbon as well trough amorphous silicon (a-Si:H) and amorphous silicon-nitride (a-SiNx:H) layers coated ribbon silicon. The layers can serve as passivating layers as well as antireflecting coating layers. The same MW-RF ECR assisted dual-plasma mode reactor was used for the hydrogenation and passivation/ARC steps. The penetration depth of hydrogen into the silicon was deduced from boron doping profile thanks to the capacitance-voltage measurements carried out on Fz silicon as a reference. We found a strong correlation between the amount of hydrogen atoms introduced and tits location in silicon with the minority carrier lifetime. We show that the optimized plasma hydrogenation step can boost the minority carrier lifetime by a factor of 3, which indicate a strong decrease in bulk defects. This is suitable for the fabrication of high efficient solar cells on ribbon silicon.

Resume : A low temperature multifrequency electron spin resonance (ESR) study has been performed on Cz-(100)Si/insulator structures with organosilicate films of low dielectric constant  grown at 300 °C using the plasma-enhanced chemical vapor deposition method (PECVD). After subjection to a short-time UV-irradiation assisted thermal curing treatment at 430 °C to remove the organic component from the low- film and obtain optimal porosity, the NL8 ESR spectrum of C2v symmetry is observed, characterized by g1 (//[100]= 1.99983, g2(//[011]=1.99274, g3=(//[0-11])= 2.00115. Along previous identification, this reveals the generation in the c-Si substrate of singly ionized thermal double donor (TDD) defects with a core contained of oxygen atoms. The generation is found to be non-uniform, the defect density piling up towards the interface. This enhanced TDD formation toward the dielectric/Si interface is hypothesized as being the result of the stress induced in the c-Si substrate underneath the interface by the presence of interfacial stress. This suggests substantial stress being involved with PECVD deposited and thermally treated organosilicate low-k films, although the quantification of the occurring stress is awaiting independent assessment. Building on the forwarded interpretation, the results obtained in this study represent a specific different illustration of the impact of stress on the formation modalities of oxygen-related TDDs in c-Si, thus providing independent support of previous reports on the influence of stress on TDD formation.

Resume : Hydrogenated amorphous silicon (a-Si:H) thin films are prepared using DC magnetron sputtering method at 200°C substrate temperature using a plasma of argon and hydrogen gas mixture. A serie of films are deposited under optimized preparation conditions of hydrogen and argon flow rates but at different total gas pressure. Their structural, optical and electrical properties are investigated. The structural properties of the films are analyzed from the infrared absorption. The total hydrogen bond concentration decreases slightly when the total pressure increases. The main effect observed is the important decrease of the polyhydride bonding group concentration (SiHn) when the total gas pressure increases. With increasing total gas pressure dark conductivity increases from 1E-9 to 5E-8 Scm-1 and photoconductivity under white light increases from 5E-6 to 1E-4 Scm-1. The density of states (DOS) is determined from constant photocurrent method (CPM). A correlation is established between DOS and hydrogen bonding configurations deduced from FTIR analysis. All the investigated films properties show a coherent variation with the total gas pressure. Keywords: a-Si:H, FTIR, DOS, CPM

Resume : Two methods of pulsed DC magnetron sputtering deposition have been used to form high rate, hydrogen-free crystalline silicon. The first method is in-situ crystalline silicon deposition. The second method is high rate amorphous silicon deposition followed by an anneal to induce crystallization. Over 20 ?m thick crystalline silicon can be formed on wafers up to 200 mm round or 156 mm square. Two vacuum deposition systems were used for substrate cleaning and deposition. The crystallinity of silicon was analysed by ellipsometry and Raman spectroscopy. All silicon samples deposited below 600 ?C are amorphous. Silicon deposited with table temperatures between 600 ?C and 650 ?C are poly-crystalline. Fully crystalline silicon is deposited in-situ at table temperatures greater and equal to 650 ?C. In-situ crystalline silicon has been deposited at 40 nm/min and amorphous silicon can be deposited at over 400nm/min subject to power density limitations for the silicon target. Up to 20 ?m thick amorphous silicon deposited at room temperature is fully crystallized by annealing in vacuum on a 1000 ?C table for 2 hours. This work demonstrates that sputtering offers significant potential for depositing the absorber layer in silicon based photovoltaics.

Resume : The concentration of defects in silicon can be changed by heat treatments or ion implantation. Some of these defects form complexes which can act as donors or acceptors and their concentration can even overcome the initial doping concentration of the material and hence cause an inversion of the doping type. One way to introduce defects and also to catalyze the formation of defect complexes is by proton implantation. For our project, a high resistivity p-type magnetic Czochralski grown silicon wafer was implanted with 4 MeV protons at a medium implantation dose. In a subsequent annealing step, different defect complexes were formed in the proton implanted region [1] and in the substrate region (where only the heat treatment caused a change in the defect concentrations) [2]. The local doping type and the relative change of the minority carrier diffusion length were measured using Electron Beam Induced Current (EBIC) [3]. Furthermore resistivity profiles were measured using Spreading Resistance Profiling (SRP). The results show several changes of the doping type, the minority carrier diffusion length and the resistivity profile for thermal budgets. In addition to some, already known phenomena, the formation of a high temperature donor complex has been observed. [1] Y. Ohmura et al. Solid State Communications, Vol 11. pp. 263-266 (1972) [2] G. S. Oehrlein, Journal of Applied Physics, 54, 5453 (1983) [3] S. Kirnstötter et al., ECS Transactions, Vol. 50, pp. 115-120 (2013) The work has been performed in the project EPPL, co-funded by grants from Austria, Germany, The Netherlands, Italy, France, Portugal- ENIAC member States and the ENIAC Joint Undertaking.

Resume : The multi-crystalline (mc) silicon wafers are the dominating substrate materials for solar cells in current photovoltaic (PV) industry. However, The average efficiency of mc silicon solar cells is far lower than that of CZ silicon cells since the electrical performance of mc silicon solar cells is restrained by the ubiquitous structural defects in them such as grain boundaries and in particular the high density of dislocations in some grains and the metallic impurity precipitate at these structural defect sites. Another disadvantage of mc silicon wafers has been recognized as the high surface reflectance after texturing. The random orientation of grains makes the alkali etching futile, and the isochemical etching for mc silicon wafers is far behind the alkali etching in reflectance, which leads to the optical losses of solar cells. The cast single crystalline silicon technique is designed to produce the single crystalline material inheriting the advantages of both CZ and mc silicon. This material can be considered as an ideal material for silicon solar cells: square, single crystalline, of low structural defect density and with low fabrication cost. Here, we have demonstrated the seed-assisted cast quasi-single crystalline (QSC) silicon technique to achieve high efficiency solar cells with low cost. Compared to multicrystalline (mc) silicon, the QSC silicon has better material properties, having higher minority carrier lifetime and fewer grain boundaries and dislocations. Furthermore, the <100> oriented QSC silicon can achieve a lower surface reflectance using alkaline texturing. Based on these two factors, the efficiency of the QSC silicon solar cells with the industrial size has been improved by up to 1% absolutely from the mc-Si counter parts. Compared to the Czochralski (CZ) silicon solar cells, the QSC cells have slightly lower efficiency but high productivity and negligible light-induced degradation. These results suggest a great potential of the QSC silicon applied in photovoltaic industry as the next generation substrate. The main problem of QSC silicon is the dislocations. Generally, the density of the scattered dislocations increases from the ingot bottom to the top. The very low density of dislocations is found to have little influence on the lifetime degradation, and as the dislocation density increasing, its influence on the lifetime degradation and therefore on the efficiency degradation becomes much larger. Another feature of dislocations in QSC silicon is the dislocation clusters, which are the most deteriorate to the quality of QSC silicon material and solar cells by degrading the fracture strength, the lifetime and all the solar cell characteristic parameters. The further efforts should be emphasized on the control of the generation and the multiplication of dislocations in QSC silicon. We have also experimentally revealed that the iron profile is characteristic of two peaks, consistent with the numerical simulation results. The first one is owing to the iron diffusion from the quartz crucible. The second one occurs at a height above the initial solid-liquid interface, which results from the formation of an iron-rich layer at the initial stage of the whole crystallization process. Nevertheless, the QSC wafers using seed-assisted directional solidification (SDS) process, providing a promising future of the application of the cast QSC technique in PV industry. Thereafter, by simply employing the inexpensive conventional cast techni- que, the industry can produce more cost-effective QSC silicon as the alternative substrate for the solar cell.

