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

Resume : Group-IV alloy semiconductors such as GeSn and SiGeSn have been much attracted in recent years. This is because high-Sn-content GeSn-related alloy has a unique electronic properties; being direct transition semiconductor, high carrier mobility, low thermal conductivity, low growth temperature, preferable stressor for strained Ge, etc. For practical application, there is a challenge for enhancing the Sn content in GeSn-related alloys. Some groups and we have been developing the synthesis of GeSn and related thin films with a high Sn content over the thermal-equilibrium solid solubility of Sn. Recently, we achieved GeSn, SiSn, and SiGeSn epitaxial layers with a Sn content as high as a few tens %. The key factors of the epitaxial growth for high-Sn-content GeSn layer are the low temperature growth, controlling chemistry, and strain engineering. Engineering dopants, vacancy, and defects for GeSn epitaxial layers is also essentially for electronic applications. In addition, interface engineering for GeSn is required for establishing electronic devices. We are developing material and process technologies for p/n junction, metal/GeSn contact, and MOS interfaces. In addition, heterostructure engineering with GeSn/SiGeSn should be developed for not only electronic but also optoelectronic applications. We recently demonstrated a type-I energy band alignment with no or small strain GeSn/SiGeSn heterojunctions, which promises wide applications for various nanoelectronic devices.

Resume : Direct bandgap materials in the mid-IR range that are compatible with the silicon based semiconductor technology are desirable components for optoelectronics and photonics. A promising candidate to achieve this goal is a germanium-tin alloy. However, the theoretically required Sn content of ~ 9 % to convert Ge to a direct bandgap material exceeds the solid solubility limit of < 1 % according to the binary phase diagram.[1, 2] Therefore thermodynamically controlled processes are not suitable for the growth of this metastable material composition. This contribution focuses on the growth of Ge1-xSnx nanostructures with Sn concentrations above the threshold of the indirect-to-direct transition. Ge1-xSnx nanowires and nanorods are synthesised in a low-temperature solution-based process using metalorganic precursors in combination with microwave-assisted heating enabling growth by kinetically driven processes.[3, 4] The process parameters allow the incorporation of up to 28 % of Sn into the Ge lattice, paving the way to potential bandgap engineering of bottom-up grown Ge1-xSnx nanowires. Independent and complementary methods including X-ray diffraction, Raman spectroscopy, and EDX elemental mapping are used to determine the Sn content, while TEM confirms the high crystallinity of the obtained products. In addition, variable-temperature XRD measurements allow detailed studies in regards to the thermal stability of the synthesised metastable alloy composition. These studies reveal distinct differences in stability and segregation behaviour depending on the composition of the material. Absorption measurements using Ge1-xSnx alloys with different Sn contents reveal direct bandgap transitions with composition-dependent bandgap energies in the energy range < 0.5 eV. The results presented in this contribution show new features of a low-temperature nucleation regime and the impact on the obtained Sn concentration and direct bandgap energy of the nanostructures as well as their thermal stability.[5] These results will help to understand the nucleation and growth process of this highly interesting material in absence of an external template. [1] Wirths, S.; Geiger, R.; Von Den Driesch, N.; Mussler, G.; Stoica, T.; Mantl, S.; Ikonic, Z.; Luysberg, M.; Chiussi, S.; Hartmann, J., Lasing in direct-bandgap GeSn alloy grown on Si. Nature photonics 2015, 9 (2), 88-92. [2] Olesinski, R. W.; Abbaschian, G. J., The Ge−Sn (Germanium−Tin) system. Bulletin of Alloy Phase Diagrams 1984, 5 (3), 265-271. [3] Barth, S.; Seifner, M. S.; Bernardi, J., Microwave-assisted solution–liquid–solid growth of Ge 1− x Sn x nanowires with high tin content. Chemical Communications 2015, 51 (61), 12282-12285. [4] Seifner, M. S.; Biegger, F.; Lugstein, A.; Bernardi, J.; Barth, S., Microwave-Assisted Ge1–x Sn x Nanowire Synthesis: Precursor Species and Growth Regimes. Chemistry of Materials 2015, 27 (17), 6125-6130. [5] Seifner, M. S.; Hernandez, S.; Bernardi, J.; Romano-Rodriguez, A.; Barth, S., Pushing the Composition Limit of Anisotropic Ge1–xSnx Nanostructures and Determination of Their Thermal Stability. Chemistry of Materials 2017, 29 (22), 9802-9813.

Resume : Most optoelectronic devices are based on direct band-gap materials i.e. group III-V, which are expensive. An urge to find an alternative to III-V materials which would be compatible with Si has gained more importance. For instance, Germanium-Tin (GeSn) alloy is a tunable band-gap material that undergoes an indirect-to-direct gap transition for Sn concentrations above ~ 10 at.% [1], and it is compatible with Si-based technologies [2]. Fundamental problems arise from the large lattice mismatch between Ge and Sn (15 %), the low solubility limit of Sn in Ge (< 1% at.), and a tendency to phase-separation. This work reports on the fabrication of germanium-tin nanowires by a mechanism known as Solid-Liquid-Solid [3]. This process consists of i) deposition of a catalyst (In or Sn) nanodrops on a substrate such as a c-Si wafer, ii) hydrogen plasma treatment of catalyst nanodrops in a Plasma-Enhanced Chemical Vapor Deposition reactor, iii) deposition of an amorphous Ge layer and iv) annealing the sample at a temperature above the eutectic point to activate the growth. The influence of different growth conditions on the NWs synthesis will be presented. In particular we will show that this fabrication process allows obtaining Ge-Sn nanowires with a uniform distribution of Sn, with concentrations in the range of 20 at. %. 1. Coppinger, M., et al., Applied Physics Letters, 2013. 102(14): p. 141101. 2. Liow, T.Y., et al., IEEE Journal of Selected Topics in Quantum Electronics, 2010. 16(1): p. 307-315. 3. Yu, L., et al., Physical Review Letters, 2009. 102(12): p. 125501.

Resume : Production of SunPower’s 25% efficient Maxeon cell technology has now surpassed 100 million wafers per year. Ensuring a stable supply of incoming silicon wafers capable of maximizing the cell architectures efficiency potential, is of paramount importance in high-volume manufacturing. In this article, we will examine three areas which impact the selection, specification and supply of defect lean silicon for high efficiency photovoltaic applications. First, we discuss how different applications impact the type of silicon selected. Here we assess the considerations that go into specifying wafers to be used for champion solar cell devices versus wafers needed to support high volume manufacturing. We also explain how different cell architectures – like n-PERT, SiHJ and IBC – influence silicon specifications. Next, we examine gaps in the academic literature which impact silicon specifications. The focus of this section is on the impact of the Auger parameterizations and Shockley Read Hall parameters for defects, on the loss analysis used to understand how silicon specifications can be adjusted to maximize device efficiency and electrical yield. Lastly, we assess the limitations of current state-of-the-art wafer metrology tools available for identifying and monitoring defects in n-type silicon, and present TEM images of an oxide precipitate.

Resume : A model for the defect underlying the light- and elevated temperature induced degradation (LeTID) of the charge carrier lifetime of multicrystalline (mc) Si is proposed. It is based on the conjunction of recent investigations of LeTID and observed similarities to phenomena in monocrystalline Si. This includes investigations of the Shockley-Read-Hall recombination parameters and a multitude of phenomenological studies. Findings from literature are presented alongside our experimental studies and discussed in terms of insights into the defect precursor species and contributing impurity species. Several metallic impurities suggested in literature to cause LeTID in mc Si can be discarded from detailed spatially resolved investigations and discussion of their solubility and diffusivity. We observe striking similarities of LeTID and the LID of floatzone (FZ) Si under similar conditions addressed to as FZ-LID. Also, LeTID has been observed in Czochralski grown Si lately. It thus appears reasonable to suspect the LID phenomena to arise from the same or similar defects in the silicon bulk. The high purity of FZ Si confines the potential defect precursor species and opens a pathway to well-defined investigations of the LeTID effect. Conjunction of the available insights suggests that hydrogen is a likely candidate for a defect precursor. In this model, the degradation upon illumination is caused by the binding of H to another precursor due to charge state changes of H and/or the second species X under carrier injection at elevated temperature. The observation of a subsequent recovery of the bulk lifetime upon prolonged perpetuation of the conditions then represents the sinking of H or X in energetically more favourable locations or complexes.

Resume : Passivating contacts of crystalline silicon (c-Si) solar cells with a poly-crystalline silicon (poly-Si) layer on a thin silicon oxide (SiOx) film lead to the decrease of recombination at the metal/c-Si interface. A promising way to form a hole-selective poly-Si contact is to deposit a boron-doped amorphous silicon layer by Plasma Enhanced Chemical Vapour Deposition (PECVD) on top of SiOx, followed by an annealing step. However, if the deposition step is not optimised, the release of hydrogen causes blistering of the poly-Si layer. The annealing temperature (Ta) has to be tuned as well to crystallize the layer and activate dopants without degrading SiOx and surface passivation. Samples were prepared by growing a SiOx (1.4 nm) in ozonized DI-H2O followed by forming an in-situ boron doped poly-Si layer (30 nm). First the PECVD conditions were adjusted in order to reduce blistering of the poly-Si and to improve its conductivity. Then the effect of Ta on the surface passivation properties of the poly-Si was studied. An optimum of implied open circuit voltage (iVoc) was found for Ta=800°C. The local photocurrent of samples was analysed by conducting probe atomic force microscopy measurements under illumination and the boron diffusion in the substrate was analysed by electrochemical capacitance-voltage profiling. Furthermore, adding a hydrogen-rich silicon nitride layer deposited by PECVD on top of the poly-Si helped achieving iVoc of 720 mV suitable for high-efficiency solar cells.

Resume : Silicon (Si)-wafer based photovoltaic (PV) technologies currently lie at the heart of PV market on the strength of their high efficiency and a continuous fall in the production costs. The efficiency improvement of Si solar cells is one of the key challenges to a further rise of the PV’s share of global electricity. However, the record efficiency is near the theoretical limit and a significant increase in efficiency is unlikely to be achieved due to the practical Auger limits. Combining Si and other materials exhibiting complementary spectral response is one of the promising approaches to improving the overall efficiency. In particular, perovskite solar cells are the most prominent candidate as a top cell in the Si-based tandem architectures, potentially leading to a boost in the efficiency beyond the Shockley-Queisser limit. In this study, we report a critical aspect associated with the recombination junction for constructing efficient monolithic perovskite/Si tandem devices. The Si bottom cell used here is the standard Al-BSF p-type Si solar cells with an efficiency of ~14%, where a transparent conductive oxide layer was deposited onto the n+ emitter for creating an intermediate recombination junction of the monolithic tandem devices. The recombination junction was deliberately modified with a thin organic layer to improve the compatibility between sub-cells, resulting in the substantial mitigation of interfacial resistance. As a result, in combination with a solution-processed perovskite top cell on the modified recombination junction, we demonstrated a monolithic perovskite/Si tandem device with an efficiency of 21.5%.

Resume : The most common substrate currently used to produce solar cells is boron doped p-type silicon. Boron doped silicon is susceptible to light induced degradation (LID), whereby boron-oxygen complexes form as recombination centres and reduce cell efficiencies. Several methods have been suggested to reduce or eliminate the effects of LID and one of them is the use of alternative p-type dopants. Indium doped silicon is a possible candidate and this study aims to ascertain its fundamental limitations. We have measured the effective carrier lifetime in indium doped samples passivated by silicon nitride and a room temperature superacid method. We find the recombination rate in the indium doped samples to depend systematically on the dopant concentration. The relatively deep acceptor level of indium (0.15 eV above the valence band) means there is a significant fraction of unionised indium present at typical operating temperatures, and we suggest that this is responsible for the recombination effect. Although this recombination is a fundamental feature of indium doped silicon, we show that, under certain circumstances, the lifetime for certain relevant doping levels can be higher than boron doped silicon after LID. Therefore, a window of opportunity may exist to use indium doped silicon in high efficiency p-type solar cells.

Resume : Electronic devices generate unwanted heat which needs to be removed. This often involves a heat front interacting with the interface between Si and some other material connected to a heat sink. Sophisticated techniques such as thin water pipes, microfluidic channels, kinetic cooling engines, nanowires, or carbon nanotubes, add complexity (and cost) to devices. The research in this area has also involved trail-and-error experimentation with high-thermal-conductivity layers (such as crystallized C) or empirical molecular-dynamics (MD) simulations which often include thermostats to control the temperature fluctuations. The present ab-initio MD study provides some atomic-level insights into the interactions between heat flow and interfaces. A simple ‘thermal circuit’ is embedded in a slab of Si, starting at some high temperature. The circuit consist of a straight wire of material X (in the present case: Ge, C, or Si itself) from which heat is periodically removed, and the temperature of the Si slab is monitored vs. time. The geometry we select is theoretically convenient and could well be an experimental nightmare. However, the physics of Si, X, and the Si|X interface are the same for an X layer, bump, plug, nano-insert, or any other geometry. Indeed, we still deal with moving heat from Si through the Si|X interface, into X, which is connected to a heat sink. We propose a new method to calculate the thermal interface (or Kapitza) resistance directly from the MD data.

Resume : Identifying and quantifying point defects in semiconductors is a complex and long standing problem. In this talk I will present an unconventional approach to elucidate the presence of different point defects, by combining ab initio thermal transport calculations and previous experimental measurements. I will illustrate this for two materials of industrial interest: silicon carbide (SiC), and gallium nitride (GaN). For cubic SiC, our comparison between predicted and measured values unveils an extraordinarily strong scattering due to boron doping. This surprising effect is due to resonant scattering, which can be traced to a broken symmetry around the substitutional B impurity. [1] For GaN, our approach allows us to provide conclusive evidence of the recent hypothesis that gallium vacancies in ammonothermally grown samples can be complexed with hydrogen. Our calculations for O-doped and Mg-O co-doped samples yield a consistent picture interlinking dopant and vacancy concentration, carrier density, and thermal conductivity, in excellent agreement with experimental measurements. [2] These results highlight the predictive capability of ab initio phonon transport modeling, and its value for understanding and quantifying defects in semiconductors. All the calculations presented were obtained with almaBTE (www.almabte.eu) [3], and supported by the H2020 programme through project ALMA. (1) Katre, A.; Carrete, J.; Dongre, B.; Madsen, G. K. H.; Mingo, N. Phys. Rev. Lett. 2017, 119 (7), 75902. (2) Katre, A.; Carrete, J.; Wang, T.; Madsen, G. K. H.; Mingo, N. arXiv:1712.08124 [cond-mat] 2017. (3) Carrete, J.; Vermeersch, B.; Katre, A.; van Roekeghem, A.; Wang, T.; Madsen, G. K. H.; Mingo, N. Computer Physics Communications.

Resume : Understanding and control of point defects are of decisive importance for present and future use of silicon in electronics and photovoltaics. Intrinsic defects, formed during crystal growth and/or device processing, interact strongly with common residual impurities like oxygen (O), carbon (C), hydrogen (H), transition metals (e.g., Fe) as well as with dopants (e.g., B and P). Most of these defect complexes are electrically active with deep states in the bandgap and in order to minimize their adverse effect on device performance, they need to be controlled at concentrations on the order of ~1010 cm-3 (or sometimes even lower). In this contribution, an overview is given of recent results on (i) the migration of the double negatively charged mono-vacancy (V2-) in n-type silicon and determination of its absolute diffusivity values, (ii) low-temperature annealing of divacancy centers through interaction with mobile diatomic hydrogen molecules (H2) in n-type material, and (iii) annealing kinetics of interstitial carbon ? interstitial oxygen centers (CiOi) evidencing formation of carbon-dioxygen complexes in moderately doped p-type material. The main characterization techniques used were deep-level transient spectroscopy, including minority carrier transient spectroscopy, and Fourier Transform Infrared spectroscopy. When applicable, the experimental data were compared with kinetics simulations results employing the theory for diffusion-limited reactions.

Resume : It is well known that thermal donors (TDs) are formed in silicon crystals during annealing at 400-500oC. TDs are considered to be oxygen clusters whose formation can be retarded and enhanced in the presence of excess vacancies [1] and silicon self-interstitials [2], respectively. Voronkov et al. suggested that the formation of the oxygen clusters is enhanced by self-interstitials emitted from oxide precipitates [2]; however, the enhancement is not quantitatively revealed. In this work, Czochralski silicon crystals were grown with different grown-in defect regions such as voids and oxide precipitates. The wafers taken from the crystals were annealed at 450oC for 4 to 40 h in a nitrogen ambient. The concentration of TDs formed at 450oC was estimated from the electrical resistivity obtained by four-point probe measurements. It is found that the formation rate of TDs increases with increasing density of oxide precipitates in silicon wafers with different grown-in defect regions. The thermodynamic model for TD formation shows that the small oxygen clusters as nuclei of TDs in as-grown crystals increase with increasing oxide precipitates, suggesting that the formation of the nuclei is enhanced due to self-interstitials emitted by oxide precipitations during crystal growth. [1] M. Tajima, et al., Appl. Phys. Lett., 65 (1994) 222. [2] V. V. Voronkov, et al., Solid State Phenom., 131-133 (2008) 387.

