Jean-Bernard Gros

A simplified monodimensional approach for modeling coupling between radiant light transfer and growth kinetics in photobioreactors

Abstract Local information is essential to photobioreactor modeling because of medium anisotropy in radiant light energy. Local available energy can be calculated using complex equations, applying the physical laws of radiative transfer and independently accounting for light absorption and scattering in the reactor. In this paper, these equations are simplified postulating monodimensional approximation for the radiation field. This simplification is established for rectangular, cylindrical and spherical coordinates, leading to simple analytical solutions for available radiant energy profiles inside the reactor. This approach provides a method of determining working illuminated volume defined as the photobioreactor volume with sufficient radiant light energy for microorganism growth. This enables the coupling between radiant light transfer and growth kinetics to be easily studied. Physical light transfer models are used to simulate volumetric biomass growth rates in a cylindrical photobioreactor with kinetic parameters obtained from batch cultures of the cyanobacterium Spirulina platensis in rectangular photoreactors. These calculations are compared with experimental data obtained on continuous cultures in a wide range of incident radiant energy fluxes. The model is found to have a good predictability and robustness.

Energy model and metabolic flux analysis for autotrophic nitrifiers

The behavior of pure cultures of nitrifying microorganisms under autotrophic growth operating conditions was investigated and the relations between their energy metabolism and their anabolism analyzed by means of metabolic network computation. The description of the metabolism of the nitrifiers is extended to their energy metabolism by introducing compartmentalization (cytoplasmic and periplasmic sides) and studying coupling between the electron transport chain and the proton gradient generation. The energy model of Nitrosomonas and Nitrobacter was developed based on the oxidoreduction reactions known to be involved. The electron transport chains and the associated proton translocation for these models are described. Several possible hypotheses are analyzed and discussed concerning the thermodynamic consistency of all the oxidoreduction reactions. For Nitrosomonas, the most delicate point is the second step of hydroxylamine oxidation. For Nitrobacter a new energy model is proposed in which NO plays an important role as node in the distribution of electrons from NO(2)(-) oxidation to the membrane electron transport chain. The compartmentalization enables us to consider a proton gradient dissipation flux as the expression of the overall energy loss in metabolic analysis (the so-called maintenance phenomena). The energy model (electron transport chain, proton gradient) is associated with an overall description of the metabolism of Nitrosomonas and Nitrobacter in terms of metabolic flux calculation. This representation demonstrates that a maintenance in nitrifiers expressed as a proton leak is no higher than for other aerobes. The yields calculated from the energy models integrated with the metabolic models of nitrifiers are consistent with the experimental yields in the literature.

HIGH PYRAZINE PRODUCTION BY BACILLUS SUBTILIS IN SOLID SUBSTRATE FERMENTATION ON GROUND SOYBEANS

2,5-Dimethylpyrazine (2,5-DMP) and tetramethylpyrazine (TTMP) were produced using Bacillus subtilis IFO 3013 grown in solid substrate conditions using ground soybeans suspended in water. Optimization studies showed that the best way to produce the two above aroma compounds involved massive enrichment of the medium with l-threonine and acetoin. The amino acid allowed 2,5-DMP formation and was added at 40 g/l at the beginning of a process, while acetoin was the precursor of TTMP and had to be added at 60 g/l only after 2,5-DMP production was terminated. The optimal cultivation temperature was 40°C, pH had to be monitored at 7.5, and aeration rate had to be higher than 0.1 VVM with a volumetric oxygen tranfer coefficient kLa close to 180 h−1. These conditions allowed recovery of 2 g/l total pyrazines, that demonstrated the efficiency of this approach.

Representation of vapour -liquid equilibria in water-alcohol-electrolyte mixtures with a modified UNIFAC group-contribution method

Abstract Achard, C., Dussap, C.G. and Gros J.B., 1994. Representation of vapour-liquid equilibria in water-alcohol-electrolyte mixtures with a modified UNIFAC group-contribution method. Fluid Phase Equilibria , 98: 71-89. A new method for the correlation and prediction of salt effects on vapour-liquid equilibria of mixed solvent systems is proposed. The model combines a term of the Debye-Huckel type with a modified UNIFAC equation and is based on the solvation concept. The model parameters are ion-specific and no ternary parameters are required. Group interaction parameters have been estimated between ions (Na + , K + , Li + , Mg 2+ , Ca 2+ , Co 2+ , Ba 2+ , Sr 2+ , NH + 4 , Cl − , Br − , F − , NO − 3 , SO 2− 4 ) and solvent groups (CH 3 OH, CH 2 , OH), while previously published group interaction parameters between solvent groups and between water and ionic species have been maintained. Application of this model to alcohol-water-salt mixtures shows that the change in vapour phase composition on adding salt to the liquid phase can be represented satisfactorily. The average expected accuracy of the total pressure or temperature is about 4% and of vapour phase mole fractions better than 4%. Since it is a group-contribution method, it has a much broader range of applicability than classical approaches based on local composition models such as NRTL or UNIQUAC

Pyrazine production by Bacillus subtilis in solid-state fermentation on soybeans

Abstract 2,5-Dimethylpyrazine (2,5-DMP) and tetramethylpyrazine (TTMP) were produced using Bacillus subtilis IFO 3013 grown on soybeans. Solid-state cultivations were carried out either in 100-ml bottles or in a fixed-bed column reactor, both systems being at 27 °C. Optimization studies showed that the best way to produce the two above aroma compounds involved two separate processes. 2,5-DMP was obtained using soybeans enriched with 75 g threonine/kg initial dry weight (i.d.w.), giving 0.85 g metabolite/kg i.d.w. after 6 days. TTMP production involved addition of 90 g/kg i.d.w. acetoin to soybeans, and 2.5 g/kg i.d.w. was recovered after 14 days. These results demonstrated the suitability of solid-state cultivation for production of high-added-value compounds.