Resume : Ultra high efficiency silicon solar cells, particularly those with rear contacts, require silicon substrates with minority carrier lifetimes of several milliseconds or higher. In these very high lifetime materials, effects pertaining to defects which usually do not limit lifetime become apparent. Studies of such defects are challenging, typically due to low defect concentrations, the relatively weak recombination activity of the defects, and experimental issues associated with separating bulk and surface recombination effects. We will present new results from an international collaboration which aims to ascertain the effects of nitrogen-related impurities on carrier lifetime in silicon. Czochralski silicon wafers were grown to contain different levels of nitrogen (from 5E13/cm^3 to 2E15/cm^3). The samples were studied by analysis of injection-dependent transient lifetime measurements, deep-level transient spectroscopy and infrared spectroscopy. Samples for lifetime measurements were studied with a range of surface passivation treatments (silicon nitride, liquid HF and a-Si) in order to isolate the effect of the bulk recombination from surface effects. A correlation between nitrogen concentration and recombination activity was found. We will discuss the degree to which the nitrogen-related defect(s) can be passivated by bulk hydrogen and possible structures of the defect(s).

Resume : Silicon materials for photovoltaic applications contain high concentrations of residual donor and acceptor impurities, which may cause reduction of ionization energies and formation of impurity bands. In spite of their significant potential impact on device performance, detailed information has not yet been obtained. We have reexamined the donor and acceptor levels in the higher concentration range on the basis of the analysis of donor-acceptor (DA) pair luminescence. We have identified the discrete line spectrum due to the DA pair luminescence based on a comparison with a theoretical spectrum using their generally accepted ionization energies in low concentration ranges in uncompensated Si. The fine structure showed no dependence on dopant concentrations in the ranges from middle 1015 to low 1017 cm-3, which leads us to suggest that the dopant impurities act as isolated donors and acceptors without noticeable reduction or broadening of their energy levels due to high doping. This conclusion is, however, inconsistent with the reduction of the ionization energies observed by Hall effect measurement. A possible reason for the discrepancy may arise from inhomogeneities in the impurity density. Simultaneous appearance of the impurity-cluster bound exciton luminescenceand discrete DA pair lines supports this hypothesis.

Resume : Solar cells based on heterojunctions between amorphous hydrogenated and crystalline silicon (a-Si:H/c-Si) are of great interest due to several advantages like high efficiency (up to 24.7 %) and low temperature fabrication process. However, the properties of the interface between a-Si:H and c-Si are one of the most important factors for heterojunction solar cells efficiency. There are many interface characterization methods with different limitations and reliabilities. We here report the results of new a Si:H/c-Si interface study based on Hall mobility measurements. Recently in was found that an inversion layer in c-Si near to the a Si:H/c- Si interface occurs in case of anisotype heterojunction. The presence of this layer was confirmed by direct planar conduction and AFM conductive probe measurements. The values of band offsets were deduced from planar conduction and its temperature dependence. While there was still some incertitude due to unknown charge carrier mobility in inversion layer. In this work the mobility of charge carriers in the inversion layer of a-Si:H/c-Si interface was determined by Hall measurements at 300 and 77 K for different surface treatment and a-Si:H growth conditions. Significant difference compared to bulk material was observed. The mobility measurement results are analyzed in terms of scattering at surface defects and compared with solar cell performance. The possibility to use the Hall measurements of surface inversion layer for interface characterization are discussed in the paper.

Resume : Si-based light sources have attracted much attention in the past years because of their potential applications in realizing on-chip optoelectronic integration. In order to circumvent the limit of bulk Si material with very low radiative recombination efficiency due to its indirect band structure, various other Si-based materials, such as nano-crystalline Si (nc-Si), porous silicon and rare-earth elements doped SiO2 films have been extensively investigated. Here, we report the light emitting devices based on nc-Si/SiO2 multilayers which are prepared by annealing ultrathin hydrogenated amorphous Si/SiO2 stacked structures at high temperature. Bright light emission is achieved under the direct-current (dc) driving conditions and the turn-on voltage of the device is as low as 5V. The frequency-dependent electroluminescence (EL) behaviors are observed under alternating-current (ac) conditions. The degradation of emission intensity is less than 12% after 3 hours for ac driving condition, exhibiting the better device stability compared to the dc driving one. Moreover, P-doped nc-Si/SiO2 multilayers are also prepared with nc-Si density of 4.2x1012cm-2. It is found that P atoms are repelled out of SiO2 and incorporated into nc-Si after high temperature annealing. The ESR experiments demonstrate the existence of Si dangling bonds (Si DBs) defects at nc-Si/SiO2 interfaces, which can be reduced by P doping. The subband light emission is observed from P-doped nc-Si/SiO2 multilayers. The

Resume : Rare earth doped Silicon oxide thin films are promising materials for future Si-based light emitting devices. Their optoelectronic properties can be tuned either by using the quantum confinement in Si-nanocrystals or by using radiative transitions of the dopant. In this contribution, we investigate the structural and luminescence properties of Cerium (Ce) doped SiOx (0

Resume : Er-based, electrically driven light emitters, which can easily be integrated into Si-based circuitries, are of great interest for a broad range of applications, especially in the field of telecommunication and sensing. This work investigates the electrical and electroluminescence (EL) properties of Er-implanted MOS structures with different designs of the dielectric stack. The dielectric stack is essentially composed of a 30 nm thick SiO2 layer and a 40 nm thick host matrix for the Er ions made of Si-rich SiO2, silicon nitride or Si-rich silicon nitride. All structures implanted with Er show intense EL around 1550 nm which is excited by impact excitation of hot electrons. We compare the different host matrices regarding the EL efficiency, the EL excitation cross section, the EL decay time, the fraction of excited Er ions, the EL quenching cross section and the operation lifetime. This comparison reveals fundamental properties of the EL mechanism and addresses the current problems of this type of Si-based light emitter to achieve the performance level of compound semiconductors with a direct bandgap.