Resume : A large quantity of digital images are generated daily for material characterization and process monitoring/control in semiconductor research and manufacturing. Various types of image capturing tools and image generation tools are used. Cameras (i.e. image sensors) are used in optical characterization in X-ray, UV, VIS, IR wavelength ranges. Electron microscopy also uses image sensors to visualize the interaction between electrons and materials of interest. Atomic force microscopy (AFM) and wafer mapping tools also generate pseudo digital images. Computer aided simulation and various analysis results are also exported as artificial digital images to visualize the results. Every tool provides its own image export and analysis functions. Interpretation of identical digital image data is often subjectively, qualitatively and/or empirically analyzed. The interpretation of the data is often quite different among researchers, scientists and engineers, depending on their experience and areas of interest. Only a very small fraction of digital image data is utilized due to the lack of user friendly image processing software. It would be beneficial to develop end user friendly, unified image processing software which can support various digital image formats (such as BMP, JPG, PNG, GIF, TIF, DM3 etc.) from various image capturing and export tools. For example, shape, size, area, surface area, volume and profile of etch pits cannot be easily extracted with ordinary characterization techniques. We often rely on our instinct and subjective judgment, even in scientific research. Useful data mining capability from ordinary digital images can greatly help researchers, scientists, engineers and students to effectively use the digital image data and communicate data objectively. Statistical analysis of data gathered from ordinary digital images will enhance our understanding of the research subject. Correlation between simulation and experimental results can also be quantified. In this paper, a newly developed end user friendly, unified image processing software (PicMan from WaferMasters, Inc.) is introduced, along with a few digital image examples in silicon carbide (SiC) and related materials research. Suggestions of the effective use of digital images in terms of quantitative and statistical analysis will be introduced using sample images. The image processing software enables end users to analyze digital images of any format, on their own PCs, very easily and accurately. Improved digital image analysis by end users will greatly improve the understanding of characterization results and shorten development cycles.

Resume : Frenkel pair (FP) (vacancy V and self-interstitial I) formation and annihilation in Si crystals was analyzed by first-principles calculation, and the impact of accumulated interstitial oxygen (Oi) atoms was studied. The formation energy (Ef) of all possible configurations of I around V was calculated after geometry optimization using 64-atom Si cubic cells. The LST-QST method was used to obtain the diffusion barrier of I from the remaining V. The four-fold coordinated defect (FFCD) → I at bridge site of V → I at the 3rd tetrahedral (T)-site from V → I at the 4th(b) [110] dumbbell (D)-site → I at the 3rd(c) [110] D-site → I at adjacent [110] D-site → … was the best diffusion path of I from V. The calculated Ef of FPs increased to ~7 eV (sum of calculated Ef of independent V and I) when I reached the 4th(b) D-site (~5 Å) from V. The formation (vibration) entropy (Sf) increased to ~15 kB (sum of calculated Sf of independent V and I) when I reached the 4th(b) [110] D-site. The calculated results indicate that the average Ef of independent V and I (3.5 eV) dominates the thermal equilibrium concentration of V (concentration of I) formed by the Frenkel mechanism for all possible configurations of I around V with an actual concentration in Si. The calculated diffusion barrier of I around V was ~0.3-0.9 eV, which is ≤ the diffusion barrier of independent I in Si. Since the thermal equilibrium concentration of V (I) formed by the Frenkel mechanism is comparable to that of independent V and I in Si, FP formation easily occurs within Si wafers when the temperature rapidly increases. The other interesting results were (1) the effective barrier of pair annihilation was evaluated almost zero, (2) the I up to the 3rd T-site and V in the cell annihilated in 0.5-1 ps at 1350oC by first-principles MD calculation, and (3) it was easier to form I with the accumulation of Oi atoms.

Resume : 3 dimensional Fin structured P type (source)-Intrinsic (gate)-N type (drain) FET (3D P-I-N FET) on silicon on insulator (SOI) has been suggested for capacitor-less 1 transistor embedded dynamic random access memory (eDRAM). The gate length and the thickness of Fin gate channel on SOI is 10 and 15nm, respectively, and the intrinsic gate is divided into the partially gated and the ungated channels. DC and transient memory characteristics are measured and analyzed with the Silvaco ATLAS3D tool. The 3D P-I-N FET sharply turns on at very low voltage, 0.75 V and the on/off current ratio is 3 order of magnitude. Memory operation successfully shows that the memory states regarding writing ?0?, reading ?0?, writing ?1?, and reading ?1? strongly depends on the current level by the injected rate through the potential barriers on the drain/source junctions, and the retention time is about 100ms. We have analyzed the potential distribution along the gated and ungated channel and discussed the relationship between memory characteristics and the potential distribution.

Resume : The continuous improvement of photovoltaic technologic performance pushes more and more towards the use of new materials and new technological approaches aimed at overcoming the physical limits of the materials currently in use, such as silicon. One approach is to integrate III / V materials on silicon in order to realize monolithic tandem structures that can allow a reduction in the costs of high efficiency photovoltaic cells and their easy integration with "traditional" photovoltaic systems. The use of this approach, however, presents the need to have different materials with the same lattice parameters. By MOVPE growth technique it is possible to "engineering" these materials by varying the chemical composition step by step in order to fit the various lattice constants. SiGe (Silicon-Germanium) or, alternatively, SiGeSn (Silicon / Germanium / Tin), have the chemical-physical properties suitable for producing materials with variable compositions such as to have both the silicon and III / V materials (such as GaP, InGaP) lattice parameter. A preliminary study for the development of MOVPE processes to grow SiGe and SiGeSn films on silicon has been carried out using a modified Aixtron G4R Epitaxial Reactor. The structural properties of materials have been studied by X-Ray and TEM; Electrochemical profiling has been used to get doping profile and PL to perform optical characterization.

Resume : In this paper, we make a comparative study on the scaling capabilities of GeSn and SiGe nanoscale Double-Gate tunneling (DGT) FETs by using accurate numerical investigation. To do so, we evaluate the figures of merit (FoMs) of nanoscale devices by carefully investigating subthreshold swing factor (SS), ION/IOFF ratio and the threshold voltage as function of channel length values. Moreover, numerical analysis is carried out incorporating the impact of the Ge and Sn mole fractions on the devices performance, where the increase in Sn content of the GeSn-layer at the source/channel junction and the influence of its thickness and position improves the ION current and ambipolarity effects. Our findings reveal that GeSn-based DG TFET outperforms SiGe and Si devices in terms of scaling capability and immunity against short-channel-effects (SCEs). Moreover, the investigated GeSn DG TFET offers superior and optimal FoMs for an appropriate Sn content value as compared to that provided by SiGe and Si-based transistors. This makes the GeSn-based DG TFET a potential alternative for high-performance nanoelectronic applications.

Resume : Recently, in-situ-phosphorus-doped (ISPD) epitaxial Si has been drawing considerable interest due to their attractive electronic properties. Despite this interest, the fundamental properties of Si:P, such as the origin of high tensile strain and the mechanism of enhanced electrical conduction by laser annealing, remain poorly understood. In this paper, we have investigated the nature of non-equilibrium incorporation of phosphorus dopant into Si substitutional sites above the equilibrium concentrations and its electrical activation, using a combination of experiments and density functional theory (DFT) calculations. Highly phosphorus-doped epitaxial Si films were grown on Si (100) substrates by using a reduced pressure chemical vapor deposition (RPCVD) system. Raman scattering and X-ray diffraction (XRD) measurements were performed to investigate composition and strain states and Hall effect measurements were also conducted to explore the electrical properties. We have quantitatively and qualitatively identified substitutional incorporation of P into Si lattices using Raman and XRD analyses. Moreover, experimental results are compared with the theoretical values under various P incorporation conditions. Our results demonstrate that the tensile strain is not generated by another complexes but by the lattice mismatch between P and Si, thereby unveiling the nature of the P incorporation in Si:P thin films.

Resume : First results on self-diffusion (SD) in amorphous Si (a-Si) were published only recently [1,2]. The lack of data is due to the difficulty to suppress the random nucleation of the crystalline phase and the solid phase epitaxial recrystallization (SPER) of the a-Si layer commonly prepared on crystalline Si substrates. Here we report experiments on SD in a-Si utilizing isotopically controlled Si multilayers that were grown epitaxially on silicon-on-insulator wafers and subsequently amorphized by Si implantation. Diffusion was performed at temperatures between 460 and 600°C. The microstructure of the isotope layer was verified by transmission electron microscopy. Si isotope profiles were recorded with secondary ion mass spectrometry to extract the diffusional broadening. The temperature dependence determined for self-diffusion in a-Si is described by an Arrhenius equation with a single activation enthalpy Q=(2.67 ± 0.12) eV. Our experiments confirm results of Noah [1] but contradict the high activation enthalpy Q=4.4 eV reported by Strauß [2]. Whereas the value of Strauß suggests a cooperative mechanism of self-diffusion in a-Si, which involves several atoms, our value suggests a more local process. A similar low activation enthalpy is reported for SPER and foreign-atom diffusion in a-Si. This similarity points to a common atomic mechanism for all these processes. [1] M.A. Noah et al., J. Appl. Phys. 117 (2015) 165306. [2] F. Strauß et al., Phys. Rev. Lett. 116 (2016) 025901.

Resume : Lithium-ion batteries, which are dominating energy storage devices for consumer electrics and electric vehicles, appear high-capacity electrode materials urgently. Si is the most promising anode alternative for its super theoretical specific capacity (~4200mAhg-1). But the problem is the huge volume change of Si during charging and discharging, which leads to structural failure and fast capacity fading. Herein, we demonstrate the synthesis of a novel structure of porous Si in carbon cages via the reaction between Mg2Si and CO2 and subsequent acid washing. Benefiting from the in-situ CVD through magnesiothermic reduction of CO2, the deposited carbon cage seals the inner Si completely and shows higher graphitization compared with that from acetylene decomposition. After acid washing, pores are created, which can accommodate the volume change of Si. As the anode of lithium-ion battery, the carbon-caged porous Si electrode shows a capacity of 1124 mAhg-1 after 100 cycles with 86.4% capacity retention at 0.25C. When the current densities increase to 1C and 2C, the capacity can still be maintained at 860 and 460 mAhg-1 reversibly. The enhanced cycling and rate performance are contributed by the built-in void for Si expansion, static carbon cage keeping the outside solid electrolyte interface (SEI) stable, and fast charge transport by the porous structure. The results indicate that porous silicon in carbon cages is prospective to work as high-performance lithium-ion battery anode.

Resume : Electrically active defects in n-type 4H-SiC Schottky barrier diodes were studied by deep level transient spectroscopy (DLTS) and high-resolution Laplace-DLTS (L-DLTS). One of the main electron carrier traps in as-grown and irradiated n-type material gives rise to a broad DLTS band at around 290K, which is commonly referred to as Z1/2. The formation and behaviour of the Z1/2 defects are important for SiC radiation detectors as they are effective minority carrier lifetime killers. The Z1/2 peak is thought to be related to electron emission from the doubly negatively charged state of carbon vacancy (VC) at hexagonal (h) and cubic (k) lattice sites in 4H-SiC. It is generally accepted that VC possess negative-U ordered acceptor levels for both lattice sites. However, so far no hard evidence has been provided for a separation of the broad Z1/2 DLTS signal into components. Using L-DLTS we show that the Z1/2 peak has two components with activation energies for electron emission of 0.56 and 0.67 eV. It is argued that these components are related to Z1(=/-)+Z1(-/0) and Z2(=/-)+Z2(-/0) transitions associated with negative-U ordered acceptor levels of Vc at hexagonal (h) and cubic (k) lattice sites, respectively. With the use of L-DLTS we have also investigated processes of electron emission and capture for both acceptor levels of VC(h) and VC(k), determined activation energies for the charge state transitions and plotted the corresponding configuration-coordinate diagrams

Resume : Experimental studies [1] of Ge nanoislands on silicon-on-insulator (SOI) substrates indicate that the strain distribution in such heterostructures is very different from that observed during Ge growth on thick Si substrates. This is accompanied by a defect-free strain relaxation mechanism through the bending of the substrate. Using atomistic Monte Carlo simulations and analytical modeling [2], we have coupled this relaxation mechanism with interdiffusion and alloying and observe composition profiles that are completely different from those observed in flat nanoislands [3]. The effect of curvature can dramatically alter the composition state of alloyed heteroepitaxial systems. We observe that both the shape of the composition profile and the Ge content of the island are strongly affected.The presence of curvature can reduce alloying in the system,particularly when kinetic effects are important. The effect of this relaxation mechanism on the optoelectronic properties of heteroepitaxial systems is now under study. [1] F. Liu et al., Nature 416, 498 (2002); F. Cavallo and M. G. Lagally, Nanoscale Research Letters 7, 1 (2012). [2] T. Leontiou, J. Tersoff, and P. C. Kelires, Phys. Rev. Lett. 105, 236104 (2010). [3] T. Leontiou and P. C. Kelires, Phys. Rev. B 93, 125307 (2016).

Resume : Carrier lifetime measurement is an essential step of the quality control in silicon wafer production due to its peerless sensitivity to metallic contamination. and other electrically active defects. However, the adequate bulk carrier lifetime testing requires decent surface passivation process. Chemical passivation methods (e.g. using iodine-ethanol solution) are widely used both in R&D and in the industry providing excellent passivation quality, even down to 1 cm/s surface recombination velocity. Chemical passivation is a manual process, thus unwanted and non-systematic failures in the process are possible. These results in reduced lifetime patterns in the recorded carrier lifetime maps. Since low lifetime patterns may correspond to both bulk contamination or locally improper passivation, the reliable distinction of the two possible origins is important. Mostly, it requires competent human operation. Nowadays, artificial neural networks (NN) are widely used for many different applications, it is an especially efficient tool for image processing and pattern recognition. We have developed a human trained neural network based image processing solution, which is capable of the identification of the origin (passivation related or bulk contamination) of the low lifetime patterns and so the automatic classification of the samples. Based on the analysis of the several thousand labeled µ-PCD carrier lifetime maps, the statistical results of the automatic routine show – compared to human operators – equal or even better reliability, better reproducibility and - by its nature - no variation due to different human operators.

Resume : Two stage electron irradiation with thermal heat-treatments after each stage has been used for oxygen-related defect engineering in Czochralski-grown silicon (Cz-Si). First, the Cz-Si samples were irradiated at room temperature with 2 MeV electrons and then heat-treated at 320 ºC for 30 hours to anneal out the VO, V2O and V3O centers and generate the VO2, V2O2 and V3O2 complexes as dominant vacancy-oxygen related defects. Subsequently, the samples were irradiated at room temperature again and subjected to 30-min isochronal annealing up to 350 ºC. Defect evolution upon the treatments was monitored by means of the local vibrational mode (LVM) spectroscopy. The FT-IR spectra were recorded with the samples being at 20 K and 300 K. From an analysis of changes in intensity of the LVM lines it has been revealed that the second irradiation results in a noticeable decrease in the concentrations of VO2, V2O2 and V3O2 complexes and an essential growth in the concentrations of the oxygen dimer and the substitutional oxygen-interstitial oxygen pair, i.e. the so-called VO2* defect (metastable state of VO2). The observed defect transformations are argued to be related to an interaction of the radiation-induced self-interstitial silicon atoms (I) with the vacancy-oxygen complexes via the following reactions: VO2 + I ? O2i, V3O2 + I ? V2O2, V2O2 + I ? VO2*.

Resume : The most common substrate currently used to produce solar cells is boron doped p-type silicon. Boron doped silicon is susceptible to light induced degradation (LID), whereby boron-oxygen complexes form as recombination centres and reduce cell efficiencies. Several methods have been suggested to reduce or eliminate the effects of LID – one of them is the use of alternative p-type dopants. Indium doped silicon is a possible candidate and this study aims to ascertain its fundamental limitations. We have measured the effective carrier lifetime in indium doped samples passivated by silicon nitride and a room temperature superacid method. We find the recombination rate in the indium doped samples to depend systematically on the dopant concentration. The relatively deep acceptor level of indium (0.15 eV above the valence band) means there is a significant fraction of unionised indium present at typical operating temperatures, and we suggest that this is responsible for the recombination effect. Although this recombination is a fundamental feature of indium doped silicon, we show that, under certain circumstances, the lifetime for certain relevant doping levels can be higher than boron doped silicon after LID. Therefore, a window of opportunity may exist to use indium doped silicon in high efficiency p-type solar cells.

Resume : Silicon wafers for photovoltaics could in principle be produced without kerf loss by high temperature mechanical deformation of polycrystalline silicon feedstock, provided sufficient control of recombination-active defects can be achieved. In this paper we report a study of microstructural evolution in polycrystalline silicon using mainly high resolution electron backscatter diffraction (HR-EBSD). The Siemens polycrystalline silicon starting material is found to be heavily textured with a stem-tree-like structure with grain sizes typically 3 to 30 µm long in growth direction and 1 to 3 µm in diameter. The material contains many Sigma 3 and Sigma 9 grain boundaries. Annealing at 1400 ˚C results in 90% recrystallization and a reduction in average geometrically necessary dislocation (GND) density from >10^14 /m^2 to approximately 10^13 /m2. Subsequent compression at 1150 ˚C by 10% produces sub-grain boundaries seen as continuous curved high GND content linear features spanning grain interiors in HR-EBSD generated GND density maps. The GND density stored in these features is approximately 10^14 /m^2, whereas the regions between them have a GND density of 5 x 10^13 /m^2. Post-deformation annealing at 1400 ˚C facilitates a secondary recrystallization process, resulting in large grains typically of 100 µm diameter with relatively straight high angle grain boundaries.

Resume : We have reported on the Cu-induced poly-crystallization used Ge/Cu/Ti/substrate structure as a staring material. The Ti layer work as an adhesive layer, however this Ti layer is difficult to form patterns and prevent for making flexible integrated senor system. In this presentation we would like to report on the poly-crystallization technique to Ge film deposited on flexible substrate without using Ti layer. We have chosen Mo for new adhesive layer instead of Ti. This is because the Mo layer is easy to form the patter by a conventional photolithography method. Details of preparation of Ge film were same as previous report. The thickness of Mo film was 15nm. We had form the various size of pattern without evaporated-Mo layer on the flexible substrate. The poly-crystallization temperature was 150C for 10h in vacuum condition. The peak of Raman shift from Mo removed part was 299cm-1 same as that with Mo layer, but its FWHM value of 12cm-1 was broader than the value with Mo layer. From the curvature measurement, electro-deposition current dependence was not observed clearly, and the film thickness dependence observed when using Ti was not observed when Mo was used. Thus, it is presumed that the change in stress is small when Mo is used. It was often observed that Gee peeling off in Mo-etched-off region and confirmed that optimizing the size of area without Mo layer was very important. [1] Y. Uchida et. al., Phys. Status Solidi C, 11(2016)864.

