Master of Engineering / Spray dryers are the core components of a milk powder production plant, where the basic configuration usually features co-current flow of milk powder and air. Spray dryers have to be cleaned frequently due to powder deposit build-up on the walls. Powder deposit build-up gives rise to lower product yields and poses a potential fire risk. If the powder deposits are scorched (from being overheated) they will contaminate, and thus compromise, the quality and consumer safety of the final product, if the powder deposits fall in and mix with it. With milk powder production rates of most industrial spray dryers ranging from 4-28 tonnes of dry powder an hour, these wall deposition problems are significant. This problem is worth investigating because the outcome of reducing or eliminating wall deposition is that a spray dryer could operate for a longer period of time without having to be cleaned. Reduction in downtime due to cleaning would give rise to increased production time and possibly a reduction in the cost of manufacturing the product. The spray dryer used in this work was a modified short-form co-current Niro unit, fabricated from stainless steel. The spray dryer had an internal diameter of 0.80 m, narrowing down to 0.06 m at the base, and a height of 2 m. A two-fluid nozzle was used to spray the process fluids (water, skim milk and grape skin extract) into the drying chamber. To measure the wall deposition fluxes on the internal walls of the spray dryer, four stainless steel plates (dimensions 110 mm by 120 mm) were inserted in place of the windows that were previously used as sight glasses. A fifth plate (dimensions 110 mm by 120 mm) and a sixth plate (dimensions 110 mm by 110 mm) were also placed on the conical section of the spray dryer at different circumferential locations. Before this work, no quantitative data on the effects of spray dryer operating conditions on the wall deposition fluxes of food material were available. This work investigated the effect on the spray deposition flux of skim milk powder on the walls of the spray dryer of (i) flow patterns in the spray dryer, by changing the degree of swirl imparted to the incoming air by using three swirl vane angles of 0o, 25o and 30o, and (ii) the stickiness of the product, through first changing the temperature of the incoming air by using three inlet air temperatures of 170oC, 200oC and 230oC; and then changing the process fluid flowrate by using three flowrates of 1.4 kg hr-1, 1.6 kg hr-1 and 1.8 kg hr-1. Previous researchers have found that the extent to which water droplets spread out in the drying chamber is affected by the amount of swirl in the inlet air. This is likely to affect wall deposition fluxes because the particles will be closer to the walls if the droplets spread out widely. The results of this work have quantitatively confirmed that the spray deposition flux increases at higher swirl vane angles, where the spray deposition flux increased from 7 g m-2 hr-1 (swirl vane angle 0o) to 12.9 g m-2 hr-1 (swirl vane angle 30o). When a swirl vane angle of 0o was used, it was observed that the cross-sectional area of the spray cloud did not change very significantly with time. However, when a swirl vane angle of 25o was used, the spray cloud was observed to “flutter”, and when the swirl vane angle was increased to 30o, the spray cloud was observed to recirculate rapidly back in the direction of the nozzle. Thus, the chance of the particles being thrown further towards the walls of the chamber is likely to increase at higher swirl vane angles. This result suggests that higher wall deposition arises because more swirl is imparted to the air entering the dryer, which in turn affects the stability of the spray cloud and, therefore, the stability of the flow patterns in the spray dryer. The stickiness of the skim milk powder is related to the temperature and moisture content of the particles. In the past, the sticky-point curve has been suggested as a semi-quantitative concept in selecting operating conditions for spray drying food material containing carbohydrates, where it has been implied that there is no significant wall deposition below the sticky-point curve. This work has quantified the spray deposition in spray dryers with respect to the sticky-point curve, where the highest spray deposition flux of skim milk powder on the walls was 16 g m-2 hr-1, and the operating point corresponding to this spray deposition flux was located at and above the sticky-point curve. Hence, both particle stickiness and flow patterns affect the wall deposition of particles in a spray dryer. This work also investigated the effect of wall properties, namely a non-stick food grade material (nylon), adhesive tape and stainless steel, on the spray deposition flux of skim milk powder on the walls. The effect of electrostatics on wall deposition was studied by grounding the spray dryer and an anti-static agent was added to the skim milk to investigate if altering the properties of the feed material could reduce wall deposition. This work has quantitatively confirmed that cohesion occurs at the same rate as adhesion for skim milk powder in this spray dryer, because firstly, decreasing the adhesion tendency of the v wall by using nylon coating had no significant effect on the spray deposition flux compared with a smooth stainless steel wall and a wall covered with a double-sided adhesive tape; and secondly the powder