Resume : For the further development of 450 mm-diameter defect-free Si crystals, one has to take into account the impact of thermal stress on intrinsic point defect properties and behavior during single crystal growth from a melt. In this study, we evaluate the impact of thermal stress between -30 MPa (tensile) and 30 MPa (compressive), which probably covers the actual values in mass production for 450 mm-diameter crystals, on the so-called Voronkov criterion of dominant point defects. The dependence of the formation and migration enthalpies of vacancy V and self-interstitial I on the stress is evaluated by density functional theory (DFT). The obtained results are used to more accurately describe the impact of thermal stress on the critical (v/G)0. The impact of compressive thermal stress on (v/G)0 is predicted to shift the growing Si crystal more vacancy-rich, as recently confirmed experimentally by Nakamura et al (ECS Solid State Letters, 3 (2014) N5-N7). Furthermore, the dependence on the thermal stress of the process window for defect-free Si growth is clarified, which is important for the development of the pulling process of future large-diameter defect-free Si crystals. Also the mechanisms behind the experimentally observed impact of the type and concentration of substitutional dopants on intrinsic point defect behavior and formation of grown-in defects (voids and interstitial clusters) in a Si single crystal growing from a melt, are clarified. DFT calculations with 216-atom supercells are carried out to obtain the formation energies of V and I at all sites within a sphere around the dopant atom with 6 Å radius for V and 5 Å radius for I. P-type (B and Ga), neutral (C, Ge, and Sn) and n-type (P, As, Sb, and Bi) dopants are considered. The formation energies of V and I around dopant atoms change depending on the types and sizes of dopants. On the basis of the calculated results, an appropriate model of intrinsic point defect behavior in heavily doped Si is proposed. (1) The incorporated total V and I concentrations at the melting point depend on the types and concentrations of dopants. (2) Most of the total V and I concentrations contribute to Frenkel pair recombination during Si crystal growth at temperatures much higher than those to form grown-in intrinsic point defect clusters. The Voronkov model, while taking into account the present improvements, clearly explains all reported experimental results on grown-in defects for heavily doped Si. The proposed improved Voronkov model uniformly explains point defect behavior for all dopant types and concentrations, taking into account also thermal stress levels.

Resume : Dynamics of Si(100)-oxidation processes at Si/SiO2 interface and in SiO2 region are investigated using classical molecular dynamics (MD) simulations with variable charge interatomic potentials. We focus on SiO and Si emissions from Si/SiO2 interface and the following incorporations into the SiO2 and/or substrate. These typical atomic processes are successfully extracted from a lot of MD simulations in which statistical procedure is partly employed. By incorporated oxygen atoms, two-folded Si atoms are formed after structural relaxation at the interface and are emitted as SiO molecules into the SiO2. The energy barrier of the SiO emission is estimated to be 1.20 eV from the enthalpy change in an MD simulation. The emitted SiO molecule is incorporated into the SiO2 network through an Si-O rebonding process with generating an oxygen vacancy. The energy barrier of the SiO incorporation is estimated to be 0.79 - 0.81 eV from the enthalpy change. The oxygen vacancy diffusion processes leading to the complete SiO incorporation are also simulated and the energy barriers are found to be relatively small values of 0.71 - 0.79 eV from the enthalpy change. It is found that the SiO emission into the SiO2 region occurs prior to the Si emission at the ideally flat Si/SiO2 interface, which is consistent with previous theoretical and experimental studies. Our results give a unified understanding of Si oxidation processes from the atomistic point of view.

Resume : A specific feature of group IV semiconductors is the diversity of interstitial-type defects. In Si interstitial clusters can nuclear not only as dislocation loops, but also as rod-like defects. While the basic structure of large rod-like defects is known for more than two decades, the detailed mechanisms of their nucleation and growth remain uncertain. One of the commonly agreed ideas is that the rod-like defect nucleation starts with the formation of one-dimensional interstitial chains that subsequently evolve into planar defects by acquiring additional chains aside. In this presentation we combine classical molecular dynamics (MD) simulations with first-principles calculations in order to investigate the energetic proficiency and dynamics of various kinds of one-dimensional interstitial chains in silicon. We demonstrate that multiple versions of interstitial chains are mechanically stable and energetically competitive. While looking quite similar when viewed along <110> direction, these chains have completely different internal structure. Moreover, at elevated temperatures, the most energetically favourable versions of interstitial chains can transform into each other and the mechanisms of transformation are elucidated by MD. Simulations show that although long interstitial chains eventually acquire the shape of a sequence of extended split interstitials, the dynamic growth of a chain by the consecutive capture of interstitials can involve very ‘non-classical’ chain shapes.

Resume : Much of initial semiconductor research was performed on Ge. In the 1960s, interest shifted to Si due to its advantages for CMOS technology, but it is now reaching physical limits. Previously problematic properties of Ge, such as the difficulty to grow stable thermal oxides are no longer relevant and the much higher carrier mobility can be exploited. To further explore its use, one needs a knowledge of Ge properties comparable to that of Si. Given decades of experimental efforts dedicated to Si, this is not trivial. High-throughput Density Functional Theory calculations are one way to speed up generation of essential knowledge of Ge properties. The present high-throughput study focuses on one specific property: the embedding enthalpy of extrinsic point defects. This has been examined for both Si and Ge, allowing us to verify our computational method through comparison with reliable experimental data for Si, while filling in missing data for both materials. All elements from periods one through six (excluding lanthanides) were put at six different positions in the Si or Ge lattice, which was allowed to relax. The embedding enthalpy for each impurity at each site has subsequently been determined. This approach has provided a substantial dataset for further analysis. Calculated results will be compared with available experimental data. Trends through the periodic table and the degree of transferability of knowledge for Si to the corresponding situation for Ge will be discussed.

Resume : Over the past decades our daily life has been revolutionized by the invention of solid-state electronic devices. The key was the preparation of high purity semiconductors as well as the ability to control the impurity level, i.e., defects on the atomic scale. Nowadays, nanoelectronic devices more and more consist of layered structures with defect densities that limit their integrity. Further miniaturization is intimately connected with an improved understanding on the properties of atomic defects to effectively control the technological processing steps to advanced functional devices. Atomistic calculations are increasingly used to predict defect types, their stability, mobility, and electronic properties but the experimental relevance is often questionable and comparison to experimental results remains essential. In this talk our present understanding on self- and dopant diffusion in silicon germanium and its alloys derived from experimental and theoretical studies is summarized. Comprehensive experimental studies on atomic transport under various experimental conditions are presented that provide valuable information about the mechanisms of self- and dopant diffusion, the type of native point defects involved, and their interaction with dopants. This information is of fundamental significance for the development of effective doping strategies of silicon-germanium alloys and will help to improve and impede the fabrication process of group-IV based semiconductor devices.