Resume : In a conventional field effect transistors (FETs), the operating voltage is limited by the Boltzmann distribution in a transistor subthreshold swing of 60 mV per decade at room temperature. However, the ferroelectric thin films can be formed by Zr doping in HfO2, (HZO), a conventional gate dielectric material, to reduce the operating voltage by forming a negative capacitance. In order to understand the characteristics of ferroelectric (FE) and anti-ferroelectric (AFE) characteristics of HZO thin films, we measured and analyzed the polarization versus electric field (PE) loops. Furthermore, to understand both phenomena such as wake up and fatigue effect, we fabricated HZO thin film on actual Si/SiO2/TiN substrate and analyzed the domain configuration and local polarization switching behavior using Piezoresponse Force Microscopy (PFM). Our work is expected to significantly contribute to the design of HZO, which will be studied in memory such as negative capacitance, ultra low power logic transistor and ferroelectric random access memory (FRAM) devices.

Resume : Tin diselenide (SnSe2) is a 2D transition-metal dichalcogenide of great interest in making field-effect transistors and optoelectronic devices. We investigated the thermal stability of magnetron sputtered SnSe2 films. This problem is important to ensure the durability of the material. SnSe2 films were sputtered at 200oC. They were heated in Ar and O2 at atmospheric pressure and temperature Ta from 200oC to 650oC. We investigated the threshold temperature for each case above which the film started to dissociate or being oxidized. The structural change of the films was monitored using X-ray diffraction (XRD), Raman scattering and scanning electron microscopy (SEM). The photo-assisted resistive response to O2 were also measured to probe the change of metal content or oxidation. XRD and Raman data showed that when heated in Ar at Ta from 200oC to 350oC, the film was dominated by SnSe2. When Ta = 400oC, it dissociated to SnSe and was completely oxidized to become SnO2 quickly. Strong photo-assisted resistive response to O2 was detected. Contrarily, the film heated in O2 remained in the SnSe2 phase. Without going through the SnSe phase, it started to be oxidized at Ta > 450oC. Complete oxidation occurred at Ta = 550oC. Aggregation of all the films occurred at high Ta. In conclusion, a SnSe2 film is thermally stable up to 400oC in Ar or 450oC in O2. The latter case is suggested to involve surface oxide formation which slows down oxidation rate to result in a better thermal stability.

Resume : The transmission line model (TLM) and circular transmission line model (CTLM) are common methods for measuring contact resistivity because of their simplicity. Using finite-element simulation, the contact resistivity measurement accuracy of TLM and CTLM with highly doped silicon was studied. In a conducted simulation using TLM and CTLM, we defined the measuring voltage at which contact resistivity is extracted and changes occur in the activated doping level and measuring position. The contact resistivity extracted for highly doped silicon and high metal resistance is much more sensitive to the position at which voltage is measured than lightly doped silicon. To increase the measurement accuracy of contact resistivity, the multi-line transmission line model (ML-TLM) was used. With ML-TLM, the ratio of contact resistance to the total resistance could be calculated, along with the metal resistance.

Resume : Oxygen related complexes are widely accepted to have a direct impact on the performance of Si-based solar cells. For instance, the complex of a vacancy with two oxygen atoms (VO2) is believed to play a crucial role at the early stages of oxygen precipitation during Czochralski–growth. Light induced degradation is also related to an oxygen-related complex with boron (BO2), although the exact atomic configuration is still under debate. Under equilibrium conditions, the concentration of the defects in the electronic and solar- graded Si is below the detection limit of most of the spectroscopic techniques. High-energy particle irradiation is used to artificially increase the concentration of the defects. In this study, we have investigated thermal stability and electronic properties of the radiation induced defects in Cz-Si by deep level transient spectroscopy (DLTS). The irradiation dose used is considerably higher as compared to that typical for DLTS studies, in order to promote formation of higher order oxygen complexes. p+n– and n+p–diodes were irradiated by 2 MeV protons with doses from 1×10^14 cm-2 to 1×10^15 cm-2 at room temperature. The samples underwent a series of annealing from 350 to 550 °C. Several DLTS peaks are observed, not found after conventional low dose irradiation. Their thermal stability and the origin, which can be tentatively attributed to vacancy-oxygen (VmOn, m,n>1) and carbon-di-oxygen (CiO2), are discussed.

Resume : Nitride LEDs development presents significant scientific and societal issues. The aim is to get low-cost, high efficiency LEDs with accurate colour-rending (typically the Colour Rending Index has to be higher than 90). Type-II InGaN-ZnGeN2 quantum wells (QWs) were proposed for the improvement of efficiciency in active regions and realizeing then devices operating in a large wavelength range from UV to IR. Due to their large band gap (from 0.8 to 6.2 eV), III-N materials, as GaN and alloys, are still used for LEDs development. Nevertheless, they present several huge limitations mainly due to the evolution of InGaN properties for higher Indium concentrations. Strain and polarization effects affect then the LED quality throught the reduction of the spontaneous emission. New high-performance devices require the development of new materials and the introduction of ZnGeN2 layers could be an alternative solution. ZnGeN2 derives from the III-nitride elements by replacing the III-group alternatively by a group II (Zn) and a group IV (Ge). Both the energy band gap and the lattice parameters of ZnGeN2 are very close to those of GaN. The crystallographic organisations are similar and the recently predicted large band offset between GaN and ZnGeN2 allows the formation of a type-II InGaN-ZnGeN2 heterostructure. Inserting a ZnGeN2 layer in a conventional type-I InGaN QW structure yields significant modifications. The strong confinement of holes in the ZnGeN2 layer allows the use of a lower In-content InGaN QW with uniform In content. In the type-II InGaN-ZnGeN2 QW designs, a thin AlGaN layer was used as a barrier for better carrier confinement. The type-II InGaN-ZnGeN2 QWs lead to a significant enhancement of the electron-hole wave function overlap as compared to those of the conventional QWs. We describe here the results of simulation of the proposed type-II QWs. We focus on the designs of type-II InGaN-ZnGeN2 QW structures for green-emitting LED applications. Different design protocols are compared by changing the width and the position of the ZnGeN2 layer. The self-consistent 6-band k∙p method is used to perform the band structure calculations, which take into account the effect of strain, the valence band mixing, and the spontaneous and piezoelectric polarisations.

Resume : Diamond is a wide band gap semiconductor (5.45 eV) with some amazing properties such as a huge breakdown field (10 MV/cm) and a high mobility (2000 cm²/V.s). Those are some of the reasons that make it a good candidate for power electronics device in comparison with other wide band gap semiconductors materials like silicon carbide and gallium nitride, which are also under study to overcome the silicon devices limitations. Diamond Schottky diodes are the most promising device but there still some points to understand such as: where is the leakage current coming from. How is distributed the electric field along the material and what is the effect of barrier inhomogeneities at the interface. Furthermore, the architecture currently used need to be improve to reach the theoretical potential of diamond power device. For all of these reasons, we proposed to define by FIB (Focused Ion Beam) a thin lamella through a full stack, i.e. Schottky metal, diamond, and back ohmic contact and analyzed it by transmission electron microscopic (TEM). The TEM study give us info on crystalline quality and help us to find the good condition to grow good diamond quality. Then, by adding wire on both contact, we propose to observed field distribution by electron beam induce current (EBIC), which is normally impossible.

Resume : In this work, the influence of plasma process on passivation quality of lifetime sample (c-Si wafer passivated by Atomic Layer Deposited Aluminum Oxide [ALD-Al2O3]) has been studied in real time by in-situ photoluminescence tool (in-situ PL). PL intensity, indicating passivation properties of sample, behaves differently between the first and afterward plasma treatment. We observe that immediately after starting plasma the PL signal abruptly shots up, followed by a gradual degradation. The PL intensity of degraded sample can be simply brought back to its previous level by annealing above 150°C, otherwise the degradation remains stable. However, from the second plasma treatment onward, a plummet of PL signal was observed instead of the slow degradation in the first plasma exposure. On top of that, the quick increase of signal just after plasma ignition is not found either. This phenomenon is even more obvious when the plasma treatment is performed at higher temperature. A model, explaining structure modification of passivation layer due to plasma process, is proposed based on plasma physics and degradation mechanism of Al2O3/c-Si interface. The model is discussed in light of Fourier Transform Infrared spectroscopy (FTIR) results on the samples before and after plasma treatment. This insight into influence of plasma on passivation layer can possibly assist the definition of a suitable plasma process to reduce the degradation e.g. during deposition of capping layer.

Resume : We have theoretically investigated the 1.55 µm p-i-n GaNAsBi-based multiple quantum wells (MQWs) using a self-consistent calculation combined with 16x16 BAC model. Their performances are evaluated in terms of optical gain and radiative current density J_rad. We have found that J_rad reduces by increasing the well thickness〖 L〗_ω. The quantum confined Stark effect as well as the doping effect on spontaneous emission and radiative current density in ideal lasers are also discussed. The optical properties of the heterostructure are improved when the MQWs are doped. The optimization of well parameters can be used as a basis for GaNAsBi-based lasers intended for optical fiber telecommunication wavelength.

Resume : Growth of III-V compounds on silicon attracts lots of attention primarily due to silicon wafers cost and a fact that it is the second most abundant element in the earth's crust. Nowadays there are two main techniques for III-V on Si growth – MBE and MOCVD. In both cases it is necessary to heat silicon up to 700-800 °C prior to the deposition for oxide removing and surface reconstruction, which could affect the properties of silicon wafers [1,2]. For semiconductor devices such solar cells the quality of silicon surface and its bulk properties in general define the solar cells characteristics. Plasma enhanced atomic layer deposition (PE-ALD) approach for III-V growth at temperatures not exceeding 400 °C [3] is a promising method for nucleation layers growth and potentially could increase III-V/ Si solar cell’s efficiency. To estimate carrier’s lifetime in silicon following methods could be used: transient photoconductance decay, photoluminescence (PL) imaging and PL decay measurements. The last one is a direct method of carrier’s lifetime estimation. This paper presents the last results on 2D measurements of carrier’s lifetime in silicon thermally treated in conditions related to different growth methods. [1] E. García-Tabarés et al. // 42nd IEEE PVSC, New Orleans, 2015, pp. 1-6. [2] R. Varacheet et al. // Energy Procedia 77 (2015) P.493-499 [3] A.I. Baranov et al. // Phys. Status Solidi A 214 (2017) 1700685

Resume : Small bandgap and high mobility charge carriers have been recently making germanium more and more interesting in several application fields for photovoltaics, radiation detection and nano-electronics. For this last application, nanoscale doping is a challenging task and monolayer doping is one of the promising route to face it especially tanks to its conformal behavior. In this work, we report about the formation of a Sb monolayer (ML) on Ge surface by means of thermal evaporation from a Sb layer sputtered on Si, used as a remote source. Thanks to RBS, we have studied the phenomenon with different thermal budget in order to characterize the evaporation and the deposition of Sb. We discovered the thermal windows in which the ML is formed on Ge: only above 600°C the Ge surface is covered with a Sb ML and it’s stable up to 780°C, while at lower temperature a thicker Sb/Ge alloy layer forms. By means of rapid thermal annealing (RTA) deposition, we demonstrate that the process is very fast occurring in the first 10s of the heating. This is a crucial result since it means that monolayer may be formed before that diffusion process takes place into the material. The use of Sb monolayer as a doping source is investigated under RTA, flash lamp and laser annealings. SIMS Sb chemical profiling have demonstrated that Sb ML acts as an effective diffusion source. These results make Sb monolayer evaporation a concrete perspective for future nanoscale controlled doping.

Resume : The technologies of formation of silicon nanowires and combined silicon pyramid structures - silicon nanowires by the method of chemical etching are researched and developed. It is shown that the method of chemical metal-catalytic etching allows to obtain structures with different morphology and distribution on the substrate surface. The obtained arrays of filamentous silicon nanocrystals with diameters of 60-250 nm and a height of 2-10 microns. The optical and temperature characteristics of such structures are studied. The prototypes of photoelectric converters based on structures with silicon nanowires are formed and their characteristics are studied.

Resume : A new set of interatomic pair potentials is reported for indium sesquioxide (In2O3) and tin dioxide (SnO2) which offer an improvement over the previously available models in terms of the lattice and defect properties. Based on the newly developed potentials, atomistic simulations including intrinsic and extrinsic defects, cluster defect formation, electron-hole formation, and defect formation energy have been studied. Our calculation shows that both anion and cation interstitials of In2O3 prefer 16c site. The calculated results of the cluster binding energies show that the tin substitution prefers d-site indium when only taken account of the first shell cations around the interstitial oxygen. After taking into account the next-nearest-neighbor shell, we find a strong preference for the tin substitute in the d-site indium at the first shells. Using hybrid quantum mechanical/molecular mechanical (QM/MM) embedded cluster calculations, we investigated the formation and ionization energies of native point defect including vacancies and interstitials in In2O3, SnO2, and the low energy configurations from the solid solutions of In2O3 and SnO2 simulation using our new potential.

Resume : In the present work, we report on epitaxial growth of ferromagnetic Mn5Ge3 thin films on (001) Ge substrates induced by Mn in-diffusion during non-equilibrium flash lamp annealing for 20 ms. The ferromagnetic Mn5Ge3/Ge samples with very sharp interface between the Mn5Ge3 layer and the Ge substrate can be used to fabricate spintronic devices. Temperature-dependent magnetization reveals a Curie temperature of 282 K which can be tuned much above room temperature by strain engineering and/or co-doping with C. The microstructural properties of the fabricated films were investigated by X-ray diffraction, cross-sectional TEM and Rutherford backscattering spectrometry. Both used material and technology are highly compatible with complementary metal-oxide-semiconductor (CMOS) technology and can be used for spintronics.

Resume : One of the perspective ways for further developments of photovoltaic is related to fabrication of multijunction solar cells grown on Si wafer. For lattice matched growth GaP is the best candidate because it has the smallest lattice mismatch with Si (less 0.4 %) along all binary III-V alloys. However GaP has an indirect band gap of 2.26 eV being a promising material for barrier layers but it is unsuitable for active layers of top junction for multijunction solar cells based on Si. A lattice-matched GaP/Si superlattice could be used as an active material for top junction grown on Si based bottom junction. According to theoretical estimation reduction of Si quantum well width to 1-2 nm in GaP/Si superlattice leads to shift of the band to 1.6 eV. Study of structural (SEM, TEM, AFM), optical (UV-vis-IR and Raman spectroscopy) and electrical properties (I-V and space charge capacitance measurements) of GaP/Si superlattice grown on Si and GaP substrates will be presented in the paper. The layers of GaP were grown using plasma enhanced atomic layer deposition approach, while Si layers were grown using conventional plasma enhanced chemical vapor deposition with high hydrogen dilution. The whole deposition process was performed at temperatures not exceeding 400 °C to minimize negative influence to Si wafer. An effect of the deposition conditions to the morphology, structural, optical and electrical properties will be discussed.

Resume : All-vanadium redox flow batteries (VRFB) are one of the promising electrochemical energy conversion and storage technology of modern era. It comprises of four vanadium species i.e. V(IV)/V(V) and V(III)/V(II) as cathode and anode redox couples, respectively. Carbon felt is most widely used electrode material for VRFB. However, the kinetics of vanadium redox reactions at the anode and cathode are slow which, therefore, requires enhancement for the operation of VRFB at high current densities. As liquid acidic vanadium-based electrolytes stored in tanks flow through the cell stack (comprised mainly of electrodes and membrane), implementation of active electrocataltsyts through introduction in the electrolyte and hence in situ deposition at the electrodes can be a convenient and commercially feasible methodology for the performance improvement of VRFB. Therefore, tin in form of two tin species (Sn(II) and Sn(IV)) were introduced in the electrolyte and their impact on kinetics of vanadium redox reactions is evaluated. As the deposition potentials of Sn(II)/Sn is -0.13 vs. SHE, therefore, the operating conditions in VRFB are suitable for in situ electrodeposition of tin. Through the coupling of redox activities of tin and vanadium, it is found a highly effective catalyst for VRFB. It remarkably increased the energy efficiency, specific discharge capacity and electrolyte utilization by lowering overpotentials and improving redox kinetics of vanadium couples, in particular at anode.

Resume : We have studied the time-resolved luminescence of Si-nanocrystals in SiO2 matrix and the influence of doping by phosphorus and boron on the luminescence properties of the samples with ps time-resolution. The NCs were fabricated by two different processes, resulting in NCs with the main luminescence band at 1.3 eV or at 1.42 eV with a lifetime of tens of us. The phosphorus doping behaves as a passivating agent in contrast to the boron doping, which shortens the radiative decay times and seemingly changes the energy structure of the nanocrystals. We discovered a new luminescence band centered at 1.9 eV visible only for high excitation intensity and with decay times shorter than the main luminescence, but still in the us range. We connect its origin to the NBOHC-defect (non-bridging oxygen hole center) in SiO2. We have found out that this defect luminescence band is heavily influenced by doping and passivation. The H2 passivation and phosphorus doping quenches this band but boron doping enhances it. At the same time the luminescence of the main band decreases at excitation intensities higher than 8 mJ/cm2. By comparing the luminescence of the NBOHC defect in the NC samples and a pure SiO2 we have discovered that the defect luminescence in the nanocrystal samples is stronger suggesting a transfer of energy between the SiO2 defect and the nanocrystals.