collected on the walls was a linear function of time with and without adhesive on the plates. Furthermore, using a nylon coated wall did not eliminate wall deposition, and the wall deposition flux was found to be the same as when a stainless steel wall was used. This result further supports the finding here that spray deposition on the walls for skim milk powder is controlled by cohesion rather than adhesion. The spray dryer operating parameters that gave rise to the least spray deposition flux on the walls were a swirl vane angle of 0o, an inlet air temperature of 230oC and a process fluid flowrate of 1.4 kg hr-1. Decreasing the feed flowrate from 1.8 kg hr-1 to 1.4 kg hr-1 (decrease by 24%), with the inlet air temperature and swirl vane angle held constant, decreased the wall deposition flux by 43% from 7 g m-2 hr-1 to 4 g m-2 hr-1. Since the spray deposition flux on the walls decreased by 43% when the feed flowrate was decreased by 24%, it might be considered that the production process is in favour of a decrease in the feed flowrate to 1.4 kg hr-1 in this dryer, and consequently a decrease in the spray deposition flux on the walls per unit production output. Finally, this work investigated if the outlet moisture content from this small spray dryer used here was equilibrium limited or controlled by drying kinetics. The findings in this work confirmed the product moisture locus concept, which implies that the outlet moisture content of the skim milk particles approaches the equilibrium moisture content (in equilibrium with the outlet gas), and that the outlet moisture content of spray-dried food material containing carbohydrates is probably not limited by particle drying kinetics, even though the spray dryer is smaller (diameter 0.8 m, height 2 m) than those used in the dairy industry, typically with a diameter of 30 m and a height of 10 m.

Master of Engineering / Spray dryers are the core components of a milk powder production plant, where the basic configuration usually features co-current flow of milk powder and air. Spray dryers have to be cleaned frequently due to powder deposit build-up on the walls. Powder deposit build-up gives rise to lower product yields and poses a potential fire risk. If the powder deposits are scorched (from being overheated) they will contaminate, and thus compromise, the quality and consumer safety of the final product, if the powder deposits fall in and mix with it. With milk powder production rates of most industrial spray dryers ranging from 4-28 tonnes of dry powder an hour, these wall deposition problems are significant. This problem is worth investigating because the outcome of reducing or eliminating wall deposition is that a spray dryer could operate for a longer period of time without having to be cleaned. Reduction in downtime due to cleaning would give rise to increased production time and possibly a reduction in the cost of manufacturing the product. The spray dryer used in this work was a modified short-form co-current Niro unit, fabricated from stainless steel. The spray dryer had an internal diameter of 0.80 m, narrowing down to 0.06 m at the base, and a height of 2 m. A two-fluid nozzle was used to spray the process fluids (water, skim milk and grape skin extract) into the drying chamber. To measure the wall deposition fluxes on the internal walls of the spray dryer, four stainless steel plates (dimensions 110 mm by 120 mm) were inserted in place of the windows that were previously used as sight glasses. A fifth plate (dimensions 110 mm by 120 mm) and a sixth plate (dimensions 110 mm by 110 mm) were also placed on the conical section of the spray dryer at different circumferential locations. Before this work, no quantitative data on the effects of spray dryer operating conditions on the wall deposition fluxes of food material were available. This work investigated the effect on the spray deposition flux of skim milk powder on the walls of the spray dryer of (i) flow patterns in the spray dryer, by changing the degree of swirl imparted to the incoming air by using three swirl vane angles of 0o, 25o and 30o, and (ii) the stickiness of the product, through first changing the temperature of the incoming air by using three inlet air temperatures of 170oC, 200oC and 230oC; and then changing the process fluid flowrate by using three flowrates of 1.4 kg hr-1, 1.6 kg hr-1 and 1.8 kg hr-1. Previous researchers have found that the extent to which water droplets spread out in the drying chamber is affected by the amount of swirl in the inlet air. This is likely to affect wall deposition fluxes because the particles will be closer to the walls if the droplets spread out widely. The results of this work have quantitatively confirmed that the spray deposition flux increases at higher swirl vane angles, where the spray deposition flux increased from 7 g m-2 hr-1 (swirl vane angle 0o) to 12.9 g m-2 hr-1 (swirl vane angle 30o). When a swirl vane angle of 0o was used, it was observed that the cross-sectional area of the spray cloud did not change very significantly with time. However, when a swirl vane angle of 25o was used, the spray cloud was observed to “flutter”, and when the swirl vane angle was increased to 30o, the spray cloud was observed to recirculate rapidly back in the direction of the nozzle. Thus, the chance of the particles being thrown further towards the walls of the chamber is likely to increase at higher swirl vane angles. This