Resume : Pushing the limits in IC-technology has raised strong interest in expanding the use of semiconductor films with more enhanced (mobility) properties ;among the popular ones are SiGe, Ge, GeSn alloys. Whereas in the past probing blanket films using 1D-metrology in terms of composition, electrical properties, strain and thermal stability was sufficient, recent evolution towards more confined and even 3D-volumes (like Finfets) has created a demand for metrology suited for very small volumes and more atomic scale observations. One obvious attempt is to abandon 1D-metrology and to focus on metrology with near-atomic 3D-resolution such as TEM-tomography and Atom probe tomography (APT). It is clear that APT is an extremely powerful method able to provide composition analysis within very small volumes ( a few nm3) with high sensitivity and accuracy as will demonstrated in this paper. This can be done not only in well defined , small trenches but also around very small crystal defects where the concentration might deviate from the bulk values. Moreover its excellent spatial and depth resolution (in many cases with the ability to resolve lattice planes) opens up the prospect to observe (off lattice site) atom migration and through statistical analysis, even cluster formation during thermal processing and strain relaxation. Despite their unique resolving power, these methods suffer from a poor productivity and a lackof statistical averaging over large areas as required in more production oriented metrology. We therefore present here also the concept of “self focusing SIMS” whereby we demonstrate that it is possible to determine, for instance, the SiGe-composition from trenches as small as 20 nm without having an ion beam with nm-resolution.

Resume : Aggressive down scaling of complementary metal oxide semiconductor (CMOS) devices call for greater gate control over channel to maintain field effect transistor (FET) current drive which brought the high-k dielectrics into picture to render higher capacitive effect and lower leakage current through increase in film thickness. High-k technology has also come across few generations where Si based dielectric is now being replaced by Hf based oxides. Furthermore, number of crystalline high-k dielectric metal oxides like Y2O3, ZrO2, Al2O3, Nd2O3, Gd2O3 have been studied in detail as the gate dielectric materials for the current and future CMOS devices. Being already crystalline, these oxides exhibit much higher thermal stability when grown epitaxially on Si substrate. However, in the present scenario when the crystalline high-k oxides are potential candidate for near future ultra-scaled devices, their long term reliability has been a serious issue to be taken care of. In this work we have carried out long term reliability study on molecular beam epitaxy (MBE) grown ternary neodymium-gadolinium oxide (NGO) thin films on Si (001) substrates with regard to their application as future gate dielectric material. Mixing Gd2O3 with Nd2O3 has an added advantage with regard to the epitaxial growth on Si substrates. Gd2O3 has a lattice constant which little constant smaller than Si while lattice constant of Nd2O3 is slightly larger. Therefore, NGO with appropriate atomic concentration would have lattice constant which is exactly matched with Si substrate. Hence, one could rule out any lattice strain at the interface that may stem from lattice mismatch. In the present work we have studied epitaxial thin film of (Nd1-xGdx)2O3 which were grown on Si (001) (1-20Ωcm) substrates using multi-chamber ultrahigh vacuum system using solid source molecular beam epitaxy (MBE) back in 2005. The detailed growth and electrical results are already available in literature (Laha et al. Appl. Phys. Lett. 88, 172107 (2006)). Standard electrical characterizations (C-V & I-V) were performed on these samples at different points of time during last 8-9 years to examine their performance as MOS capacitor as well as electrical degradation over time, if any. Recently (in 2014) the same NGO layer have been characterized using PHI2000 X-ray photoelectron spectroscopy (XPS) system in order to observe any degradation at the interface during last 8 years. The XPS results of the current data explicitly shows the NGO film grown onto Si substrate still retain excellent bulk and interface without being reacted with underlying Si substrate. As observed in 2005, the recently measured O1s spectrum shows two peaks. The intense peak at 530 eV can be attributed to oxygen in Gadolinium oxide whereas the solder at 531.4 eV indicates a binding energy much higher than the metal oxide but also quite lower than the same corresponding to SiO2 (532.7 eV). It can be attributed to the oxygen bonded in a silicate like Gd-O-Si structure. Such bonding occurring at the metal oxide and Si interface is reported several times in literature. The Si 2p spectrum showing the peak with maximum intensity at 99.05 eV definitely corresponds to bulk Si and the solder appearing at approximately 0.6 eV higher energy to that is due to the spectral decomposition into the Si (2p)1⁄2 and (2p)3⁄2 spin-orbit split components. There is also a broad peak at around 101.9 eV which is at lower energy than the position of the peak due to SiO2 (102.9 eV). So this broad peak is attributed to sub-oxide (Si+1 and Si+2) binding states which reinforces the argument of the formation of a silicate like interface. The electrical responses of the devices were initially measured in 2005 soon after the growth of epitaxial layer and also in 2009 and finally those data are compared to the same measured in 2013. No post growth treatment had been carried at any time during these years. C-V and I-V measurements carried out on these samples show promising electrical characteristics till today. A comparison of C-V characteristics measured at 100 kHz shows that there is no change in the order of the Cox values of the capacitors during these years. The CET calculated for the C-V data of the years 2005, 2009 and 2013 are 1.37, 1.92 and 2.42 nm respectively. The flat band voltage has also increased from 0.6 V to 0.87 volt. Interestingly the mid gap Dit decreased over this long period of time by almost an order of magnitude (1012 eV-1cm-2 to 1011 eV-1cm-2) although the leakage current increased more than an order. These apparently contradicting results can be explained in the following manner. As the samples were never thermally treated, the existence and increase of trap charges within the oxide is quite possible and they can play a role to increase the leakage current and flat voltage though these increases were never high enough to cause considerable degradation. We argued before the formation of a silicate like interface and it can be responsible for the decrease of the number of the unsatisfied Si bonds at the interface which naturally decreases the mid gap Dit values. It is reported earlier that the dielectric constant of the oxide is higher with silicate like interface than others such as oxide like interface. The minimum decrease in CET values over such a long period of time indicates the stability of the dielectric constant of the oxide system and silicate like stable interface is supposed to be responsible for that. In summary, we showed the stability of the oxide interface with the Si substrate over a long period of 8-9 years through XPS and electrical measurements. Therefore, recent measurement guarantee that these epitaxially grown lanthanide oxide could be the excellent candidate for future ultra-scaled CMOS devices.

Resume : In this work we review published work and present data obtained by us recently on vacancy reactions in crystalline silicon, germanium, and SiGe alloys. We combine results from deep level-transient spectroscopy, infrared absorption spectroscopy and ab-initio modeling studies to determine the structure, electronic properties, reconfiguration barriers and diffusion paths for vacancy complexes. In particular we consider the divacancy (V_2) and trivacancy (V_3) defects in Si and Ge crystals and SiGe alloys. Interactions of the V_2 and V_3 defects with oxygen and group-V impurity atoms in these materials are considered and the structure and properties of the resulting complexes are reviewed.

Resume : Beta-FeSi2 is one of promising semiconductors for the next generation IR telecom system, since resource and environmental problems will surely come to established fiber telecom technology based on photonics using InGaAs systems. However, we need further enhancement of light intensity from beta-FeSi2. In this study, we investigated enhancement of IR photoluminescence (PL) of the beta-FeSi2 nanocrystals (NCs) by doping Cu atoms. Cu films were deposited on Si layers densely embedding the NCs with the average size of ~10 nm. The diffusion rate of Cu atom was controlled by annealing. The most appropriate condition brought the largest intensities of both the intrinsic A band emission at 0.803eV by 214% and the accepter related C band emission at ~0.77eV by 582% in comparison with those of non-doped NCs. Rutherford backscattering spectrometry (RBS) revealed increase of Cu atoms in the layer with NCs with increasing the anneal time. The C band enhancement surely comes from increase of accepter levels relating to increase of Cu atoms doped in beta-lattice. Photocarrier injection measurements (PCI) is helpful to investigate a mechanism of the A band emission enhancement. The PCI measurement consists of the following processes; photocarrier excitation in Si, its migration and injection of minority carriers into the NCs and radiative recombination of carriers. Moreover, we will show results of excitation frequency dependent PCI that is of importance for discussion of the origin.