Resume : Due to the quantum confinement effect, the band gap energy tunability of silicon nanocrystals (Si-NCs) is directly related to the size of these nanostructures, which has been largely studied for its implementation in photovoltaics and light-emitting devices. Chemical deposition methods have typically been employed for matrix-embedded Si-NCs fabrication, being the utilization of toxic gases a serious handicap aiming at mass production. In contrast, physical deposition methods such as electron beam evaporation (EBE) encourage the direct deposition of Si-NCs without chemical reactions being involved in the process. In this work, Si-NC superlayers have been fabricated by means of EBE, by depositing in an oxygen-rich atmosphere alternative layers of SiO2 and Si on SiO2 substrates. The stoichiometry of the layers was determined by X-ray spectroscopy, which confirmed the controlled silicon oxidation in order to attain silicon-rich oxide (SRO) layers. A post-deposition annealing process was carried out at different temperatures in order to achieve the Si precipitation in the form of nanocrystals. Transmission electron microscopy and Raman-scattering measurements confirmed the presence of crystalline Si-nanoprecipitates. Photoluminescence (PL) spectra from the Si-NCs samples could be deconvolved into two contributions, whose dynamics suggests that two different luminescent centres are responsible for the optical emission of the samples. The present PL results will be discussed considering the possible origin of the luminescent centres, published data in similar EBE-grown SiNCs and also in high-quality samples obtained with chemical deposition methods.

Resume : SiC is a wide bandgap indirect semiconductor, whose rigidity and biocompatibility suggest its use in the form of nanoparticles for applications in bio-imaging and bio-medicine. The SiC nanoparticles commonly show a broad blue-green photoluminescence (PL) under excitation by UV or violet lasers. In this work we study such emission from single SiC nanoparticles prepared by etching and filtering of micron-sized SiC particles [1]. The resulting colloidal dispersions (without any stabilization by surfactants) are size-separated using the gradient centrifugation. For the micro-spectroscopy experiments, the fractions of suspension are deposited on a cleaned Si substrate. In general smaller and bigger nanoparticles show luminescence peaking around 450 and 530 nm, respectively, revealing existence of two main types of luminescence centres. Both of them are broad, even at low temperatures indicating the presence of strong electron-phonon coupling (large Huang-Rhys parameter). The cryogenic micro-spectroscopy experiments reveal vibronic structure in the spectra which will be compared with theoretical models of surface defects. (The work was performed within the Visegrad group + Japan project 8F15001, co-financed by the International Visegrad Fund). [1] D. Beke et al., Nanoscale 7 (2015) 10982.

Resume : In 1998, we demonstrated that a grown-in defect free CZ-Si crystal can be realized over the entire length of the 150mm diameter crystal by v/G control, i.e., the ratio of growth rate (V) to the axial temperature gradient (G) in the crystal near it’s melting point. In this method, using a hot-zone with a uniform G distribution in the radial direction, v/G is maintained around a critical value at which the amount of vacancies is balanced with that of Si interstitials during crystal growth, so that the generation of grown-in defects is suppressed. After that, this concept of v/G has been applied to mass production of 200mm and 300mm diameter CZ-Si crystals and nowadays they have become the standard crystals for various devices. On the other hand, FZ-Si crystals had been used mainly for power devices such as IGBTs. In 2008, we developed 200mm diameter MCZ-Si crystals with very low oxygen concentration as an alternative material, because of limitation of FZ-Si wafer supplier and difficulty of the larger crystal development. However, MCZ-Si crystals contain grown-in voids, because simultaneous control of very low oxygen level and v/G was difficult. In order to develop lower cost and higher quality MCZ-Si wafers for future Si power devices, as-grown crystals free of oxygen precipitates as well as voids must be achieved. In this presentation, we will introduce nitrogen- and hydrogen-doping technologies to solve the problem and discuss the advantages and disadvantages of each method.

Resume : Crystalline silicon has been the workhorse of the photovoltaics industry since the early days. Earth-abundant, stable and well-known, with high minority-carrier lifetimes, it has enabled fabricating solar cells with ever-increasing power conversion efficiency. Today, the key driver of the technology is to reduce the cost per watt-peak and per area. Layer-transfer processes hold the promise of a breakthrough for such cost reduction, by fabricating 2 to 4 times thinner wafers, without kerf losses, while short-cutting energy-intensive steps of the wafer production value chain. The most prominent of these processes today, that is making its way to industrial production, are epitaxial foils from porous-silicon based lift-off. These foils are grown by epitaxy on a porous wafer surface, which enables both a monocrystalline growth and the controlled detachment of the foil. To compete with the current wafer-based industry, a few challenges are still to be met: to catch-up the high quality of the present Cz material with a (almost) perfect layer-transfer yield, and to manage processing thinner wafers into high-efficiency solar cells. This contribution focuses on the first challenge, of reaching high, stable and uniform minority-carrier lifetimes. Epitaxial foils show indeed a wide range of lifetimes and the factors influencing it are diverse, and seemingly not yet all identified. In this presentation, you will join our investigation into the main epitaxial foil lifetime killers.

Resume : High power semiconductor devices require silicon wafers with good radial resistivity uniformity. Float Zone (FZ) silicon wafers from neutron transmutation doped (NTD) ingots are used for achieving the lowest radial resistivity variation (RRV) compared to gas doped ingots. A radial neutron flux gradient is present for all NTD reactors due to neutron attenuation in the rotating ingot. The result is higher resistivity in the center than at the rim under assumption of insignificant initial doping concentration. For conventional 8’’ FZ NTD wafers, a RRV value of 5% can be achieved. The radial resistivity profiles for gas doped wafers have many different shapes and depend on the FZ growth condition. Growth parameters influencing the radial resistivity variation are among others: magnetic fields, dopant gas flow, crystal rotation speed, alternating rotation direction, induction coil design, eccentricity of crystal, RF frequency and crystal pull speed. By tuning the FZ growth parameters the shape of the radial resistivity profile is designed to have higher resistivity at the rim than the center. Meaning that the gas doped profile is the inverse of the NTD profile. An almost flat radial resistivity profile is achieved by tuning the gas doping concentration and the NTD irradiation dose. Using the approach of two profiles cancelling out each other, wafers from NTD ingots in the resistivity range from 50-800 ohmcm have been produced with RRV values less than 2% for all wafers.

Resume : Flash lamp annealing (FLA) is an innovative annealing method already used in semiconductor industry, for flexible electronics and for thin, functional coatings on glass. Due to the short time scale of milliseconds, FLA is cost and time effective, suitable for temperature-sensible substrates and allows the exploitation of non-equilibrium crystallization processes. In this contribution we present a new approach in which magnetron sputtering is combined with FLA. In detail, thin polycrystalline Si films have been fabricated and characterized with respect to their structural, optical and electrical properties. Special focus is set on the non-equilibrium crystallization process within the millisecond time scale. Furthermore, strategies to avoid thermal stress, to minimize defects and to obtain layers with a low electrical resistivity are discussed.

Resume : In this paper a review of the different 2D and 3D defects observed in the 3C-SiC hetero-epitaxy on silicon substrate is reported together with the different growth strategies for defects reduction. In particular we will discuss the origin and the mechanism of defects reduction of the voids at the 3C-SiC/Si interface, micro-twins, pyramidal twins and stacking faults. Both thin layers of few nanometers and thick layers of more than 100 microns were analysed. Furthermore the effects of compliance substrates, both on strain reduction and defects reduction is reported in detail.

Resume : Wide-ranging control of carrier lifetimes in epilayers is a key technique for 4H-SiC bipolar devices, such as PiN diodes, BJTs and IGBTs. Relatively long carrier lifetimes in E0-E1 μs range and short lifetimes less than 100 ns may be requested for drift and buffer epilayers, respectively [1]. In this paper, we surveyed methods to control carrier lifetimes by intentional impurity doping in 4H-SiC epitaxial growth. Control of V doping concentrations from mid E12 to mid E15 cm^-3 was achieved by changing the VCl4 flow rates and C/Si ratios of the source gases in 4H-SiC epitaxial growth. The lightly N V-doped epilayer (N=1E15, V=7E12 cm^-3) shows carrier lifetimes of 0.9-4 μs at RT-250 ºC. Conversely, the highly N V-doped epilayer (N=5E18, V=7E14 cm^-3) shows short carrier lifetimes of ~20 ns or less at RT-250 ºC. Inserting a thin N V-doped buffer epilayer between a substrate and n- drift epilayer in the PiN diodes was confirmed to be effective to prevent the formation of stacking faults from basal plane dislocations in the substrate. We will also discuss the effects of other impurities on limiting carrier lifetimes in 4H-SiC epilayers. This work was supported by Council for Science, Technology and Innovation (CSTI), Cross-ministerial Strategic Innovation Promotion Program (SIP), “Next-generation power electronics/Consistent R&D of next-generation SiC power electronics” (funding agency: NEDO). [1] T. Tawara et al., J. Appl. Phys. 120, 115101 (2016).

Resume : It has been recently attracted that formation of single photon sources (hereafter ‘surface SPSs’) that emit brightly single photons at room temperature was confirmed in SiC semiconductor material by oxygen annealing process. In SiC semiconductor industry, since high-quality and large diameter wafers are in mass-production and various device processes are matured, it can be said that SiC is very preferable as a SPS material for practical use. However, there are some issues to be solved, such as dispersive emission wavelengths between 600-800 nm and a broad emission spectrum from the surface SPS. According to ab-initio studies, it is suggested that the wavelength variation is due to the different distances between surface SPS and stacking fault because the stacking fault affects the structure and/or charge state of the surface SPS. On the other hand, we found that “line-shaped faults” were generated during oxidation of 4H-SiC epilayer and double Shockley stacking faults (2SSFs) were formed and extended by UV laser irradiation from the line-shaped fault. Accordingly, the emission characteristics for various surface SPSs were observed and they were compared to the surface SPSs at 2SSF free region. As a result, although the emission wavelength and peak width were comparable between near and without 2SSF, it is found that in terms of radiation intensity the surface SPSs tend to have stronger radiation.

Resume : Among the materials of the spintronic, great hopes are placed on the diluted magnetic semiconductors (DMS) [1]. Making a room-temperature diluted magnetic semiconductor has been the object of an intense research activity in the past 15 years, with mostly disappointing results. Here, we show evidence for room-temperature magnetism in Fe-implanted 6H-SiC samples annealed at different temperatures. We successfully disentangle the magnetism coming from the presence of Fe rich clusters to that of the fully diluted matrix thanks to thorough analyses of the nanoscale distribution of Fe, Si and C atoms by Atom Probe Tomography. We made the 3D reconstructions of as-implanted and 900°C annealed sample where only the Fe atoms are represented. In the sample some clusters clearly appear in the annealed sample. This reveals the presence of Fe rich nano-precipitates, with a disparity in their size ranging from 1 to 4 nm (the largest). Knowing the atomic composition of each cluster and using the SQUID magnetometry, we could calculate the mean value of the magnetic moment of each Fe ion located inside the cluster phases and diluted in the matrix. We provide unambiguous evidence for the presence of magnetism all the way to room-temperature in the diluted matrix of several samples. This result reinforces the assumption of intrinsic interactions between diluted Fe atoms, inducing ferromagnetism in Fe-implanted 6H-SiC. [1] T. Dietl, H. Ohno, F. Matsuka, J. Cibert, D. Ferrand, Science 287 1019 (2000).

Resume : The heteroepitaxial growth of 3C-SiC on silicon (100) is a well-known but difficult topic which is still studied by many research groups. In the other hand, the reports on the reverse system, i.e. Si heteroepitaxy on 3C-SiC, are much scarcer despite some potential applications. This is due to the fact that the Si layers grown on 3C-SiC(100) are almost always either polycrystalline or columnar with [110] oriented grains. In the present study, a robust and reproducible CVD process is developed in order to nucleate and grow (100) fully-oriented Si layers on top of 3C-SiC(100). It implies a post-growth modification of the 3C-SiC surface by pulse insertion of precursors during cooling, that led to a change in Si nucleation, favoring (100) Si islands instead of elongated (110) ones. Then, the standard CVD process is used to form a second 3C-SiC(100) layer on top, leading thus for the first time to a fully (100) oriented Si/SiC/Si/SiC stack. TEM observations confirm that (100) unique orientation of the stack, despite the generation of a high density of crystalline defects at each interface due to the 20% lattice mismatch. These defects densities significantly decrease along with layer thickness for both Si and SiC materials. APB defects present in the 3C-SiC layer do not propagate inside the Si(100) layer due to the four-fold symmetry of this material.

Resume : The atomistic kinetics allowing the growth of group IV crystals (e.g. epitaxial graphene or SiC) is barely known and this lack of fundamental knowledge often results in a difficult process and material quality control in term of defects and surface morphology. As a consequence, the development of the growth process processes for the production of group IV crystals is often based on time consuming and expensive design of experiments campaigns. We have developed stochastic simulations to study at atomic level the growth kinetics of materials characterized by the sp2 and sp3 bonding symmetries. The model can be coupled to the continuum simulation of the gas phase status generated in the equipments to simulate a variety of growth techniques (e.g. Chemical and Physical Vapour deposition, sublimation). Kinetics proceeds by nucleation and growth of ideal or defective structures and their balance depends critically on the process related parameters. Quantitative predictions of the process evolution can be obtained and readily compared with structural characterization of processed samples. In particular we can describe the surface state of the crystals and defects' generation/evolution (point and extended defects, as stacking faults) as a function of the initial substrate conditions and the process's parameters (e.g. temperature, pressure, gas flows). We discuss simulation aided real process development considering similarities and differences different synthesis procedures.

Resume : Silicon carbide (SiC) is a chemically inert wide band gap semiconductor, and promising new material for bioimaging, targeted drug delivery, nanosensing, optoelectronic and for heterogeneous photocatalysis as well, especially in nano size. Biocompatibilities of bulk SiC and SiC nanoparticles (NPs) have been proven by several research teams, and the aqueous solutions of SiC nanocrystals are exceedingly promising candidates to realize bioinert nonperturbative fluorescent nanoparticles for in vivo bioimaging. Thus, size control and the identification of luminescent centers in SiC NPs are of immediate interest. Despite its semiconducting nature, quantum confinement on SiC nanoparticles has not yet been proven because of the lack of size-selective synthesis methods below 10 nm. We studied the stain etching process used for SiC NPs synthesis and constructed a new description for electroless wet chemical etching of semiconductors. We claim that chemically generated excitons play significant role in the pore formation. By understanding the chemistry of electroless wet chemical etching, it is possible to control the size of the particles composing the pore wall, in order to produce SiC NPs in different size distributions below 10 nm.

Resume : The Si terminated nH-SiC(0001) (n=4 or 6) surface has been widely studied since 1989, and various reconstructions after Si exposure have been observed. The most common ones are the (√3×√3)-R30° and the (3×3) with Si adatom coverage of 1/3 and 13/9 monolayer (ML) over the Si atomic plane of the nH-SiC(0001) surface. We propose a new procedure for nh-SiC(0001) surface preparation based on a progressive enrichment in Si. Doing so, we identify two new reconstructions, namely the giant (12×12) stable below 650°C and the (4×8) stable until 830°C , which are intermediate between the (√3×√3)-R30° and the (3×3). From the positions of the Si adatoms determined by STM, atomic models are built introducing a new type of Si adatom that bridges two Si surface atoms of the nH-SiC(0001) surface. We also build two new atomic models of the previously observed (2√3×2√3)-R30° and (2√3×2√13) reconstructions. The stability of these models is tested by molecular dynamics and the results are consistent with the experimental observations. The reconstruction sequence with increasing Si coverage is (√3×√3)-R30°, (12×12), (2√3×2√3)-R30°, (4×8), (2√3×2√13) and (3×3). The four middle reconstructions have a Si coverage from 1 to 1.179 ML, which renders them very sensitive to the methods of the SiC surface preparation. This study rationalizes the preparation of nH-SiC(0001) surface, which is of crucial importance to improve its use for elaboration of opto-electronic and high-power electronic devices.

Resume : Silicon nanowires (NWs) are attracting the interest of the scientific community as building blocks for a wide range of future nanoscaled devices. We demonstrated the realization of a 2D random fractal array of vertically aligned Si NWs by using metal assisted chemical etching without any lithographic process and by a cheap, fast and maskless approach compatible with Si technology. We were able to control and tune the optical properties of the system by realizing different fractal geometry through the optimization of NW spatial arrangement [1]. In-plane multiple scattering and efficient light trapping related to the fractal structure were observed [2]. NWs achieved by this technique exhibited a very bright room temperature PL and EL, tunable with NW size in agreement with the occurrence of quantum confinement effect. An innovative Si NW-based optical biosensor is realized, which exploits the PL properties for the detection of proteins in a wide range of concentrations, down to the femtomolar limit, demonstrating the great potentiality of this material for biosensing [3]. 1. Light: Science & Applications 5 (4), e16062, 2016 2. Nature Photonics 11,170-176, 2017 3. ACS Photonics 10.1021/acsphotonics.7b00983, 2017

Resume : Self-diffusion experiments in silicon vertical nanowires with diameters of 70 and 400 nm annealed at 850 and 1000 °C are reported. The nanowires are structured epilayers of isotopically controlled silicon. These structures were first epitaxially grown on top of silicon substrate wafers. Nanowires were then fabricated using a nanosphere lithography process in combination with inductively coupled plasma dry reactive ion etching (ICP-DRIE) at cryogenic temperature. Three-dimensional profiling of the nanosized structure before and after diffusion annealing was performed by means of atom probe tomography (APT). Self-diffusion profiles obtained from APT analyses are accurately described by the solution of Fick’s law for self-diffusion across buried interfaces. The data obtained for silicon self-diffusion in nanowires are equal to results reported in the literature for bulk silicon crystals. This demonstrates that finite size effects and high surface-to-volume ratios do not significantly affect silicon self-diffusion. Furthermore, it shows that the properties of native point defects determined from self-diffusion in bulk crystals also hold for vertical nanoscale silicon structures with diameters down to 70 nm.

