Effect of indirect dependencies on maximum likelihood and information theoretic blind source separation for nonlinear mixtures

Two major approaches for blind source separation (BSS) are, respectively, based on the maximum likelihood (ML) principle and mutual information (MI) minimization. They have been mainly studied for simple linear mixtures. We here show that they additionally involve indirect functional dependencies for general nonlinear mixtures. Moreover, the notations commonly employed by the BSS community in calculations performed for these methods may become misleading when using them for nonlinear mixtures, due to the above-mentioned dependencies. In this paper, we first explain this phenomenon for arbitrary nonlinear mixing models. We then accordingly correct two previously published methods for specific nonlinear mixtures, where indirect dependencies were mistakenly ignored. This paper therefore opens the way to the application of the ML and MI BSS methods to many specific mixing models, by providing general tools to address such mixtures and explicitly showing how to apply these tools to practical cases.     

Multicolor up conversion emission and color tunability in Yb/Tm/Ho triply doped heavy metal oxide glasses

Multicolor and white light emissions have been achieved in Yb³⁺, Tm³⁺ and Ho³⁺ triply doped heavy metal oxide glasses upon laser excitation at 980 nm. The red (660 nm), green (547 nm) and blue (478 nm) up conversion emissions of the rare earth (RE) ions triply doped TeO₂–GeO₂–Bi₂O₃–K₂O glass (TGBK) have been investigated as a function of the RE concentration and excitation power of the 980 nm laser diode. The most appropriate combination of RE in the TGBK glass host (1.6 wt% Yb₂O₃, 0.6 wt% Tm₂O₃ and 0.1 wt% Ho₂O₃) has been determined with the purpose to tune the primary colors (RGB) respective emissions and generate white light emission by varying the pump power. The involved infrared to visible up conversion mechanisms mainly consist in a three-photon blue up conversion of Tm³⁺ ions and a two-photon green and red up conversions of Ho³⁺ ions. The resulting multicolor emissions have been described according to the CIE-1931 standards.     

Mechanisms of scale formation and carbon dioxide partial pressure influence. Part I. Elaboration of an experimental method and a scaling model

Scale formation in industrial or domestic installations is still an important economic problem. The existence of a metastable domain for calcium carbonate supersaturated solutions and its breakdown are observed under conditions rarely well defined. In most cases it is the pH rise caused by the carbon dioxide loss that involves calcium carbonate precipitation. Before studying this problem, we suggest in this first part, a new model for the evolution of the calcocarbonic system that takes into account the hydrated forms of CaCO3: CaCO3 amorphous, CaCO3 x 6H2O (ikaite) and CaCO3 x H2O (monohydrate). According to this model, the precipitation of any one of these hydrated forms could be responsible for the breakdown of the metastable state. After this first step, the solids evolve into dehydrated forms. At first, the metastable domain spread of the calcium carbonate supersaturated solutions was studied by the elaboration of computer programs in which the formation of CaCO3(0)(aq) ion pairs was taken into account. These ion pairs are supposed to evolve through dehydration to form the various calcium carbonate solid form precursors. This thermodynamic study was then compared to the experimental methods of the critical pH. Here the pH rise was caused by adding sodium hydroxide under different conditions for sodium hydroxide addition speed, agitation mode and ageing of solutions. For the highest speed of sodium hydroxide addition, the CaCO3 ionic product reached the value of the amorphous calcium carbonate solubility product, and the reaction of the amorphous calcium carbonate precipitation was of the homogenous type. Decreasing the reagent's addition speed caused an extension of the titration time. Then, the breakdown of the metastable state was obtained with the CaCO3 x H2O heterogeneous precipitation. This clearly illustrates the probable ageing of the precursors of the solid states that are considered in this model.     

Prevention of volatile fatty acids production and limitation of odours from winery wastewaters by denitrification

The effect of the addition of nitrate to winery wastewaters to control the formation of VFA in order to prevent odours during storage and treatment was studied in batch bioreactors at different NO(3)/chemical oxygen demand (COD) ratios and at full scale in natural evaporation ponds (2 x 7000 m(2)) by measuring olfactory intensity. In the absence of nitrate, butyric acid (2304 mgL(-1)), acetic acid (1633 mgL(-1)), propionic acid (1558 mgL(-1)), caproic acid (499 mgL(-1)) and valeric acid (298 mgL(-1)) were produced from reconstituted winery wastewater. For a ratio of NO(3)/COD=0.4 gg(-1), caproic and valeric acids were not formed. The production of butyric and propionic acids was reduced by 93.3% and 72.5%, respectively, at a ratio of NO(3)/COD=0.8, and by 97.4% and 100% at a ratio of NO(3)/COD=1.2 gg(-1). Nitrate delayed and decreased butyric acid formation in relation to the oxidoreduction potential. Studies in ponds showed that the addition of concentrated calcium nitrate (NITCAL) to winery wastewaters (3526 m(3)) in a ratio of NO(3)/COD=0.8 inhibited VFA production, with COD elimination (94%) and total nitrate degradation, and no final nitrite accumulation. On the contrary, in ponds not treated with nitrate, malodorous VFA (from propionic to heptano?c acids) represented up to 60% of the COD. Olfactory intensity measurements in relation to the butanol scale of VFA solutions and the ponds revealed the pervasive role of VFA in the odour of the untreated pond as well as the clear decrease in the intensity and not unpleasant odour of the winery wastewater pond enriched in nitrates. The results obtained at full scale underscored the feasibility and safety of the calcium nitrate treatment as opposed to concentrated nitric acid.     

Solar‐grade boron emitters by BF3 plasma doping and role of the co‐implanted fluorine

We investigate the electrical properties and dopant profiles of boron emitters performed by plasma immersion ion implantation from boron trifluoride (BF₃) gas precursor, thermally annealed and passivated by silicon oxide/silicon nitride stacks. High thermal budgets are required for doses compatible with screen‐printed metal pastes, to reach very good activation rates. However, if good sheet resistances and saturation current densities may be obtained, we met strong limitations of the implied open‐circuit voltage of the n‐type Czochralski silicon substrates, which is incompatible with high‐efficiency solar cells. Such limitations are not encountered with beamline where pure B⁺ ions are implanted. Efforts on the passivation quality may improve the implied open‐circuit voltage but are not sufficient. We provide experimental comparison between beamline and plasma immersion allowing us to discriminate the causes explaining this observation (implantation technique or ion specie used) and to infer our interpretation: The co‐implantation of fluorine seems to indirectly impact the lifetime of the core substrate after thermal annealing. Copyright © 2015 John Wiley & Sons, Ltd.