Resume : A simple approach to make untra-long Ge nanobelts under a large compressive strain is investigated with Si compatible technologies of hololithography, etching and oxidation of the SiGe on SOI substrate. Oxidation processes are performed on the SiGe stripes on SOI substrate fabricated by hololithography and reacitve ion etching. At the first stage of oxidation, Ge-riched SiGe nanobelts on the SOI substate are formed by Ge condensation. Scurled buddles of ultra-long SiGe nanobelts are lifted off from the substrate with HF etching process. At the second stage, the free-standing SiGe nanobelts are further oxidized to be condensed to nealry pure Ge nanobelts with significantly size reduction. The volume expansion of the surrounded SiO2 applies large compressive stress on the nano-belts in the directions perpedicular to the nano-belts, which enables engineering of the band gap in Ge nanobelts to extend the application range of nano devices.

Resume : Thin-film crystalline silicon solar cells have a great potential for high quality, cheap and high efficiency solar cells. Low cost of the manufacturing process is in progress. To reduce the cost of manufacturing process, the plating process was analyzed by using the simulation of stress applied to silicon substrate. In this study, the occurring of metal-silicon interface was modeled to predict the quantitatively of stress affecting inside the silicon substrate. We predicted through the changing of simulation what the factors will give a large effect on the value of the results. The stress characteristic factor was specified as a parameter in the major factor of the substance to be plated. Mechanical properties of plating materials have changed density, coefficient temperature expansion and Young’s modulus. Density is not expected to give greater effecting on stress. In the case of temperature, it was affected directly on the thermal expansion of the material. Lead to the peeling phenomenon occurred from the silicon substrate, to control the main factor, the CTE and Young’s modulus is expected. Acknowledgements: This work was supported by the New & Renewable Energy Program (No. 20123010010160) and the Human Resources Development Program (No. 20104010100580) of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korean Ministry of Trade, Industry and Energy (MOTIE).

Resume : The procedure elaboration of SiC epitaxial buffer layer preparation on silicon wafers is acknowledged to be among the most prospective technological tasks avaluable for creation of substrates suitable for GaN and AlN epitaxial films growth. The method of solid-state epitaxy in spite of the great difference in parameters of its lattices allows to form SiC/Si epitaxial heterostructure with 4H hexagonal SiC polytype on the cubic matrix [1]. This work is devoted to study by methods of X-ray diffractometry and ion beam spectrometry peculiarities of the orientational conformity of matrix and film crystalline lattice and the composition of interface between the film and wafer. Our investigations showed that the film lattice is aligned by (00l) axis in parallel with (111) axis of Si matrix. The technology of the heterostructure preparation leads to formation of double-layer coating with Si1.0C1.0 composition and 60 nanometers thickness for the upper sublayer and with Si1.0C0.6 and 42 nm thickness for the lower one. The interface between film and matrix is the silicon pore zone with thickness near 200 nm and pores concentration variation from 45% to zero. [1] S.A. Kukushkin, A.V. Osipov. New method for growing silicon carbide on silicon by solid-phase: Model and experiment // Phys. of Sol. State. v50(7). 2008. pp. 1238-1245.

Resume : In this work we report on the preparation of GeSi nanostructured films and the investigation of structure correlated with electrical behaviour. The amorphous GeSi films with Ge:Si composition of 55:45 were deposited by magnetron sputtering and were nanostructured by annealing in N2 at 700, 800 and 900 oC. XRD and TEM were used for investigation of film structure and electrical measurements (current-voltage, I–V and current-temperature, I–T characteristics) were performed. Two different structures were evidenced in the annealed/nanostructured films, one in films annealed at 700 oC and the other in films annealed at 800 and 900 oC. The 700 oC annealed films consist of GeSi nanocrystals (NCs) with sizes of 7–15 nm separated by amorphous regions of 1–2 nm. The 800 and 900 oC annealed films contain NCs with sizes of 10–30 nm and 20–60 nm, respectively, boundaries between them being completely crystallized. The NCs contain inside many defects as stacking faults and nanotwins, their density being higher in the 900 oC annealed films. The films have different electrical behaviours according with their structures. The 700 oC annealed films present superlinear I–V characteristics controlled by the high field-assisted tunnelling of carriers between NCs through the amorphous regions, while the 800 and 900 oC annealed films present linear I–V characteristics, typical for polycrystalline films.

Resume : We report the results of conductance studies in 5 – 40 µm thick Si whiskers with boron impurity concentration in the vicinity to metal-insulator transition (MIT) in the temperature range of 4.2 – 300 K, frequency range of 1 – 1×10⁶ Hz and magnetic fields up to 14 Т. The negative magnetoresistance in transverse magnetic field has been observed, and its value depends on the whisker diameter. The impedance studies allowed us to determine the whiskers’ impurity, and the charge carrier concentrations were about 5.0×10¹⁸ сm^-3 and 5.2×10¹⁸ сm^-3 for 5 µm and 40 µm samples, respectively. In the present talk we will discuss the observed differences in the whiskers’ conductance and magnetoresistance in terms of Kondo effect, which occurs at the vicinity to MIT in the crystals.

Resume : The phenomenon of total internal reflection on a slab-air interface causes that most of radiation stays trapped in the structure. Several approaches have been developed to overcome this limitation, such as random or periodical patterning of the slab surface. In the latter approach, two-dimensional (2D) photonic crystals (PhCs) with properly-designed dimensions are fabricated on the top of waveguiding layers. Then, the Bragg-diffraction of guided modes, referred to as leaky modes, allows the extraction of light into surroundings [1]. In the present study, we employed 2D PhCs with a goal to increase the extraction efficiency of light from silica (SiO2) layers embedded with nanometer-sized light-emitters, silicon nanocrystals (SiNCs). Dimensions of the photonic structures were computed by employing a computer simulation based on Rigorous coupled-wave analysis method such that a spectral overlap of the PhC leaky modes with the emission spectrum of SiNCs was achieved [2]. We fabricated columns on the top of the silica layer ordered either into square or hexagonal lattices with a typical lattice constant of the order of hundreds of nanometers. For the square lattice, we measured up to 8-fold enhancement of the luminescence extraction efficiency in the direction normal to the sample plane. For other directions, signal from the PhC was up to 3 times higher than that from a planar reference layer. [1] Ondic et al., ACS Nano 5 (2011); [2] Ondic el al., Appl. Phys. Lett. 102 (2013)

Resume : Group-IV doped silicon oxide systems are intensively studied as future materials for optoelectronic devices but, relatively, only few reports dealt with the Si- or Ge-nanoclusters (ncs) embedded in alumina or hafnia hosts. Our study represents a comparative investigation of Si- and Ge-doped oxides fabricated by RF magnetron sputtering. The effect of deposition conditions and post deposition treatment on the formation of Si- and Ge-NCs as well as on the defects in the films was investigated by means of X-ray diffraction, Transmission electron microscopy, Raman scattering and photoluminescence. It is observed that formation and crystallization of Si-ncs in silica host occurs at higher temperatures (1050-1150oC) than that in alumina or hafnia hosts (900-1000oC). At the same time, Ge-ncs are formed and crystallized at lower temperatures. The mechanism of clusters formation and light emission varies with the different matrices and a competition between different radiative channels is observed. Exciton recombination inside NCs dominates in Si-or Ge-doped-SiO2, whereas radiative recombination via interface or host defects gives main contribution in doped alumina or hafnia hosts. The possible photonic or microelectronic applications of the investigated composites are discussed.