Resume : The solar energy that irradiates to Earth in a single day is more than the total energy consumption of the all world in a year. Therefore, it is particularly important to collect sunlight effectively and convert it into heat, electricity and light. In the present work, we proposed a 3D Copper-silicon nanowire structures for solar light-to-heat conversion device. The bottom-up synthetic strategy was used to fabricate a cross-interconnected multistage nanowire (NWs) structure into a three-dimensional porous membrane. Firstly, a cross-interconnected amorphous silicon (a-Si)/CuO core-shell NWs structure was formed by depositing the a-Si film around the CuO NWs. Afterwards, crystalline silicon NW branches were grafted on the core-shell NWs at the bottom while Cu3Si nanocrystals (NCs) were embedded in core-shell NWs at the upper layer. We observed the cross-interconnected multistage NWs structure by SEM and further confirmed the core-shell structure for the NWs and Cu3Si NCs by TEM. The significantly suppressed light reflection was achieved within the cross-connected multistage NWs structure, which made the captured light was absorbed fully at any angle on the surface of the material. Thereby, the formed absorber can efficiently realize wide-spectral absorption (>93%). This novel 3D absorber structure could provide a portable solution in water desalination and solar thermal photovoltaic field. This work is partly supported by NSFC (61735008).

Resume : Extensive theoretical studies demonstrated that kinked silicon nanowires (k-SiNWs) display superior mechanical, electrical and thermal properties compared to their straight counterparts. Yet, on the experimental front, a simple, accessible and controlled fabrication method to explore and exploit these theoretical predictions is not readily available. The interplay between the nanoscale properties of such k-SiNWs and their target applications requires precise tuning such as the number of kinks, their location, the length of the kink arm as well as crystallographic orientation. This work relies on the simplicity of the metal-assisted chemical etching of Si for the controlled fabrication of k-SiNWs. Our approach uses repetitive etch-quench sequences to form the kinks. The number of kinks is given by the number of etch-quench sequences and the etching times control the lengths of the arms of the kink. We further emphasize on the multiple etching opportunities with a broad kink angle spectrum dictated by the physical parameters of the Au catalyst masks. In particular, we investigate the effect of 10, 15 and 30 nm Au mask thicknesses associated with opening diameters of 100, 120 and 150 nm. Thin catalysts are associated with porous k-SiNWs and sharp kink angles, while employing thick catalysts typically results in solid k-SiNWs. The opening diameters are found to greatly influence the etching rates. A peculiar effect of the Au catalyst is to alter the mass transport of the chemical species to an extent that scales with the catalyst thickness. These observations are discussed within a theoretical framework. The elaborated experimental and theoretical framework provide an accessible handle for fabrication of application-tailored k-SiNWs.

Resume : In the last decade, silicon has been intensively investigated in the lithium ion batteries field because it provides the higher theoretical specific capacity known. However, their extraordinary properties come accompanied by a rapid material degradation, which results in a short lifetime of the batteries. To overcome these issues, numerous researches had focused on material nano-structuration or the use of amorphous silicon compounds. In the present work, we aim to merge both approaches by finding an amorphous silicon nanostructured material for its use in the Li-ion batteries. The nano-structuration process selected was the Metal Assisted Chemical Etching (MACE). This versatile technique is one of the most extended methods for the manufacturing of highly oriented silicon nanowires (SiNWs). Its controllability, scalability and high amount of nanomaterial obtained, make it stand as one of the most promising methods for industry scale SiNWs production. However, this method usually requires crystalline Si substrates (c-Si). Moreover, once the MACE procedure is completed, the SiNWs grown have to be detached from the c-Si substrate and then transferred to the batteries anodes. To simplify the process and obtain non-crystalline nanowires, the SiNWs were directly produced on hydrogenated amorphous silicon (a-Si:H), deposited on top of copper foils. The as resulting a-SiNWs/Cu electrodes by this one-step procedure, were used on Li-ion batteries. Their analysis shown an outstanding capacity and better overall performance than the starting a-Si:H material. These novel electrodes were further improved using several coatings, extending the lifetime of the material and the capacity retention.

Resume : Germanium recently attracted a renewed interest in various fields of material science such as photonics, opto- and nano-electronics, owing to its high carrier mobility as well as to its compatibility with silicon technology. However, Ge-based devices often requires well-defined doping profiles and very high activation levels (>1020cm-3), which are challenging for most of dopants due to their high diffusivity and low solubility . For this purpose, ion implantation followed by pulsed laser melting (PLM) is the most promising technique as it induces ultra-fast liquid phase epitaxial regrowth that allows confinement of diffusion while enhancing dopant incorporation. Latest studies on p- (by means of B or Al) and n-type (P, As) doping by PLM following ion-implantation into bulk Ge or Ge-on-Si epilayers will be presented. Thanks to advanced chemical (1D and 3D), electrical and structural characterizations with nanometer resolution, fundamental mechanisms such as non-equilibrium segregation, diffusion, clustering, strain and defects evolution will be discussed with special care on strategies for improving the electrical activation, together with issues about contaminations and thermal stability.

Resume : The fabrication of highly doped and high quality Ge layers is currently a challenging and hot topic in nanoelectronics, photonics and radiation detectors. We explore for the first time n-type doping of Ge by sputter deposition of a thin layer of pure Sb on the Ge surface followed by pulsed laser melting (PLM) diffusion annealing cycles. A broad characterization has been performed, based on secondary ion mass spectrometry, channeling-Rutherford backscattering spectrometry, Van der Pauw-Hall, high resolution x-ray diffraction, infrared reflectivity, scanning electron microscopy. We show that PLM promotes an efficient diffusion of high Sb concentrations in the Ge melted subsurface layer, followed by an ultra-fast epitaxial regrowth. As a result, excellent surface morphology and crystalline quality is obtained, with record active Sb concentrations well above 1020cm-3 over 100-150 nm. Key properties such as substitutional fraction, electron mobility, residual strain, infrared reflectivity and plasma frequency are also characterized and discussed. These results demonstrate Sb deposition followed by PLM as a simpler and cheaper doping method, and with lower thermal budget, than the other methods commonly employed such as ion implantation or in-situ doping of Ge epilayers, and at the same time able to achieve record activation levels with no residual damage and excellent electrical and optical properties, relevant for Ge based future advanced devices.

Resume : Group IV substrates are widely used for photovoltaic applications. Multijunction GaInP/GaAs/Ge solar cells are of the great interest due to their supreme efficiency that has already exceeded 40%. Further development of these structures is mainly focused on III-V layers while bottom sub-cell based on the p-n junction in the substrate is disregarded. However, Ge sub-cell contributing up to 10% of total solar cell efficiency is not studied enough. We have previously shown that diffusion of group III atoms during epitaxial growth of III-V layers leads to the formation of undesirable potential barrier for charge carriers at the III-V/IV heterointerface. Thus, charge carrier lifetime which depends on impurity level can be also affected by diffusion processes. In this paper the influence of III-V epitaxy on the minority carrier lifetime in Ge substrates is discussed. III-V/Ge heterostructures that reproduce bottom sub-cell of MJ solar cell both before and after thermal annealing as well as the structures with AlAs diffusion barrier are analyzed. The results on minority carrier lifetime obtained by different capacitance techniques are presented. Diffusion related defects in Ge substrate are described using DLTS.

Resume : The control of electrically active dopant distributions in nanosized semiconductor devices is crucial as it affects the device performance. Compared to the earlier two-probe spreading resistance profiling technique that allows measuring dopant profiles with penetration depths of several microns, the quantitative measurement of active dopant concentrations in the submicron and nanometer range remains challenging. Scanning spreading resistance measurements (SSRM) conducted by means of an atomic force microscope is a powerful technique to measure the local resistance in the nanometer range. However, quantitative data on the concentration of active dopants rely on calibrated reference samples, complex data reduction, and a sophisticated sample preparation. In this work n- and p-type dopant profiles in germanium (Ge) are analyzed with SSRM. The obtained active dopant concentration is compared to the total concentration determined by means of secondary ion mass spectrometry. Numerical simulations of dopant diffusion in Ge reveal that differences in the total and active dopant concentration are consistently explained by the formation of inactive dopant-defect complexes. Our results demonstrate that quantitative SSRM is highly suited to study the mechanisms of dopant diffusion on the nanoscale.

Resume : Silicon and germanium nanowires (SiNWs and GeNWs) are anticipated for the realization of next-generation metal-oxide-semiconductor field-effect transistors. Impurity doping is one of the key techniques for the NWs devices [1], while the retardation of carrier mobility due to impurity scattering has to be taken into account. Core-shell NWs composed of Si and Ge are key structures for realizing high mobility transistor channels, since core-shell structures separate the carrier transport region from the impurity doped region, resulting in the suppression of impurity scattering [2,3]. Si/Ge and Ge/Si core-shell NWs were rationally grown on a Si substrate by CVD. The TEM and EDX images of i-Ge/p-Si (i: intrinsic, p: p-type) core-shell NWs clearly show the formation of i-Ge/p-Si core-shell NWs and clear lattice fringes in the shell regions. To confirm the selective B-doping in the shell region, we performed Raman measurements. The electrical activity of B atoms can be clarified by the Fano effect [2], which is due to coupling between discrete optical phonons and the continuum of interband electron excitations in degenerately doped p-type Si. The Si optical phonon peaks observed for i-Ge/p-Si NWs shows an asymmetric broadening toward higher wavenumber, whereas no asymmetric broadening was observed for the i-Si shell. This asymmetric broadening is due to the Fano effect, showing that B atoms are electrically activated in the Si shell, resulting in the formation of p-Si shell. Precise analysis using Raman spectroscopy were further performed. We finally got a conclusive evidence of hole gas accumulation in Ge/Si core-shell NWs, which is important for the realization of high speed next generation transistor channels [3]. References [1] N. Fukata, Adv. Mater 21, 2829 (2009). [2] N. Fukata et al., ACS NANO 6, 8887 (2012). [3] N. Fukata et al., ACS NANO 9, 12182 (2015).

Resume : Light emission from Si/Ge based nanostructures has attracted much attention in the field of Si-based photonics because of its potential to combine photonic processing and electronic processing in a single chip. So far, we have demonstrated high density formation of quantum dots (QDs) consisting of Si clad and Ge core by controlling thermal decomposition of SiH4 and GeH4, alternately, on thermally-grown SiO2 and their unique charge storage characteristics associated with type II energy band alignment between the Si clad and the Ge core, that is, holes are store in the Ge core but electrons in the Si clad. And we also reported that, in the case that a single layer of QDs having a 6nm Ge-core and a 3nm-thick Si-clad in average size on SiO2 was excited by 976nm photons, stable PL signals consisting of four Gaussian components were detected in the energy region from 0.66 to 0.88eV at room temperature, and that the components are attributable to radiative recombination through quantized states in QDs as verified from dot size dependence of PL peak energy and temperature dependence of PL properties. In this work, based on the fundamental study of photoluminescence, we have designed and fabricated stack structure of QDs sandwiched with ~2.0nm-thick SiO2 as a bottom tunnel oxide on p-Si(100) and ~10nm-thick SiO2 as a control gate oxide and studied electroluminescence through Si substrate in alternate injection of electrons and holes to QDs from the Si substrate under cold light illumination. Capacitance voltage characteristics measured at 1MHz confirm that, under the cold light illumination, photo-generated electrons in the peripheral region of the area masked with metal gate flow into the inversion region formed underneath the gate oxide and respond to the pulsed gate bias even at high frequencies. No EL signals was detectable with zero gate bias while, with application of continuous square-wave bias with peak-to-peak amplitudes of ±1.0V and higher at 500 kHz, EL signals having similar components to PL signals were observed from the backside through the c-Si substrate even at RT. With an increase in the bias amplitude from ±1.0 to ±4.0V, the EL intensity was increased by about one order of magnitude and higher emission energy components become significant with almost no change in the peak energy of each component. Notice that, by applying square-wave bias of -4 to 0V, EL was observable. The result indicates electron charging in QDs at -4V induces hole injection even in 0V half-cycle. From the bias amplitude dependences of the EL intensity, in which positive and negative bias amplitude are changed independently, we have found that hole injection is a rate-limiting factor for EL at ±2V and below because of a smaller tunneling rate of holes than electrons while electron injection becomes a major factor for EL at ±3V and over as a result that electron emission from QDs become significant in positive half cycle.

Resume : The nucleation and growth of Mn doped GeSi islands grown by self-assembly via Stranski-Krastanov growth mode on Si(100) at Ge and Mn joint deposition at 400 and 450 C were studied. The Mn content in quantum dots was varied from 5 to 22%. Samples were studied with AFM, STM, Rutherford backscattering (RBS), Raman scattering, and Energy-dispersive X-ray spectroscopy (EDXS), electron spin resonance (ESR) and magnetometer MPMS-XL «Quantum Design». The Mn deposition was found to enhance diffusion coefficient of surface atoms and as a result the enlargement of islands and decreasing their density takes place. At high Mn density (22%) the Mn-catalyzed nanowires growth apparently realized, when rate of the islands growth in one direction strongly exceeds another one. RBS studies of heavily Mn doped (22%) of the GeMnSi/Si structures, grown at 450 C and caped with Si at 400 C have shown that the peak of Ge depth distribution is shifted to surface as compared with Mn peak position. It apparently means that Mn diffusion to the tensile strain Si below Ge islands takes place and Mn inclusions produced under quantum dots in GeMnSi/Si heterostructures. In Mn undoped Ge/Si quantum dots structures tensile strain due to the 4% lattice mismatch between Ge and Si causes splitting in the nearby Si of the sixfold-degenerate ? valleys into the fourfold-degenerate in-plane ?xy valleys and twofold-degenerate ?z valleys along the [001] growth direction. The lowest conduction band edge is formed by the ?z valleys yielding the triangle potential well for electrons in Si near the apex of pyramidal shape Ge nanocrystal. The direct evidence of the electron localization place was obtained by ESR measurements which have shown that values of the parallel and perpendicular g-factor components exactly coincide with electron g? and g? in strained bulk Si, which confirms a top priority of strain induced electron localization in the Ge/Si QDs structures. The ESR measurements at low (5%) Mn doped GeMnSi/Si structures have shown signal similar to Mn undoped structures. As Mn content is increasing the intensity of ESR signal was found to decrease and g-factor shifted to lower values. The decreasing of ESR signal can be explain by the lowering of tensile strain around GeMnSi quantum dots and disappearing of potential well for electrons in Si at Si/Ge interface. The nature of g-factor shifting is not clear at present. This shifting correlates with increasing of magnetic moment of the GeMnSi/Si quantum dot structures with increasing of Mn content, so it may be influence of magnetic field of ferromagnetic Mn inclusions. To clarify the nature of g-factor shifting the additional experiments should be done. This work was funded by Russian Fund for Basic Research, (Grant No. 16-02-00397).

Resume : Photoluminescence (PL) measurement after the introduction of C-related radiation damage attracts considerable attention to quantify C in the concentration range lower than the detection limit of the IR absorption method. However, the necessity of the sample cooling down to liquid He temperature has been obstacle to the popularization of the technique. In this paper we demonstrate that the C-related C- and G-lines are observable at liquid N temperature with broader line width by using the optimized low-dispersion spectroscopic apparatus. The dopant-impurity-related lines which are dominant in the near band-edge region at liquid He temperature disappear at liquid N temperature, resulting in the simple spectrum consisting of only the C- and G-lines and the band-edge emission. Positive correlations were found between their intensity ratios to the band-edge emission and the C concentration estimated by the IR method and/or PL at 4.2 K. The broadening of the C- and G-lines brought about higher sensitivity and faster measurement, and the extinction of the dopant-related lines and little effect by temperature fluctuation lead to the improvement of the quantification accuracy. These are the great practical advantages for quantifying low-level C in Si in addition to the convenience of liquid N temperature. We also demonstrate that a broad emission band is observable, even at room temperature, in the vicinity of the C-line and discuss its origin and applicability for quantification of C.

Resume : Point defect complexes in silicon have a significant impact on the electrical and optical properties of the material. The interstitial carbon?oxygen complex (CiOi) is one of the dominant electrically active complexes in irradiated p-type Cz-Si with a deep level at 0.36 eV above the valence band edge (EV). CiOi is stable up to 400 °C, where it anneals out leading to formation of a new defect. This defect has a deep level at EV+0.39 eV and is tentatively identified as an interstitial carbon?dioxygen (CiO2i) complex. The CiOi is suggested to dissociate into Ci and Oi and the released Ci?s can be trapped by oxygen dimers (O2i?s) forming the stable CiO2i complex. Light-induced-degradation (LID) of Si solar cells can be attributed to interstitial boron?dioxygen (BiO2i) complexes, but no spectroscopic signal has been detected for BiO2i. CiO2i is expected to have comparable atomic structure to that of BiO2i, and understanding the evolution of CiO2i may provide insight into the thermal behavior of BiO2i leading to a resolution of the LID issue. In the present study, the kinetics of the tentative CiO2i complex has been investigated employing deep level transient spectroscopy (DLTS) analysis of proton irradiated boron-doped Cz-Si samples. Thermal anneals were carried out on the samples in a dry N2 ambient for temperatures in the range of 450-550 °C. The annealing of CiO2i obeys first-order kinetics and appears to occur via dissociation mechanism with an activation energy around 2.5 eV.