Metabolic flux distribution in Corynebacterium melassecola ATCC 17965 for various carbon sources

The distribution of carbon in the metabolic network of a bacterial cell was estimated by a mass‐balance‐based intracellular flux computation method. It was applied to the growth phase of Corynebacterium melassecola, a glutamic acid producing bacterium, using experimental production yields of biomass, lactate and acetate measured during batch cultures on glucose, fructose, and various mixtures of both sugars. This flux computation method identifies the direction of the 86 reactions that ensure proper metabolic function during the growth phase of C. melassecola. Flux ratios allow comparison of calculated and relevant experimental yields. The results highlight the key influence of the biomass production yield Y  X−O 2 on the overall distribution of carbon; the proportion of carbon drained in the pentose‐P pathway fell from a value in the range of 54% to 47% on media containing glucose (Y  X−O 2 = 1.75 to 1.56 g X/g O2) to 37% on fructose medium (Y  X−O 2 = 1.36 g X/g O2). The highest maintenance requirement was calculated on fructose medium (Jm = 290 mol ATP/100 mol fructose) which must be connected to a lower efficiency of cell multiplication observed on this substrate. Another important result was that the significant decreases in experimental values of production yields and rates observed on fructose medium which were related to the operation of the FBPase. In particular, it was estimated that, as long as the proportion of glucose in the carbon source remains above 22% (78% fructose), the operation of the FBPase is not necessary and the bacteria exhibit behavior similar to that observed on glucose alone; this result is consistent with experimental observations.

A physical chemical UNIFAC model for aqueous solutions of sugars

Abstract A new physical chemical model is proposed to describe the thermodynamic properties of binary water-carbohydrate mixtures. The chemical part of the model takes into account the conformational equilibria of sugars and the solvation equilibria between water and sugars. The physical part of the model is a modified UNIFAC model where three new main groups (pyranose ring, furanose ring and osidic bond) are introduced to describe carbohydrate molecules. The coefficients of the model (six interaction parameters and one parameter describing the partial molar Gibbs energy of the hydrogen bond, all temperature dependent) are adjusted on seven binary water-sugar systems (glucose, mannose, galactose, fructose, sucrose, lactose, maltose). The experimental data include water activity, osmotic coefficients, activity coefficients, freezing and boiling temperatures, excess Gibbs energy and excess enthalpy. The model allows carbohydrate isomers to be distinguished and the excess Gibbs energy of systems not used in the coefficients estimation (xylose, raffinose) to be predicted.

Excess properties and solid-liquid equilibria for aqueous solutions of sugars using a UNIQUAC model

Abstract Catte M., Dussap C.-G., Achard C. and Gros J.-B., 1994. Excess properties and solid-liquid equilibria for aqueous solutions of sugars using a UNIQUAC model. Fluid Phase Equilibria, 96: 33-50. A modified UNIQUAC model was used to describe thermodynamic properties of binary water-carbohydrate mixtures. Interaction parameters are determined for glucose, fructose and sucrose. A new equation was developed to describe carbohydrate solubility in water from knowledge of dilution enthalpy rather than fusion enthalpy, and particularly when sugar crystallizes as a hydrated form. For all thermodynamic properties (water activity, osmotic coefficients, excess Gibbs energy, excess enthalpy, activity coefficients, boiling temperature, freezing temperature and solubility) the model agrees with experiments when reliable data are available. Good predictions are also obtained for water activity and osmotic coefficients in ternary systems (water-sucrose-glucose).

Measurement of effective diffusivities of ionic and non-ionic solutes through beef and pork muscles using a diffusion cell

A diffusion cell in conjunction with a new mathematical model was used to measure the apparent diffusion coefficients of sodium chloride, potassium nitrate and glucose through beef semitendinosus and pork longissimus dorsi muscles. The apparent diffusion coefficient was obtained within less than 180 min with a coefficient of variation less than 10% in most cases. The influence of temperature can be adequately explained by the Arrhenius-type law with activation energy between 19 and 26 kJ/mol depending on the diffusant. However, these values were not significantly different. Finally, we showed that the apparent diffusion coefficient of sodium chloride is not affected by the direction of the muscle fibre to the axes of diffusion while glucose diffusion parallel to the muscle fibre can be three times greater than diffusion at a right angle to the muscle fibre.

Kinetic analysis of cooking losses from beef and other animal muscles heated in a water bath — Effect of sample dimensions and prior freezing and ageing

Cooking loss kinetics were measured on cubes and parallelepipeds of beef Semimembranosus muscle ranging from 1 cm × 1 cm × 1 cm to 7 cm × 7 cm × 28 cm in size. The samples were water bath-heated at three different temperatures, i.e. 50°C, 70°C and 90°C, and for five different times. Temperatures were simulated to help interpret the results. Pre-freezing the sample, difference in ageing time, and in muscle fiber orientation had little influence on cooking losses. At longer treatment times, the effects of sample size disappeared and cooking losses depended only on the temperature. A selection of the tests was repeated on four other beef muscles and on veal, horse and lamb Semimembranosus muscle. Kinetics followed similar curves in all cases but resulted in different final water contents. The shape of the kinetics curves suggests first-order kinetics.