Resume : In this study, SixGeyO1-x-y thin films are deposited by sputtering Si/Ge target. During the growth, several parameters are adjusted in order to control chemical content of the deposited films. Two-types of SixGeyO1-x-y film are grown. The first comprises 23 at.% Si, 25 at.% Ge, and 52 at.% O and the second contains 29 at.% Si, 19 at.% Ge, and 52 at.% O extracted by X-ray photoelectron spectrometry (XPS) measurement. Thermal annealing is performed at 500-700oC for 1-3 hours in vacuum. After annealing at 5000C for 3 hours, Si0.29Ge0.19O0.52 film shows two significant XRD peaks corresponding to Si (1 1 1) and Ge (2 2 0) orientation, respectively. TEM result confirms the existence of Si/Ge nano-crystals within the annealed films. The size of particles dispersed from 2 to 8 nm. The results suggest that both Si and Ge crystals are grown by annealing SixGeyO1-x-y films. With excitation at 405 nm, the 5000C-3h annealed Si0.29Ge0.19O0.52 film emit blue-light. The PL band is centered at ~483 nm. For the Si0.23Ge0.25O0.52 films, green PL is observed after annealing at 500oC for 1 hour. When annealing temperature is increased to 700oC, the PL peak is red-shifted to 635 nm and the light color is red. Our result indicates that SixGeyO1-x-y film may emit light and the emission is wavelength-tunable by adjusting chemical content as well as annealing condition. Since the SixGeyO1-x-y film could emit blue, green, or red light, it’s also potential for white-light applications.

Resume : Grain boundaries (GBs) in silicon (Si) are frequently detrimental to the performance of the Si based photovoltaic and electronic devices, via impurity gettering even when the gettering levels are low. Therefore, quantitative evaluation of the gettering ability for various GBs is indispensable to improve the performance of Si based devices. For the purpose, we have developed an analytical method to determine high-resolution three-dimensional impurity distribution at a large-angle GB by atom probe tomography (APT) combined with transmission electron microscopy (TEM), with a low impurity detection limit (lower than 10^17 cm^-3) about two orders lower than the limit by TEM, simultaneously with a high spatial resolution (< 1 nm) comparable to the resolution of TEM [1]. Our analytical method enables us to determine GBs even when their gettering abilities are negligible, and it is applied to sigma-3{111} and sigma-9{114} GBs, that are the dominant GBs in Si. Arsenic (n-type dopant), gallium (p-type dopant) and oxygen (neutral impurity) atoms are agglomerated at sigma-9{114} GBs, while they are not agglomerated at sigma-3{111}GBs. It is speculated that, those impurities are agglomerated so as to reduce the strain GB energy due to bond distortion, rather than to reduce the electronic GB energy. The gettering ability is discussed in terms of bond distortion. [1] Y. Ohno, et al., Appl. Phys. Lett. 103 (2013) 102102.

Resume : Silicon carbide (SiC) is an important wide band gap semiconductor in high-temperature, high-power applications. With continuing advances in high quality growth (3” wafers now being commercially available) and well established fabrication protocols, nano-fabrication and high-Q photonic crystals are now possible. SiC is thus a key material for future optical and electronic devices. Over the last few years a range of native defects have been found to exhibit properties that may be utilised as solid-state quantum bits. Optical read-out of an ensemble of spin states is possible even at room temperature through magnetic resonance experiments. We have shown that single defects can also be addressed. These defects are non-classical light sources that operate at room temperature. They can be formed via electron irradiation and under various conditions display interesting blinking behaviour. We observe other native defects with a broad range of wavelengths from the visible to the near infrared range, each with unique properties. We consider the luminescence of these defects with both electrical and optical excitation.

Resume : Silicon nanowire (SiNW) ensembles with different architectures have been realized using wet chemical etching of bulk silicon wafers with an etching hard mask of silver nanoparticles that are deposited by wet electroless deposition on polystyrene pattered silicon surfaces. Strong visible (red-orange) room temperature photoluminescence has been observed in wet chemically etched heavily (10^20 cm-3) and lowly (10^15 cm-3) doped SiNWs. Our observations strongly suggest that visible light emission at room temperature of SiNWs is a result of the rough sidewall structure composed of nanoscale features that make quantum confinement most probable. The anti-reflective properties of SiNWs, extremely high absorption and broader band gap structure can be the keys for the next generation solar cells based on silicon nanowires. In this paper the (n)a-Si:H/(p)c-Si heterojunction solar cells based on top-down wet chemically etched silicon nanowires with different geometries and a-Si:H growth condition will be discussed. The detailed microstructure and optoelectronic properties of the amorphous/crystallijne interface using capacitance, photoluminescence, electron microscopy, electrical measurenets etc. will be presented and discussed during our presentation. This work was supported by German Research Foundation(DFG) in the framework of the ?Semiconductor Nanostructures for Next Generation Photovoltaic? SI1893/4-1 project.

Resume : Top-down silicon nanowires (SiNWs) with diameter between 20 and 200 nm were grown on crystalline Si (c-Si) substrates with the crystallographic orientation of surfaces along (100) by metal-assisted wet chemical etching. Near-infrared photoluminescence (IR-PL) of SiNWs under excitation with a cw Nd:YAG laser at 1.064 μm was investigated. The spectral position of IR-PL of SiNWs was the same as for the corresponding c-Si substrates, but the PL intensity of SiNWs was significantly higher than that for the substrate. The IR-PL band can be attributed to the inter-band radiative recombination of photo-excited charge carriers in the volume of SiNWs. The intensity of IR-PL in SiNWs was found to be controlled by nonradiative recombination centers on SiNW’s surfaces. It was revealed that the electronic quality of employed c-Si wafers (3 independed silicon wafer suppliers) influenced the morphology and optical properties of the formed SiNWs. The proposed IR-PL approach can be applied as an effective diagnostic tool to monitor the recombination of charge carriers in SiNW-based photovoltaic devices. This work was supported by the Russian Foundation for Basic Research (grants No. 14-02-31544) and German Federal Ministry of Education and Research (BMBF) in the framework of the “NanoSemi” RUS 12/053 project.