Resume : Transition metals (TM’s) are common contaminants in silicon solar cells produced by block casting or using lower grade polysilicon feedstock. Although certain metals can be removed by gettering, slow diffusing TM’s can remain and have a detrimental effect on cell efficiency even at low concentrations. One method to reduce their effects is passivation using hydrogen; however this process is generally ineffective for TM’s in p-type silicon, as the metal and hydrogen atoms are typically both in positive charge states. Recent studies have suggested that under certain conditions, such as e.g. illumination, the charge state of hydrogen in silicon can be altered, and its bonding with similarly charged species can be enhanced. In this work we have studied TM-H bonding in p-type silicon intentionally contaminated with Mo and hydrogenated via remote H plasma. We have used DLTS, Laplace DLTS and capacitance-voltage measurements to study the effects of heat-treatments at moderate temperatures (<250 °C) under illumination with an 830 nm LED on the formation of Mo-H complexes. We find that in illuminated samples, the concentration of electrically active Mo is reduced in regions where hydrogen is present, with no reduction seen in samples annealed in the dark. The passivated fraction of Mo can be recovered after dark annealing at temperatures >200 °C. Apparently, under illumination bonding of Mo and H atoms in p-type Si occurs. Possible mechanisms of the observed effects are discussed.

Resume : Nanoelectromechanical system (NEMS) switch have advantages such as low static power consumption and strong robustness to the harsh environment. However, NEMS switches require high operating voltages, which make it be difficult to use in practical applications. We recently demonstrated that a the one of the NEMS switch types, FinFACT, can reduce the operating voltage using vacuum effect. The FinFACT is similar to a separated double-gate FinFET which have no gate oxide but an air gap. The movable channel (fin) of the FinFACT switches on and off through the electrostatic force between the gate and the fin. When a high voltage is applied to gate-1, fin adhere to it due to an electrostatic force between each other. Similarly, a high gate-2 voltage pull the fin which was attached to gate-1 to detach from it. But this is not easy work. When the gate-1 and fin are attached by electrostatic force, The van der Waals force and the Capillary force are generated on interface. So, high operating voltage is necessary to detach the fin from gate-1. One of the causes of strong adhesion force is capillary force, which can explain that a thin liquid layer between the two contacted surface can work as an adhesive. Also, the liquid bridging between the two adhesion surfaces drastically increase capillary force by a large increase in the real contact area (RCA) Therefore, it is necessary to remove the moisture in the device as a method to reduce the adhesion force and ultimately to reduce the device operating voltage. This is achieved through the vacuum chamber. In vacuum, it was confirmed that the required operating voltage was reduced when the fin was removed from one gate. This experimental result suggest one of the obvious solutions to the NEMS switch that is experiencing difficulties in practical applications due to its high operating voltage. Therefore, lower operating voltage can be applied by ambient engineering, such as a chip level–vacuum package

Resume : The internal gettering technique is becoming more important regarding Si-based Very Large Scale Integration (VLSI) manufacturing processes because device features are continually shrinking. In addition to well-investigated metals, such as Fe, Ni, and Cu atoms, other atoms that diffuse very slowly in Si, such as W, Hf, Ti, and Mo, have become gettering targets. Since the efficiency of internal gettering is determined by the competition of target metal atoms between the gettering site and wafer surface, an understanding of both the gettering mechanism and the stability of metal atoms near the Si (001) wafer surface is necessary. In the present study, we used first-principles calculation to investigate the stability of 12 metal atoms (Ti, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Hf, Ta, and W) near the (001) surface of Si wafer. The formation energy of each metal atom at the interstitial tetrahedral sites in each atomic layer was obtained by using a calculation model that included the (001) Si surface. The main results are as follows: (1) Ti, Zn, and Hf atoms were the most stable at the outermost surface, while the others were most stable in the second layer. (2) The formation energy of these 12 metals decreased by about 0.9-2.4 eV near the surface compared with that of the Si bulk. (3) The formation energy of these 12 metals was almost constant at layers deeper than the fifth layer from the (001) Si surface. Additional calculations for the impact of ultra-thin oxide films for the stability of metal atoms etc. are ongoing.

Resume : An essential factor for achieving low-loss insulated-gate bipolar transistors (IGBTs) is controlling the bulk lifetime of phosphorus (P)-doped n-type silicon (Si) crystals. In order to do this, recombination centers, such as vacancy-vacancy (V-V) and vacancy-phosphorus (V-P) pairs are introduced by electron beam irradiation. In the case of using Czochralski (CZ) Si wafers, an interstitial carbon-interstitial oxygen (Ci-Oi) pair is also considered to affect the bulk lifetime. One of the technological problems is that the carrier lifetime is extended under the device function. This is probably due to the inactivation of deep energy levels of V-V and V-P pairs because of the interaction of Ci, Oi, and Ci-Oi pair. The purpose of this study is to understand the mechanism behind the formation of recombination centers (V-V and V-P pairs) and also the structural change of these pairs with the interaction of Ci, Oi, and Ci-Oi pair. To achieve this, first-principles calculation was performed on possible reactions of intrinsic point defects (vacancy V and self-interstitial Si I), dopant (P), and impurities (C and O), which exist in the P-doped CZ-Si crystal. Specifically, binding energy was calculated for each reaction of Cs+I, Ci+Oi, P+V, P+Ci, P+Oi, V+V, V+Ci, V+Oi, P+CiOi, and V+CiOi. The main results are as follows: (1) The effect of one V is limited to the other V at the first neighboring site, whereas the effect of P extends to V at more than the ninth neighboring site. (2) A Ci-Oi pair interacts with V and dissociates into single Cs and Oi. (3) Complexes of Ci with V, Oi, and P are formed, and the binding energy of Ci to V is the highest, whereas to P, it is the lowest. The results for the interaction between V-V, V-P, and Ci, Oi, Ci-Oi etc. will also be presented.

Resume : This research deals with preparation, characterization and performance analysis of methyl ammonium lead halide (CH3NH3PbI3) based perovskite solar cell. The rectangular active area of the cell was 280 mm2. The electron transport layer (ETL) was prepared using meso-porous TiO2 and the main solar absorption layer CH3NH3PbI3 was prepared by two steps spin coating method. N, N-dimethyl formamide (DMF) and dimethyl sulfoxide (DMSO) were used as solvents for PbI2. To facilitate complete reaction, dipping and additional lead halide such as PbCl2 were used. The hole-transport layer was made of spiro-MeOTAD with varied coating thickness. Initially the ETL was optimized under different high temperature annealing condition ranging from 450 to 600°C. The TiO2 mesoporous structure was optimized and good formation was confirmed by the XRD spectra. In addition, overall coating thickness, surface morphology and crystal structures were optimized by implementing changes in the concentration of the precursors used and spin coating conditions. Our main objective is to engineer interfaces between layers focusing on reduction in the shunt resistance and charge recombination. Considerable power conversion efficiency has not been achieved yet (less than 5% PCE till now) as compared to the reported value but based on the present results, an appreciable improvement is expected further.

Resume : Quasi-chemical reactions between intrinsic point defects and dopants in irradiated Si and Ge of n- and p-type are often believed to be similar. It would be of particular interest to compare the experimentally determined formation rates of such complexes with group-III and V impurity atoms. This is important for modeling of radiation resistance of both materials to damage at room temperature. In this report direct information on losses of shallow donor/acceptor states of the impurities due to formation of impurity-related complexes in Si and Ge under fast electron and proton irradiation is analyzed. The collected data show a very pronounced difference in the behavior of n- and p-Ge under 1 MeV electron- and 15 MeV proton irradiation as compared to that observed in Si of n- and p-type. The reasons of the observed dissimilarity in the behavior of Si and Ge under the irradiation are discussed.

Resume : In the work a change in the specific electrical conductivity σ of the n-type Si(111) surface during the implantation of Ba, Na, and Li ions with energy E0 = 1keV in different doses was studied. Implantation of ions (regardless of the type of ions) to a dose of D = 8•1014 cm-2 practically does not lead to a change in σ. This is probably due to the deep penetration of the implanted ions and their small contribution to surface reducibility. Although it should be noted that implantation of Ba and alkaline elements with a dose of D~1014 cm-2 leads to an increase in the electron concentration at the donor levels and to the beginning of the splitting of the donor levels [1]. However, at these doses, the surface region of Si(111) is strongly disordered, which leads to a decrease in the electrical conductivity of the surface, which compensates for the contribution of an increase in donor concentration to the growth of σ. The validity of such a mechanism is indicated by the minima on dose dependences σ. With an increase in the dose of implanted ions, a sharp increase in σ is observed up to D = 1017 cm-2. Subsequent thermal heating leads to an increase in the electrical conductivity of ion-implanted samples. And at a certain temperature (for example, at T = 720 K in the case of implantation of Ba ions), bending is observed, which indicates the crystallization of the ion-implanted layer and the formation of a thin nanosized film [2]. An estimate of the thickness of a silicide film by layer-by-layer Auger analysis showed that for an ion energy of E0 = 1 keV, the film thickness is 5-6 nm (or 50-60 Å). 1. A.S. Rysbaev, A.A. Rysbaev, J.B. Khujaniyazov, A.M. Rakhimov, L.Kh. Rakhimova.// Uzbek Journal of Physics. 2013.Vol.15 N1-2.c 26-32. 2. A.S. Rysbaev, J.B. Khujaniyazov, A.M. Rakhimov, I.R. Bekpulatov. Formation of Nanosize Silicides Films on the Si(1111) and Si(100) Surfaces by Low_Energy Ion Implantation. // Technical Physics, 2014, Vol. 59, №. 10, pp. 1526–1530.

Resume : The change of recombination properties of Co-60 gamma irradiated Czochralski grown (Cz) silicon of n-type with concentration of free electrons ~ 10^14–10^16 сm^-3 after 20-min isochronal annealing in the temperature range 180–380 °С, in which the formation of V2O complexes due to a transformation from V2 and their annealing occurs, are investigated in detail. It is found that, at room temperature and for low excitation level, the nonequilibrium charge carrier lifetime (τ) significantly decreases in the range ~ 180–280 °C and this effect is stronger in low-resistivity n-Si. At higher temperatures of annealing, the τ increases and recovers to ~ 80–90 % of the initial value at ~ 360–380 °С. It is shown that τ change in the range ~ 180–380 °С is caused by the V2 related defects and has a clear anti-correlation with a change in the V2O concentration. Assuming that these V2 related defects are the V2O complexes, the Shockley-Read-Hall model was used for the description of experimental data. It was determined that the formation of V2O complexes is characterized by the activation energy of 1.25±0.05 eV and frequency factor (10±5)×10^8 s^-1, their annealing have activation energy of 1.52±0.08 eV and frequency factor of (2±1.5)×10^10 s^-1. The values of hole capture cross-section by the singly and doubly charged acceptor states of V2O are obtained: (5±2)×10^-13 and (10±4)×10^-12 cm^2, respectively.

Resume : In the work change in the specific electric conductivity of the n-type Si(111) surface during the implantation of Ba, Na and Li ions with energy E0=1 keV in different doses was studied. Implantation of ions (regardless of the type of ions) to a dose D=8•1014 cm-2 practically does not lead to a change in σ. This is probably due to the deep penetration of the implanted ions and their small contribution to surface reducibility. Although it should be noted that the implantation of Ba ions and alkaline elements with a dose of D1014 sm-2 leads to an increase in the electron concentration at the donor levels and to the beginning of the splitting of the donor levels. However, at these doses, the surface region of Si(111) is strongly disordered, which leads to a decrease in the electrical conductivity of the surface, which compensates for the contribution of an increase in donor concentration to the growth of σ. The validity of such a mechanism is indicated by the minima on dose dependences σ. With an increase in the dose of implanted ions, a sharp increase in σ is observed up to D=1017 sm-2.

Resume : Polycrystalline silicon (poly-Si) thin films were fabricated by aluminum-induced crystallization of silicon suboxide (SiOx) synthesized by gas-jet electron beam plasma chemical vapor deposition (GJ EBP CVD) method for the first time. SiOx films (90 – 400 nm) were deposited using SiH4, H2 and O2 gases on borosilicate Corning XG glass substrates covered by aluminum (Al) layers. According Fourier-transform infrared (FTIR) spectroscopy measurements, stoichiometric coefficient of initial films was 0.23. In addition, the films contained small quantity of bonded hydrogen. Thereafter, glass/Al/SiOx samples were annealed in a classical tube furnace at 450 – 550 C during 4 – 20 hours under vacuum. Annealing process led to Al and SiOx layers exchange and formation of large Si grains (few hundreds of microns). Layer exchange was confirmed by examination by scanning electron microscopy of cross-sections of the initial and annealed samples obtained on crystalline silicon wafers with SiO2 sublayer. The influence of the annealing temperature and duration on the sample properties (the fraction area of Si grains, Si grain size, and their crystalline orientation) was investigated by optical microscopy, Raman spectroscopy, and X-ray diffraction method. The results suggested that SiOx is a promising, alternative material for low-temperature formation of large-grained poly-Si films. This study was financially supported by the Russian Science Foundation, project # 17-79-10352.

Resume : Fabrication of Ge nanocrystals in a multilayer system in which the nanocrystals are separated by a dielectric matrix such as stoichiometric SiO2 or Si3N4 gives the possibility to control the size and density of nanocrystals independently. These type of superlattice multilayers have potential application in third generation solar cells where tunneling of charge carriers between nanocrystals is essential for charge transport. In this paper, we report on the fabrication of (SiGeN/SiO2) and (SiGeN/Si3N4) multilayers via plasma enhanced chemical vapor deposition (PECVD) and post annealing in Ar ambient. A detailed structural study have been done by Raman spectroscopy, high resolution transmission electron microscopy (HRTEM) and deep X-ray photoelectron spectroscopy (XPS). Raman spectroscopy and HRTEM micrographs confirm crystallization of nanocrystals annealed in appropriate conditions. Depth profiling with XPS shows the composition of samples before and after annealing. Our analysis show that stoichiometric dielectric separation layers perform as a good barrier for diffusion of Ge atoms during annealing. Roughness of films also preserve unchanged during annealing with fairly high roughness. Possible origin of waviness of films observed in TEM micrographs of samples with compressive stress is also discussed. Origin of the photoluminescence have been discussed.

Resume : Selective contacts used in crystalline silicon solar cells require both good surface passivation and very good carrier conductivity, i.e. low ohmic contact resistances, to reach high-efficiency solar cells. Traditionally selective contacts are made using high doped regions by thermal and/or laser process diffusion stages, which are expensive and high-cost energy. Alternatively a cost-effective approach to obtaining good selective contacts involves the use of transition metal oxides (TMO) for electron or hole transport layers, ETL and HTL respectively. In order to improve the conductivity of that selective contact a dipole interlayer can be used. Using dipole interlayer a better energy band alignment can be achieved by reducing the work function of the electrode. In this work, we report on the use of a polymer and a hydroxide dipole (δ) to improve ohmic contact with the following configuration Al/ δ /TMO/n-Si. We use as dipole poly [(9,9-bis(3′-(N,N-dimethylamino) propyl)-2,7-fluorene)-alt-2,7-(9,9-dioctylfluorene) ] (PFN) and barium hydroxide [Ba(OH)2]. Best results were obtained using PFN dipole interlayer, with contact resistances about 2.5 mΩcm2. Finally, crystalline silicon solar cells using V2O5 as HTL and TiO2/PFN as ETL were fabricated with Voc= 0.61 V, Jsc= 30.5 mA/cm2, FF=76% and 14% conversion efficiency.

Resume : Radiation-induced defects degrade performance of Si detectors used for many applications. In this work we present results of a study of a radiation-induced bistable center, which has been detected in p-type boron doped silicon crystals by means of DLTS. The DLTS measurements show that in the lowest energy configuration the defect is not electrically active in p-type Si. However, upon injection of minority charge carriers, the center transforms into a metastable configuration with deep energy levels at Ev + 0.45 and Ev + 0.54 eV. The reverse transition to the main configuration occurs in the temperature range 50–100 ºC with the activation energy of 1.25 eV. The defect is thermally stable at temperatures < 250 ºC. It is argued that this defect can either be a complex of a self-interstitial silicon atom (I) with an interstitial carbon atom (Ci-I) or a complex consisting of I with an interstitial boron atom (Bi-I). Density functional theory calculations of the Ci-I complex have been performed. Two stable configurations of the neutral Ci-I defect have been found. The calculated binding energy of the Ci-I center in the minimum energy configuration is 1.65 eV. This value is consistent with the experimental results on the thermal stability of the bistable defect detected by DLTS. The metastable configuration is higher in energy by about 0.3 eV compared to the most stable one. The calculated electronic and vibrational characteristics of the Ci-I defect will be reported.

Resume : From all the research that has been collected since the discovery of graphene, no doubt, it is the material of the future. Excellent electrical and thermal conductivity, elasticity and strong covalent bonds between neighboring C atoms would make graphene a perfect candidate in semiconductor electronics; however, for this application an energy band gap must be opened. In this work we focus on controlling the size of a band gap while creating defects on zigzag and armchair graphene nanoribbons adsorbed on a single graphene sheet. Zigzag and armchair structures are created of different widths, adsorbed on graphene at different distances from each other and doped with boron or nitrogen at varying densities. Energy variations and spin properties are tested in order to build a quantum model and present how the energy fluctuates depending on those variables. To achieve that, we plan to perform the first-principle calculations using Density Functional Theory (DFT) in the CRYSTAL package. Our goal is to create a tunable band gap while simply changing the parameters listed above. We hypothesize that combining semiconducting and magnetic properties in one system, we could construct a device that could be used in quantum information processing, solar cells, spintronic devices or field effect transistors.

Resume : A new light source is not just simple illumination of light but it should be in harmony with its surrounding environment and have an interior decoration effect when it is turned off, and it should provide the sunlight-like or moonlight-like lighting when turned on. For this, the illumination style is expected to be design-in or built-in light. Also it would be a flexible illumination with light quality being glare-free, tunable color temperature, thin and lightweight. In this study, cadmium-free quantum dots were used as the light source of flexible illumination. Quantum dots hybrid illumination is optimal for indoor lighting as it allows high color rendering index. Quantum dots sheet was prepared for the quantum dots illumination and its brightness uniformity and color rendering index were higher than 98% and Ra 92, respectively.