Resume : Colloidal semiconductor quantum dots make good candidates for different applications in, e.g., photovoltaics and light emission. For these purposes, a broad spectral tunability, high radiative rates and efficient emission and absorption properties are desired. Direct bandgap materials offer these characteristics but are often toxic, scarce and/or (chemically) instable. Si does not have these issues, but optical properties of Si QDs are commonly inferior. This changes upon butyl termination: a direct bandgap-like structure appears due to a combined effect of quantum confinement and Si-C bonds at the surface. An enhanced radiative rate, broad spectral tunability and efficient fast emission have been experimentally observed and confirmed by theoretical modeling[1]. Here, we investigate intraband transitions in butyl-terminated colloidal Si QDs. For that purpose, we use ultrafast transient induced absorption (TIA) spectroscopy. The time-dependent evolution of absorption spectra provides information on relaxation and recombination processes within the excited states of QDs. This is done by probing intraband transitions of free carriers generated by the pump pulse. We demonstrate that the direct bandgap-like character of the material is experimentally confirmed by TIA. We discuss implications of these findings for new applications of these “direct bandgap” colloidal Si QDs for photovoltaic and light emitting devices. [1]K. Dohnalova et al., Light: Science and Applications 2013

Resume : In-situ photoluminescence (PL) spectroscopy was used to study silicon nanowires (SiNWs) growth during metal-assisted chemical etching (MACE) of crystalline Si (c-Si) wafers in solutions based on H2O2/HF. SiNWs were grown on с-Si substrates with different crystallographic orientations of surfaces ((100) and (111)), conduction types (p, n), and doping densities (from 0.1 to 5 Ω•cm) covered by Ag nanoparticles (deposition from AgNO3/HF based solutions). Ultraviolet (337 nm) and near-infrared (902 nm) pulsed laser sources were used for the PL excitation. In-situ PL measurements were carried at the wavelength of 1130 nm (1.1 eV, band-gap of c-Si) at room temperature. The PL emission at 1130 nm is attributed to the interband radiative recombination of photo-excited charge carriers in the volume of c-Si and SiNWs. In addition, the prepared SiNWs were studied by means of scanning electron microscope, ex-situ PL and Raman scattering spectroscopy. The prepared SiNWs was found to exhibit also a PL band in the spectral range of 400-700 nm, which is explained by radiative recombinations of excitons confined in small Si nanocrystals on SiNW’s surfaces. A strong influence of c-Si substrate properties and etching parameters on the PL intensity of SiNWs was revealed. The obtained results allow us to draw conclusions about processes, occurring during the growth of SiNWs.

Resume : Combination of optical and electronic components for reduction of heat release and increase of data transfer rate is one of the most important problems of modern communication technologies . This necessitates the development of of CMOS-compatible optical materials. Ge is one of the basic materials for electronics has indirect band gap thus giving no luminescence based on effective recombination. This defines general application of germanium as passive elements for IR-optics, AIIIBVI solar cells, etc. On the other hand small energy difference between direct and indirect transitions (0.34 eV) enables one to increase the possibility of pair recombination through quantum confinement by dielectric matrices. The work is focused on formation of germanium nanostructures in porous anodic alumina matrices and investigations of their luminescent characteristics. It was shown that under excitation with 473 nm laser samples exhibit intensive luminescence with the wavelength of 484 nm. To describe the nature of Ge luminescence in porous matrix an extensive study of the chemical composition and the structure of composites was performed. It was shown that the luminescence is caused by the formation of defect Ge(II) emissive centers. The work is supported by UB RAS (projects № 12-C-2-1024, 14-2-NP-179) and Russian Ministry of Education and Science.

Resume : The samples of InAs quantum dots embedded in InGaAs/GaAs QWs were created at different temperatures of QD growth from 470°C up to 535°C by self-assembled Stransky-Krashtanov growth mode using the MBE technique. The photoluminescence dependence on temperature and excitation power of measurements and the non homogeneity of photoluminescence along the wafers in QD structures have been studied.The PL non homogeneity is characterized by 2 reasons: a) the PL intensity variation without the change of PL peak positions in structures with QDs obtained at 470 and 490°C is related to the variation of QD or nonradiative center concentrations along the wafer, and b) the PL intensity variation with the change of PL peak positions in structures with QD obtained at 510, 525 and 535 °C is related to the variation of QD sizes along the wafer. This effect is due to more deep localized energy levels in QD in the center of wafer in comparison with the energy levels in QD at the periphery. The study of PL intensity variation in QD obtained at 470, 490, 510, 525 and 535°C has shown that 3 reasons exist: a) the high concentration of nonradiative recombination centers in capping layer In0.15Ga1-0.15As for QD obtained at low temperature of 470°C, b) the dispersion of the density and the size of QD obtained at 490, 510 and 525°C, c) the high concentration of nonradiative recombination centers in the barrier layer GaAs for QD obtained at 535°C.The dependence with temperature and excitation power of the PL spectra of InAs quantum dots embedded in InGaAs/GaAs quantum well has been studied.A set of equations for dynamic excitons were resolved to analyze the mechanism of thermal decay of the PL in InAs quantum dots. It was shown that the exciton nature of carriers is important to capture processes and thermal escape to and from the quantum dots. It was observed that at low temperatures the thermal decay of the ground state and excited states PL spectrum is attributed to the decreased flow of excitons to quantum dots due to thermal escape of excitons from the GS of the WL (120-200 K) or from the GS of the InGaAs capping layer (200-250K) to the GaAs barrier with subsequent NR recombination centers. At high temperatures (250-300K) the thermal activation energy of the PL decay depends on the density of the quantum dots and the quality of the structures. In structures with high emission (QDs grown at 490, 510 and 525°C) the activation energy is close to the difference between the energy bandgap of GaAs and GS quantum dots. In structures with low emission (QDs grown at 470 and 535°C) the thermal activation energy is close to the difference between the GS of the InAs QDs and the GS of the InGaAs capping layer.

Resume : Group-IV semiconductor nanowires (NWs) are attracting the interest of a wide scientific community as building blocks for a wide range of future nanoscaled devices. We show that metal-assisted chemical etching of Si substrates is a powerful technique to obtain nanometer-size, high density and low-cost Si NWs with high and controllable aspect ratio. NWs obtained by this technique have exactly the same structure and doping of the substrate and present quantum confinement effects. Photoluminescence (PL) emission is very bright and tunable with NWs size according to quantum confinement. Light emitting devices based on Si NWs, showing an efficient electroluminescence emission at room temperature under low voltage excitation, have also been realized. The same synthesis approach has been also used for the realization of Si/Ge NWs, in order to obtain different confined structures of both Si and Ge inside each NW. PL emission properties of Si/Ge NWs are presented and discussed. Moreover we demonstrate that the design of new textures of NW materials and the optimization of size and spatial arrangement play a key role on the improvement of the optical properties, such as light trapping and multiple scattering phenomena. The relevance of the reported results and their perspectives to open the route towards novel applications of NWs in photonics, are finally discussed.