Resume : Al2O3 is regarded as a promising alternative of SiO2 as a gate dielectric, since it has a higher dielectric constant and a higher thermal conductivity than those of SiO2. Although there are several methods of depositing Al2O3, the deposition by atomic layer deposition (ALD) is of large interest because of its controllability, stability, and uniformity. Furthermore, it is well known that the post-deposition thermal treatment crystallizes ALD Al2O3 in part and enables it to obtain chemical stability against acids exposed during manufacturing processes. It is considered that other characteristics can also change; however, the thermal property of the annealed film has not been reported. We deposited Al2O3 films by ALD in the thickness range from 10 nm to 50 nm, and annealed half of them at 900℃ in 100% N2 ambient for 1 hour. We measured the temperature oscillation by using the 3ω method for both annealed films and as-deposited films, and then extracted the thermal conductivity. As a result, the thermal conductivity of the annealed film proved to be higher (3.7 Wm-1K-1) than that of as-deposited (1.8 Wm-1K-1). Besides these measurements, some physical analyses were also undergone as for the 50 nm thick Al2O3 film; the density and the crystallinity were measured by XRR and by TEM respectively. According to these results, the as-deposited amorphous Al2O3 were changed to γ-Al2O3 by annealing.

Resume : ZnO is one of the most promising transparent conductive oxide (TCO) to develop optoelectronic devices for a variety of applications. Besides, ZnO thin films can be used as a TCO in organic light emitting (OLED) devices. In this work, we aim to synthesis ZnO thin films by RF magnetron sputtering in order to use them as TCO in the OLED while controlling their optical guiding properties. Thus, the objective of this project is to obtain thin ZnO films transparent, conductive and of high optical quality. ZnO were deposited by RF magnetron sputtering technique on glass and silicone substrates (using a ZnO ceramic target, power of 200 W and working pressure of 0.5 Pa, with 40 sccm of Ar). To improve the film transparency, ZnO were prepared in reactive atmosphere (10% of O2 in gas mixture) at room temperature and 200°C. The SEM observations of ZnO films deposited on silicon show a typical columnar structure of the films. The film deposited at 200 °C has a denser structure. The AFM observations indicated a roughness of 3.6 and 7.8 nm for ZnO films deposited at room temperature and 200°C, respectively. These results are confirmed by m-lines spectroscopy, which shows that the ZnO layers behave as monomode waveguides with very low losses (< 0.1 dB/cm). The analysis of the optical anisotropy indicates that ZnO filsm are uniaxial with the optical axis perpendicular to the surface of the substrate confirming the columnar growth revealed by SEM. Work continues on studying Al-doped ZnO.

Resume : Indium Tin Oxide (ITO) is widely used in solar cell devices for its excellent electrical and optical characteristics, such as high transparency (around 80%) in the ultraviolet-visible range and good conductivity (10^4 S cm^-1). Research on Transparent Conductive Oxides is always up to date because of the continuous need for better solar cells performances. In this work we have grown thin (70-100 nm) ITO layers by DC sputtering without adding oxygen in the deposition chamber. We have used different substrate temperatures during growth and have afterwards thermally annealed the samples at different temperatures up to 300°C to investigate the effects on the electrical and optical properties of the material. We have found out that the different growth/annealing conditions induce changes in the optical properties of the samples, especially on the films absorption, both in the visible and in the infrared range, as well as in the conductivity and carrier concentration. In addition to this, also the effect of introducing different amounts of hydrogen in the sputtering gas mixture has been investigated. The absorption changes have been evidenced by Quantum Efficiency measurements performed on heterojunction solar cells.

Resume : In recent years, contact resistance has become an important aspect in device-delay analyses. In order to lower the RC delay of a device, its contact resistance must be lowered. In an effort to minimize contact resistivity, a high doping concentration is required. Recently, it has been reported that the presence of phosphorus in silicon–phosphorus epilayers causes a silicidation delay in nickel silicides or changes the silicide alignment in the case of erbium silicides.[1, 2] However, there is insufficient evidence to confirm the occurrence of the above-mentioned effect of phosphorus on silicides in in-situ highly doped silicon–phosphorus epilayers. We investigated phase transformation of titanium silicide in silicon epilayers doped with various concentrations of phosphorus. The investigation reveals that titanium silicide undergoes a change of phase from C49 to C54 at a specific annealing temperature, which varies based on the doping concentration. At the high concentration substrate, the transition temperature from C49 to C54 increased by 50℃. The surfaces as well as microstructures of the films thus obtained are reported. It has been shown that despite the silicidation delay that occurs in highly doped substrates, Ti-silicide films demonstrate similar volume expansion across all doping concentrations. In addition, it was confirmed that the contact resistance is reduced as the doping concentration of the substrate increased. Through this, it was found that the high doping concentration of the substrate caused the silicidation delay but the physical and electrical characteristics of the formed silicide were not changed. [1] A.P. Peter, et al, “Characterization of ultra-thin nickel–silicide films synthesized using the solid state reaction of Ni with an underlying Si:P substrate (P: 0.7 to 4.0%)” Microelectronic Engineering 157 (2016) 52-59 [2] Jinyong Kim, et al, “Doping and strain effects on the microstructure of erbium silicide on Si:P” Journal of Alloys and Compounds 727 (2017) 728-734

Resume : Silicon nanocrystals (Si-NCs) represent possible non-toxic alternative to light emitting III-V and II-VI based NCs. By engineering their surface passivation, the band structure and emission mechanism are affected - where oxide capping leads to slower radiative rate and size-insensitive defect states inside the bandgap; organic capping linked by Si-C bond leads to enhanced radiative rates and larger absorption cross-section without affecting the bandgap [1-2]. In this work we investigate photoluminescence (PL) anisotropy of oxide and organically capped Si-NCs of various sizes, to distinguish between band-like (degenerate) and localized (discrete) state's radiative transitions. Si-NCs are excited using polarized ps-pulsed laser and the degree of emission polarization is measured in time. The de-polarization is possible via several channels: degenerated transition levels, transfer to an emissive defect site or particle diffusion. In both, oxide and organically capped Si-NCs, we find non-zero PL anisotropy, which indicates that radiative transition occurs from discrete state. The smaller Si-NCs appear to have higher initial polarization and depolarize slower as well, which could be related to higher degree of energy levels discretization. The larger Si-NCs, on the other hand, depolarize faster, possibly owing to higher degeneracy of the levels. [1] K. Dohnalova et al., Light: Sci. and Appl. (2013) 2, e47. [2] A. N. Podubny and K. Dohnalova, Phys. Rev. B 90, 245439 (2014).

Resume : Si/Ge layered structures with low thermal conductivity in the stacking direction (z) are investigated using the participation ratio obtained by phonon mode calculations. We focus on various types of Si/Ge layered structures and perform normal mode analyses to calculate the participation ratio (A. Bodapati et al., Phys. Rev. B 74, 245207 (2006)) for each phonon mode. From the participation ratios, low thermal conductivity structures are roughly predicted using genetic algorithm. In order to improve prediction accuracy of low thermal conductivity structures, we decompose the participation ratio into the stacking and in-plane direction components and analyze the stacking direction component (Pz). Pz's are averaged in the frequency ranges with their lower and upper limits independently optimized. The descriptor (D) for thermal conductivity is defined using the averaged Pz with taking stacking periods into account. Thermal conductivities in the stacking direction (kz) are calculated using perturbed molecular dynamics simulations (M. Yoshiya et al., Mol. Simulat., 30, 953 (2004)). A very strong correlation is found between the kz and D. Our results show that the participation ratio can be used to find Si/Ge layered structures with low thermal conductivity among a huge number of possible structures without direct calculations of thermal conductivities. Our approach is effective to accelerates the designing of Si/Ge materials with high thermoelectric performance.

Resume : The current LED technology is limited by a strong efficiency drop in the green-yellow-emission range known as the green gap. The total efficiency of a commercial white light LED is therefore limited. The semiconductor alloy Zn(Sn,Ge)N2 is similar to the InGaN system and is promising for the enhancement of optoelectronic devices because of its adjustable band gap and lattice parameter. We report on the study of ZnSn{x}Ge{1-x}N2 thin films with x = 0 to x = 1, grown by reactive co-sputtering. At x = 0, ZnGeN2 is promising to improve the LED efficiency in the green-yellow-emission range, due to its large and direct band gap of 3.3 eV, as well as a good lattice match with GaN [1]. At x = 1, ZnSnN2 is promising in regard of solar cell technology because of its good electrical properties and its direct band gap of 2.1 eV [2]. No phase separation occurs through the whole range of the alloy and we observe a linear relationship between composition and properties like optical band gap, lattice parameter and peak positions in FTIR and Raman spectroscopy. Electrical properties are also investigated. [1]: L. Han, K. Kash and H. Zhao, J. Appl. Phys. 120(10): 103102 (2016). [2]: T. Veal, N. Feldberg, N. Quackenbush, W. Linhart, D. Scanlon, L. Piper and S. Durbin, Adv. En. Mat. 5(24): 1501462 (2015).

Resume : The higher carrier mobilities of Germanium (Ge) than Silicon (Si) combined with its optical and thermoelectronic properties have promoted interest in Ge. However, a controllable fabrication methodology of Germanium-on-insulator (GOI) structures is essential. Appropriate GOI structures must be fabricated as thin layers on Si substrates with high crystal qualities. Liquid-phase epitaxy (LPE) is a suitable method for epitaxy of Ge since this method offers shaping of Ge as desired during the growth. The previous works on GOI obtained by LPE have confirmed the expected epitaxial growth of Ge on Si with various insulating layers and capping layers [1, 2]. The length of the epitaxially grown c-Ge was shown to be as long as 1 cm [2]. However, there is still need for further studies in the field of LPE growth of Ge on Si substrates. In this work, we studied the effects of structural parameters such as insulating layer thickness, Ge layer thickness, capping layer thickness and temperature profile of LPE process on the epitaxy of Ge on c-Si substrates. The insulating layer and the annealing temperature were the main determining factors of a successful epitaxial growth. The geometrical shape and thickness of Ge layer were also observed to have high influence on the quality of the c-Ge. Besides the structural parameters, we observed the effect of different deposition methods for the insulating and capping layer. The crystallinity of Ge structures were confirmed with the Electron Back-Scattered Diffraction (EBSD) measurements. Further investigation was done with Raman analysis showing the crystal quality of Ge and the strain on the Ge. [1] S. Balakumar et. al, ‘Fabrication Aspects of Germanium on Insulator from Sputtered Ge on Si-Substrates’, Electrochemical and Solid State Letters, 2006 [2] H. Chikita et. al., ‘In-depth analysis of high-quality Ge-on-insulator structure formed by rapid-melting growth’, Thin Solid Films, 2013

Resume : The Monolayer Doping technique (MD) uses organic molecules as precursors for the surface doping of semiconductors. Precursors are grafted on the surface to form an adsorbed layer acting as a source of dopant for a subsequent thermal annealing process. Owing to its surface conformity, low process costs and a strict control of dopant amount, the MD approach is increasingly adopted as a novel doping process for 3D nanostructures. In this work, we studied the adsorption behavior of two different phosphorus molecular precursors (diethyl 1-propylphosphonate, DPP, and octadecylphosphonic acid, ODPA) on H-terminated Ge powder by means of Diffuse Reflectance Infrared spectroscopy (DRIFT), demonstrating in agreement with existing literature that DPP undergoes physisorption on Ge, whereas ODPA is chemisorbed via Ge-O-P bond formation. X-ray photoelectron spectroscopy (XPS) and Nuclear Reaction Analysis (NRA) analyses confirmed this scenario on Ge surface showing the presence of an oxidized Ge interlayer and quantifying the precise amount of adsorbed P. Different thermal and laser diffusion annealing tests were carried out and Secondary Ion Mass Spectrometry (SIMS) was applied to investigate the diffusion profile of P in Ge. These results highlight the importance of controlling Ge surface oxidation during the deposition step, confirming that a few Ge oxidized atomic layers can act as a diffusion barrier for P that only the high temperature involved in laser annealing process can overcome.

Resume : Minority carrier lifetime is the key material parameter for quantifying the electronic quality of silicon substrates used in photovoltaic cells. In multicrystalline silicon (mc-Si), a material used in the majority of manufactured PV cells, the lifetime is limited by recombination centres associated with metallic impurities in many forms, including point-like defects, precipitates and impurities bound to or precipitated at structural defects such as dislocations or grain boundaries. In this work we demonstrate how some metallic impurities, including interstitial iron, are sufficiently mobile that they can be redistributed by annealing at low temperatures (< 600 °C). We have used this to develop a variety of low thermal budget processes to enhance lifetime, including low-temperature internal gettering and low-temperature saw damage gettering. Our processes are demonstrated in a series of experiments carefully designed to take into account sample microstructure and artefacts arising from surface passivation. We find low-temperature internal gettering can improve lifetime by a factor of ~7 in bottom “red zone” wafers. Low-temperature saw damage gettering, whereby impurities are trapped in near surface regions before they are etched away, can also result in large lifetime improvements (a factor of ~4), and it has the potential to be added easily to the beginning of a commercial cell process.

Resume : The integration of highly doped Ge on Si with a controlled amount of tensile strain is crucial for several applications in advanced nanoelectronic and photonic devices. However, obtaining n-type doping above 5x1019cm-3 and in-plane biaxial tensile strain above +0.2-0.25 % with conventional growth and annealing methods is highly challenging. Here we report on the combination of in‑situ doping of Ge-on-Si epilayers and pulsed laser melting (PLM) to improve the activation of phosphorous in germanium. Secondary ion mass spectrometry measurements indicate that the box-like profile of as-deposited epilayers is preserved during PLM with minimal P out-diffusion. By improving the growth and PLM conditions an activated uniform n-doping concentration well above 1x1020cm-3 over 2-300 nm thick layers has been achieved, as measured by infrared reflectivity and differential VdP-Hall. Photoluminescence demonstrates clear bandgap narrowing and an increased ratio of direct to indirect bandgap emission confirming the high doping densities achieved. Finally, High Resolution X-Ray Diffraction and Raman measurements show that, thanks to the extremely high thermal gradients achieved by PLM, the in-plane residual thermal strain is increased well above +0.3%, favoring the electron population of the gamma-valley.

Resume : The Schottky junction-based technology is largely used elsewhere in the Si based power devices and photodetectors but is very new to the Au/ITO/Si photovoltaic technology. In this context, in this paper a new Figure of Merit (FoM) parameter which combines both electrical and thermal stability performances is proposed. Moreover, the impact of intermediate ITO ultra-thin film in enhancing the thermal stability of Au/Si Schottky Diode is presented. The proposed design thermal stability and electrical characteristics are investigated and compared with those of the conventional Schottky structures in order to reveal the device performance thermal stability. It is found that the amended Schottky diode design has a profound implication in improving the device reliability against thermal variations. Our experimental analysis is carried out incorporating the impact of the ITO thickness on the device performance, where an appropriate ITO thickness value can improve the electrical and thermal stability performances. This makes the Au/ITO/Si structure a potential alternative for high-performance and reliable power electronic and photovoltaic applications.

Resume : The main disadvantage of bulk silicon lies in its indirect bandgap and its consequent poor performance as a light emitter. Although silicon nanocrystals have been known to emit light much more efficiently than bulk silicon for quite some time now, most types of silicon nanocrystals still retain the indirect bandgap (and, consequently, low emission rates). In this contribution, we summarize our work on silicon nanocrystals engineered into a direct bandgap material using the cooperation of quantum confinement and tensile strain, which arises simply from methyl surface capping being attached to the silicon core of the nanocrystal [1]. Due to the direct bandgap, these band-engineered silicon nanocrystals possess radiative exciton lifetimes, and the corresponding radiative rates, improved by a factor of about 10,000, being on a par with other direct semiconductors. The focus of this contribution will be optical properties of this material, including luminescence spectroscopy of single nanocrystals [2,3]. The improvement in the luminescence of these nanocrystals proves wrong the notorious statements about the unsuitability of silicon as a light-emitting material due to its indirect bandgap. [1] Kusova et al. Adv. Mater. Inter. 1 (2014), 1300042. [2] Kusova et al. Light Sci. Appl. 4 (2015), e336. [3] Kusova et al. Phys. Rev. B 93 (2016), 035412.

Resume : The metal-insulator-metal (MIM) diode has been investigated with significant attention for various electronic device applications such as infrared (IR) photo detectors, rectennas for energy harvesting, hot electron transistors, high-frequency mixers, and selector or switching component in semiconductor memory. In most of the previous results about the MIM diode, the work function difference between the top and bottom electrode is regarded as the most essential parameter to obtain high asymmetry in current density. However, it requires the use of noble metal electrodes to ensure a sufficient work function difference for the high asymmetry, which might induce difficulties in integration process. Additionally, the relatively high applied bias voltage needed for a highest asymmetry in current density inevitably requires thick dielectric films. Therefore, another way of introducing the asymmetry in current conduction should be addressed in order to fulfill the requirements for the MIM diode while minimizing operating voltage and then the thickness of the total stack. In this study, we investigate the electrical properties of MIM diode made with TiN/HfO2-Al2O3-HfO2/Ti or Pt structure for achieving the rectifying behavior by exploiting the asymmetry of the bulk-limited current conduction. The current density is decreased as the Al2O3 layer is located far from the electron injection electrode (i.e., cathode). Also, the work function difference of electrode does not affect the rectifying behavior of the MIM diodes. Asymmetry of current conduction is attributed to suppression of bulk-limited current conduction by the interposed Al2O3 layer. Physicochemical analysis confirms that rectifying behavior arises from bulk-limited current asymmetry.