Resume : Nanowires are among the most promising nanostructures that offer a multi-functional potential covering a broad range of nanotechnological developments in the strategic sectors of electronics, energy and biology. Due to size-related specific behaviours of nanoscale objects, innovative heterostructures can be formed in nanowires, not only from different materials but also from different polytypes of the same material. In this paper, we report on a stress-induced martensitic phase transformation in Ge nanowires that is believed to result from a nanoscale size effect. <111>-oriented Ge nanowires with standard diamond structure (3C) undergo a phase transformation toward the hexagonal 2H-allotrope namely the lonsdaleite phase. The phase transformation occurs heterogeneously along the length of the nanowire. It results in an unprecedented heterostructure with embedded Ge-2H domains distributed all along the Ge nanowire. Occasionally, the 4H-allotrope was also identified. This novel allotropic Ge heterostructure may have very interesting semiconductor and optical properties opening new possibilities of applications of group-IV materials in next-generation devices. Since 3C and 2H phases have different band structures and charge density, the transformation may alter significantly the electrical, thermal and optical properties of the nanowires. The literature suggests a type-I alignment band. Additionally, the 2H-allotrope is expected to present a small direct band-gap. Thus, the 3C/2H heterostructure may be optically active in the infrared region and can be a promising candidate for mid-IR detection. Enhanced optical emission and absorption are also expected from confinement effects in the nanostructures. Finally, the periodic formation of phase boundaries should result in a strong reduction of thermal conductivity while electronic transport could be preserves, what assigns 3C/2H Ge nanowire-based devices as very promising for thermoelectricity. In the perspective of the mentioned applications, we have studied the thermal stability of the 2H domains. The recrystallisation under annealing was followed in real time by in-situ TEM up to 650?C. Instability appears between 500 and 600?C. Some domains remain stable with the 2H allotropic form up to 650?C while other domains fully recrystalize toward the 3C phase (not depending on the size). This recrystallization can be quite sudden and appears in a small thermal range. This transformation was observed on nanowires with diameter above 20 nm. It is worth noting that for lower diameters the nanowires sublime under vacuum at 600?C. Electron beam induced current (EBIC) and optical absorption measurements are in progress to investigate electrical and optical properties of these promising allotropic heterostructured Ge nanowires.

Resume : Silicon germanium nanowires (SiGe NWs) have acquired today a prominent role in several cutting-edge research topics in nanoscience [1], thanks to the latest relevant advances in synthesis, processing and characterization. Their physical properties are strictly related not only to the size of the system (like the corresponding pure Si and Ge wires), but also to the relative composition of Si and Ge atoms, and to the geometry of Si/Ge interface. Substituting some of the atoms of a pure Si NW with Ge in random as well as ordered configurations of different compositions, can strongly affect band gap, effective mass, phonon and electron scattering processes. As a consequence, SiGe NWs are the target of the most exciting technological applications in the field of high performance nanoelectronics, thermoelectrics, superconductivity and spintronics. Furthermore their investigation aids in deeper insights into basic research in material science. In this presentation I will review both the achieved milestones and the current research efforts, focusing on theoretical modeling. The investigation of the matter at nanoscale with experimental techniques is often complicated by several factors not always well controlled, which can hide the right comprehension of the basic properties. In this context theoretical modeling and simulations are extraordinarily important as they can complement or augment experimental observations. [1] M. Amato, et al., Chem. Rev. (2013) DOI: 10.1021/cr400261y.

Resume : Converting indirect bandgap semiconductor into a direct bandgap one is a challenging, but highly desired task. For example, direct bandgap form of silicon would enable much more efficient and lightweight solar cells, detectors, displays, cheap light sources and monolithically integrated micro-electro-photonic chips. For germanium, direct bandgap properties were achieved by doping combined with strain. For silicon, however, the same approach did not lead to similarly spectacular results. In our research, we found another approach - using electronegative capping, electronic density is manipulated in the core of silicon nanoparticles (SiNP) in such a way that “direct bandgap states” emerge, i.e. conduction band states located in Gamma valley at a size-tunable band-edge energy. Resulting material offers full spectral tunability from UV/blue to red with about 1000 times higher radiative rate and 10-100 times stronger band-edge absorption, both when compared to the “standard” SiNPs capped with hydrogen and/or silica oxide. In the presentation we will show experimental evidence of this structural transformation and support our findings by theoretical modeling. Since the achieved material is non-toxic, i.e. does not pose health or environmental risks, it can form a basis for the new type of emerging applications, where NPs are used as the active emitting/absorbing element, such as quantum dot-displays, solar cells, functional coatings, in medicine or in many other areas.

Resume : Self-assembled growth of Ge quantum dot lattices in oxide matrices prepared by the quite simple magnetron sputtering deposition method allows the preparation of a variety of structures tunable by their shape, size and arrangement. The driving mechanism for the self-assembly was attributed to the surface morphology features originating from the quantum dots? growth. Here we present specifically that the matrix type is another critical factor that enables the control of the self-assembly process, the tuning of the ordering type and degree of regularity of quantum dot systems. The effectiveness of the matrix factor is demonstrated through the analysis of quantum dot arrangements in amorphous silica, alumina and mullite matrices. Using the same deposition conditions, different ordering types and degrees of disorder were found in the quantum dot systems based on different matrices. The matrix factor is shown to be driven by the different matrix tendencies to smooth the surface during the films growth. The obtained results are relevant for the understanding and tailoring of the self-assembled growth of quantum dot lattices in amorphous systems.

Resume : Carrier mobility is one of the most important parameter of any semiconductor material, determining its suitability for applications in a large variety of electronic devices including field effect transistors (FETs). Today the capabilities of modern planar Si FET devices are almost exhausted and re-searchers are seeking either new device architectures or new materials. Here we report an ex-tremely high room temperature 2D hole gas (2DHG) drift mobility of 4500 cm2V−1s−1 at a carrier density of 1.2×1011 cm-2 obtained in a compressively strained Ge quantum well (QW) heterostruc-ture. They were grown in an industrial type reduced pressure chemical vapor deposition (RP-CVD) system on a standard Si(001) substrate. The low-temperature Hall mobility and carrier density of this structure, measured at 333 mK, are 777,000 cm2V-1s-1 and 1.9×1011 cm-2, respectively. These hole mobilities are the highest not only among the group-IV Si and Ge based semiconductors, but also among p-type III-V and II-VI materials. The obtained room temperature mobility is sub-stantially higher than those reported so far in strained Ge QW heterostructures and reveals a huge potential for further applications of this material in a wide variety of electronic devices.

Resume : The renewed interest in Germanium as base material for electronic applications has stimulated extensive experimental and theoretical studies. Successful integration of Ge in nanoelectronic devices requires fundamental understanding of ion-implantation-induced target modification and damage. In this contribution the temperature dependence of ion-beam mixing induced by 310 keV gallium (Ga) ion implantation in crystalline and preamorphized germanium (Ge) is reported. Isotopically enriched multilayer structures of alternating 70Ge and natGe layers are used to visualize the self-atom mixing. The distribution of the implanted Ga atoms and the ion-beam induced self-atom mixing was determined by means of secondary ion mass spectrometry. Different temperature regimes of self-atom mixing are observed. At temperatures up to 423 K the mixing is independent of the initial structure whereas at 523 K the intermixing of the preamorphized Ge structure is about twice as high as that of the crystalline material. At 623 K the intermixing of the initially amorphous Ge structure is strongly reduced and approaches the mixing of the crystalline material. The temperature dependence of ion-beam mixing is consistently described by competitive amorphization and recrystallization processes.