Resume : The growth of III-V compound semiconductors on lattice matched substrate is well reported. But the growth of III-V semiconductors on silicon platform is a challenging task due to lattice mismatch between them. Because of the lattice mismatch direct epitaxial growth of III-V semiconductor on silicon substrate could not be achieved other than the low dimensional structures (Quantum Dots, Nanowires etc.). In this work the growth of indium gallium arsenide is studied on silicon substrate followed by a gallium arsenide and an indium phosphide buffer layers. InxGa1-xAs layer is grown by Metal Organic Chemical Vapor Deposition (MOCVD) technique on p-Si (100) substrate having a carrier concentration of 1016 cm-3. InxGa1-xAs layer growth was carried out in a horizontal atmospheric pressure reactor. The source of indium and gallium were metal-organics, i.e. tri-methyl indium (TMIn) and tri-methyl gallium (TMGa) respectively. The source of arsenic was arsine gas (AsH3) and that of phosphorus was phosphine gas (PH3). The flow rate of TMIn, TMGa and AsH3 were 10, 6.8 and 45 sccm respectively for the growth of the InxGa1-xAs layer. The carrier gas was high purity hydrogen with a flow rate of 4 slpm. Growth of InxGa1-xAs was carried out at a temperature of 625 0C. The buffer layers are grown at 5500C. The grown layer is studied by UV-Vis-NIR reflectance spectroscopy, grazing incidence X-ray diffractometry (GIXRD) and X-ray photoelectron spectroscopy (XPS). Composition, band gap and thickness were estimated from UV-Vis-NIR spectroscopy. The crystalline quality of the grown film has been investigated by GIXRD. The binding energy of the electron in the XPS spectrum gives the elemental composition of the grown film. It also confirms the growth of InxGa1-xAs layer.

Resume : Surface passivation is a key requirement for high-efficiency silicon solar cell structures, and the passivation quality remain stable during different processing steps of the cell fabrication. Atomic layer deposited (ALD) alumina is found to be efficient in passivating Si surfaces. However, formation of interfacial SiO2 that contains positive fixed charges deteriorate the field-effect passivation characteristic of alumina. We present a method to avoid the formation of interfacial layers and achieve good quality of surface passivation in the present investigation. In this context, Intrinsic hydrogenated amorphous silicon (a-Si:H) films were prepared by rf magnetron sputtering for passivation of n-type silicon surface at room temperature. The films were deposited in H/Ar plasma with the gas ratio in the range, 0-0.5. The microstructural and optical characteristics of films with respect to processing conditions were characterized using appropriate methods. The interface state density (Dit) was tailored to minimum so as to achieve high quality passivation by tailoring the process conditions. The as deposited a-Si:H films showed minority carrier lifetime s < 20 ms and s increased to 314 s at an injection level of 10E15 cm-3 after annealing at 200 oC for 20 min and approached close to the bulk value, 500 s over extended annealing periods, indicating the saturation of dangling bonds at the interface of a-Si:H and silicon. The hydrogen content in the films was found to be < 9 %. The implied Voc changed from 557 mV to 591 mV after depositing the a-Si:H layers at an optimized condition. The method has shown excellent passivation quality and found to be environment friendly.

Resume : The best silicon based photovoltaic solar cell performance is obtained today with heterojunctions of hydrogenated amorphous silicon (a-Si:H) on crystalline silicon (c-Si). While a-Si:H offers outstanding passivating properties of the c-Si surface carriers it still causes some losses due to parasitic absorption when used as a front emitter layer, and high resistivity. Such drawbacks could be avoided if one replaces a-Si:H by a wide bandgap III-V compound. Gallium phosphide is a good candidate since it has a larger bandgap than a-Si:H (2.3 eV instead of 1.7-1.8 eV). Moreover GaP has a very low lattice mismatch to Si (0.36%) which allows direct planar epitaxy on Si. However, the epitaxial process may lead to the creation of defects at the interface and in the c-Si. In this work, we have use Admittance Spectroscopy (AS) and Deep Level Transient Spectroscopy (DLTS) to investigate trap states in GaP/Si heterojunction diodes. We have characterized various GaP/Si heterojunction devices with different hydrogen implantation conditions to change the GaP/Si interface properties. From AS no clear defect signature could be identified due to the very low defect concentration. However, DLTS being more sensitive allows us to reveal the existence of a trap state which electronic properties depend on the GaP/Si interface passivation conditions.

Resume : Light emission from Si, would allow integration of electronic and optical functionality in the main electronics platform technology, but this has been impossible due to the indirect band gap of Si. In this talk I will discuss 2 different approaches, using unique properties of nanowires, to realize light emission from Si-based compounds. In the first route we focus on the fabrication of defect-free GeSn compounds. GeSn has been shown to exhibit a direct band gap at Sn concentrations above 12.5% in the infrared part of the spectrum (around 0.5 eV) [1]. However, in bulk layers the strain between the Ge and the GeSn layer is released by the introduction of defects near the interface affecting the optical properties of the layer. In the nanowire geometry the lattice strain can be effectively relieved in the radial direction, which is exploited to grow Ge/GeSn core shell nanowires with high (13%) Sn content. In this talk the growth mechanism is discussed, the structural properties are investigated by Electron Microscopy and Atom Probe Tomography and the temperature dependent optical properties are studied. In the second route we concentrate on Si and Ge with a different crystal structure. It has been predicted that SiGe alloys with the hexagonal (2H) crystal structure have a direct band gap. It has been shown that by using the VLS nanowire growth mechanism it is possible to fabricate III-V semiconductors, which normally crystallize in the cubic phase, can now been grown with a 2H crystal structure [2]. Here, we employ crystal structure transfer, in which we use wurtzite GaP as a template to epitaxially grow SiGe compounds with the hexagonal crystal structure [2]. We show that with this method we can grow defect free hexagonal SiGe shells and branches with tunable Ge concentration. The structural and optical properties of these new crystal phases will be discussed. References [1] Assali, S. et al. DOI: 10.1021/acs.nanolett.6b04627, [2] Hauge, H.I.T. et al. DOI: 10.1021/acs.nanolett.5b01939.

Resume : Scanning electron microscope (SEM) induced nanowire (NW) attraction or bundling is a well-known effect, which is mainly ascribed to growth parameters or material dependent properties, such as electrostatic or material dependent dipole interaction during or after growth, e.g. by polar surfaces or van-der-Waals interaction. However, there have also been recent reports of electron beam induced nanowire bending by SEM or TEM imaging, which is not fully understood by the current models, especially when considering the electro-dynamic interaction between NWs. With this contribution we aim to help advance the understanding of this phenomenon, by introducing an electro-dynamic model based on capacitor and Lorentz force interaction, where the active NW bending is stimulated by an electron beam induced electromagnetic force between the individual wires. The model considers geometrical and material specific electrical and mechanical NW parameters, as well as the influence of the electron beam source parameters. Further, the NW bending or attraction on electrically conductive as well as insulating substrates is taken into account. While the model is validated using in-situ observations of electron beam induced GaAs NW bending by SEM imaging, in dependence of the NW exposure time to the electron beam as well as the NW separation distance, the model is fully applicable to group IV semiconductors such as Si. The result demonstrates an excellent agreement with our simulation.

Resume : Semiconductor quantum dots (QDs) are 0-dimensional structures where quantum confinement effects provide unique material properties, which can be exploited in photonic and memory applications. QDs embedded in a dielectric can be formed in different ways, either by ion implantation of Ge/Si in SiO2 [1-3] or by chemical vapor or molecular beam epitaxial (MBE) deposition techniques [4,5]. This is commonly followed by a heat treatment to form the nanoclusters in the dielectric medium. A concern is that during the thermal step, not only QDs are formed but possibly also defect structures, which can compromise the performance of the targeted devices. It was shown before by capacitance-based techniques that for example a high density of states can be formed at the Si/SiO2 interface [2,3]. It is the aim of the current work to report on a systematic Deep-Level Transient Spectroscopy (DLTS) study of QDs formed in SiO2 on n-type Czochralski silicon substrates by MBE, followed by a high-temperature vacuum anneal. Silicon QDs with average diameter of 7 and 14 nm and density of ~1011 cm-2 are compared with similar size Ge QDs. An aluminum metal gate has been thermally evaporated, resulting in a Metal-Oxide-Semiconductor (MOS) structure for electrical evaluation. Initial results on Ge QDs have been recently reported [5,6] and can be summarized as follows: the presence of Ge QD results in a marked change in the 1 MHz capacitance-voltage (C-V) characteristics. The positive flat-band voltage shift indicates the presence of negative charge, associated with the QDs. Increasing counterclockwise hysteresis has been found as well, suggesting significant electron trapping effects, associated with the dots. From DLTS, two types of electron traps have been derived: interfaces states, resembling the well-known dangling bond centers, with activation energy which changes with the Ge QD size. A second band in the spectrum, occurring around room temperature, may be more related with inversion-layer build-up. The latter can be also related to surface generation by interface states. A comparison with Si QDs helps to unravel the role of Ge in the formation of specific dangling bond states and in the change in the C-V characteristics. A first conclusion is that both for Ge and Si QDs, the described effects become more pronounced with their size and in a second instance, the impact of Ge on the deep-level spectra is higher than that of Si. A more detailed quantitative analysis will assist in a better interpretation in the observed phenomena and should lead to guidelines for process optimization. References [1] J. von Borany, R. Grötzschel, K.H. Heinig, A. Markwitz, W. Matz, B. Schmidt and W. Skorupa, Appl. Phys. Lett. 71, 3215 (1997). [2] E.S. Marstein, A.E. Gunnæs, A. Olsen, T.G. Finstad, R. Turan and U. Serincan, J. Appl. Phys. 96, 4308 (2004). [3] R. Beyer, J. von Borany and H. Burghardt, Microelectron Eng. 86, 1859 (2009). [4] J. Valenta, M. Greben, S. Gutsch, D. Hiller and M. Zacharias, J. Appl. Phys. 122, 144303 (2017). [5] M. Aouassa, H. Vrielinck and E. Simoen, ECS Trans. 80 (4), 181 (2017). [6] M. Aouassa, H. Vrielinck and E. Simoen, accepted by ECS J. Solid State Sci. Technol.

Resume : We report well-aligned p-Si/n-InP core-shell radial nanowire heterojunction array based solar cell. The p-Si nanowires were prepared at room temperature by metal assisted chemical etching of p type Si (100) wafer using silver nanoparticles whereas n-InP layer was deposited on p-Si nanowire templates using horizontal-type atmospheric pressure metalorganic chemical vapor deposition to obtain a core-shell radial heterojunction. A 100 nm transparent conductive oxide layer was deposited onto top of n-InP layer by sputtering. Transmission electron microscope images confirm the formation of Si/InP core-shell radial nanowire heterostructure. From the studies of reflectance spectroscopy, higher absorption of visible photons has been found. Current-voltage measurements on the radial core-shell nanowire heterojunction based solar cell have been taken under dark and an AM 1.5 solar radiation at room temperature. The device is found to provide a conversion efficiency of 4.39% with an open circuit voltage of 0.56 V and a short circuit current density 14.26 mA/cm^2 under AM 1.5 solar radiation. The core-shell radial heterojunction solar cell on nanowire arrays shows great improvement of the performance in comparison with conventional nanowire based solar cells. Our study provides new insights into the Si/InP core-shell nanowire based heterojunction which can have potential applications in fabricating nanoscale devices on Si platform for photovoltaic application.

Resume : We study the application of the Modulated PhotoCurrent (MPC) technique [1] to investigate crystalline silicon (c-Si) / hydrogenated amorphous silicon (a-Si:H) heterojunctions with the objective to characterize interface defects and band bending, and to determine the influence of a 2DEG in the strong inversion layer that was previously revealed from other techniques [2]. Modulated PhotoLuminescence (MPL) is typically used to determine effective lifetime values in the silicon wafer [3]. Here we built an original set-up to couple both techniques,so that we can compare the phase-shifts of MPC and MPL measured simultaneously and study the frequency dependence. This allows us to distinguish the lifetime of carriers at the interface from the one in the bulk, and thus to estimate the recombination rates at the c-Si/a-Si:H heterojunction. A substantial part of the work is devoted to 2D TCAD simulations for a better understanding of the behavior of the MPC and MPL measurements. The simulation results are compared with MPC and MPL experiments performed on a set of samples with an intrinsic (i)a-Si:H passivating layer at the (n)c-Si/(p)a-Si:H interface and with i-layer thicknesses varying in the range of 0-50nm creating different levels of inversion. [1] J.-P. Kleider, C. Longeaud, in Solid State Phenomena 44-46 (1995), pp. 597-646. [2] O. Maslova et al, Appl. Phys. Lett. 103, 183907 (2013) [3] R. Brüggemann and S. Reynolds, J.Non-Cryst. Solids 352, 1888–1891 (2006)

Resume : Scanning Spreading Resistance Microscopy (SSRM) is known as a powerful 2D carrier mapping technique due to its capability to provide quantitative analysis at high resolution. The sample preparation for SSRM involves cross sectioning to expose the device of interest, typically by cleaving or polishing. However, for confined geometries such as FinFETs and nanowires the use of Focused Ion Beam (FIB) milling is indispensable, for localizing or contacting purposes. This deliberate exposure of sample to Ga ion beam results in implantation of Ga ions in the sample which, in turn, modifies the electrical properties of the device being studied and, consequently, thwarts the precise quantification of carriers. Therefore, it is crucial to understand the electrical impact that FIB imparts to the Si. Here, we leverage the inherent destructive nature of SSRM to study FIB induced electrical damage along the depth. We discuss the impact of ion beam energy, ion dose and incidence angle on the SSRM measurements for doping concentrations ranging from 1E20 to 1E16 atoms/cm3. The measured depth profiles of resistance indicate that the actual damage is much deeper than what is expected from TRIM calculations and the electrically affected depth varies significantly with doping concentrations. Both underlying physical mechanism and potential way to mitigate the damage will be discussed. This work also highlights the importance of understanding the impact of FIB for characterization beyond SSRM.

Resume : Along the scaling path, the microelectronic industry faces numerous challenges in the growth of group IV semiconductors, such as Ge and GeSn. Especially Ge(1-x)Sn(x) is a promising candidate for source-drain applications and as a stressor in Ge channel devices. However, the material properties of scaled devices can strongly deviate from their bulk counterparts and therefore reliable characterization methods are needed. Measuring the electrical properties of real device structures is particularly important to verify the degree of activation and material quality, both during and after processing. The sheet resistance, sheet carrier density and mobility are well-suited parameters to provide that information. In this study, we demonstrate the capabilities of a micro four-point probe technique in taking advantage of thermoelectric properties for characterization of nm-wide semiconductor fins. This is done by performing a four-point probe measurement with an AC current providing us with the first and second harmonics, which are related to the sample resistance and Seebeck coefficient, respectively. From the Seebeck coefficient, we can determine the charge carrier density directly. Subsequently, the carrier mobility can be extracted using the resistance signal and the Seebeck coefficient combined. We will perform these measurements on fins of B-doped Ge(1-x)Sn(x) on relaxed Ge, which are epitaxially grown in trenches. In this way, the precision of the method can also be evaluated.

Resume : One of the key issues for the miniaturization of semiconductor nanodevices is the precise control of their doping. Today, the doping spatial distribution must be controlled with a precision higher than 1 nm while atomic concentrations below 1% have to be measured. This calls for the use of high resolution techniques such as scanning transmission electron microscopy (STEM) associated with Energy Dispersive X-ray spectroscopy (EDX). Measuring low concentrations with STEM/EDX requires a high signal-to-noise ratio. This is difficult to obtain since probe current normally decreases abruptly when small probe sizes are used. In addition, the quantification of the results represents one of the most critical steps of the analysis. Different physical phenomena occurring between the X-ray emission and the EDX detection must be taken into account since the artifacts and parasites can have a significant impact on the final result. Both issues must be addressed at the same time In this work, we address both issues to successfully apply STEM/EDX for the investigation of dopant mapping in advanced nanodevices. First, we use a cold FEG, Cs-corrected electron microscope (JEOL ARM) equipped with a high-collection angle EDX detector to obtain atomic resolution with high probe currents. Second, we have developed a dedicated algorithm to improve the reliability of the quantified spectra. Finally, our approach will be applied to the dopant mapping in FinFET and ultra-thin SOI/SiGeOI test structures.

Resume : The implant damage is a major challenge for the extension doping of narrow fins and gate all around devices. Hence extension-less devices are now considered as an alternative. However, they lead to other challenges (i.e. thinner spacers, gate overlap control). One of the alternative solutions is the use of the solid-source doping (SSD) concept. It has been demonstrated already that Phosphorus doped Silicate Glass (PSG) could lead to boosted performances into sub-10nm FF devices. Within this work, we have performed a systematic analysis of P and B dopant diffusion and activation in Si for PSG and BSG layers respectively. In particular, we have analyzed their vertical and lateral diffusion in confined volumes using SOI test-structures Combining various measurement techniques, we have demonstrated that confinement allows us to reach higher active concentration levels (above 1e20 cm-3) but presents lower activation rates. We attribute this to the presence of oxidation induced excess interstitials. Moreover we observe a reduced lateral diffusion, particularly in the case of B. We attribute this to the fact that during the annealing step, excess interstitials can diffuse towards the SOI sidewalls where they are annihilated reducing the lateral diffusion, in particular for B as boron-interstitial clusters (BICs) present a higher binding energy for interstitials. This annihilation effect is increased substantially for sub-10nm dimensions. Hence we believe that SSD is particularly adapted to confined volumes presenting boosted dopant concentration and reduced lateral diffusion. It should be evaluated as a potential solution to process extensions in N7 and N5 technology nodes.