result suggests that higher wall deposition arises because more swirl is imparted to the air entering the dryer, which in turn affects the stability of the spray cloud and, therefore, the stability of the flow patterns in the spray dryer. The stickiness of the skim milk powder is related to the temperature and moisture content of the particles. In the past, the sticky-point curve has been suggested as a semi-quantitative concept in selecting operating conditions for spray drying food material containing carbohydrates, where it has been implied that there is no significant wall deposition below the sticky-point curve. This work has quantified the spray deposition in spray dryers with respect to the sticky-point curve, where the highest spray deposition flux of skim milk powder on the walls was 16 g m-2 hr-1, and the operating point corresponding to this spray deposition flux was located at and above the sticky-point curve. Hence, both particle stickiness and flow patterns affect the wall deposition of particles in a spray dryer. This work also investigated the effect of wall properties, namely a non-stick food grade material (nylon), adhesive tape and stainless steel, on the spray deposition flux of skim milk powder on the walls. The effect of electrostatics on wall deposition was studied by grounding the spray dryer and an anti-static agent was added to the skim milk to investigate if altering the properties of the feed material could reduce wall deposition. This work has quantitatively confirmed that cohesion occurs at the same rate as adhesion for skim milk powder in this spray dryer, because firstly, decreasing the adhesion tendency of the v wall by using nylon coating had no significant effect on the spray deposition flux compared with a smooth stainless steel wall and a wall covered with a double-sided adhesive tape; and secondly the powder collected on the walls was a linear function of time with and without adhesive on the plates. Furthermore, using a nylon coated wall did not eliminate wall deposition, and the wall deposition flux was found to be the same as when a stainless steel wall was used. This result further supports the finding here that spray deposition on the walls for skim milk powder is controlled by cohesion rather than adhesion. The spray dryer operating parameters that gave rise to the least spray deposition flux on the walls were a swirl vane angle of 0o, an inlet air temperature of 230oC and a process fluid flowrate of 1.4 kg hr-1. Decreasing the feed flowrate from 1.8 kg hr-1 to 1.4 kg hr-1 (decrease by 24%), with the inlet air temperature and swirl vane angle held constant, decreased the wall deposition flux by 43% from 7 g m-2 hr-1 to 4 g m-2 hr-1. Since the spray deposition flux on the walls decreased by 43% when the feed flowrate was decreased by 24%, it might be considered that the production process is in favour of a decrease in the feed flowrate to 1.4 kg hr-1 in this dryer, and consequently a decrease in the spray deposition flux on the walls per unit production output. Finally, this work investigated if the outlet moisture content from this small spray dryer used here was equilibrium limited or controlled by drying kinetics. The findings in this work confirmed the product moisture locus concept, which implies that the outlet moisture content of the skim milk particles approaches the equilibrium moisture content (in equilibrium with the outlet gas), and that the outlet moisture content of spray-dried food material containing carbohydrates is probably not limited by particle drying kinetics, even though the spray dryer is smaller (diameter 0.8 m, height 2 m) than those used in the dairy industry, typically with a diameter of 30 m and a height of 10 m.

L’application aux procédés agroalimentaires des approches d'écoconception combinant modélisation et optimisation multiobjectif est freinée par un manque de modèles de procédé intégrant les caractéristiques du produit. Ce travail consiste à développer un cadre méthodologique d'écoconception de procédés agroalimentaires combinant modélisation et optimisation, avec pour support l'exemple du procédé d'évaporation du lait. Le procédé est modélisé à l’aide d’un simulateur de procédés commercial, auquel sont intégrés des modèles de propriétés du produit et de coefficient d'échange sélectionnés dans la littérature. Le nettoyage est pris en compte à l’aide d’un outil de calcul des inventaires d'une séquence de nettoyage en place. Des critères économiques et environnementaux sont calculés pour analyser les solutions de conception du procédé.L'optimisation multiobjectif est réalisée à l'aide d'un algorithme génétique, et une méthode d'aide à la décision permet d'identifier les meilleures solutions de compromis. Les potentialités du cadre méthodologique sont illustrées dans trois études d’écoconception. Les perspectives de ce travail portent sur l’écoconception de la production de poudre de lait incluant des opérations à membranes et du séchage, l’utilisation de méthodes d’intégration énergétique, et la prise en compte de la qualité du produit.L'optimisation multiobjectif est réalisée à l'aide d'un algorithme génétique, et une méthode d'aide à la décision permet d'identifier les meilleures solutions de compromis. Les potentialités du cadre méthodologique sont illustrées dans tro / The application of eco-design approaches to food processes is yet hampered by a lack of process models which incorporate product attributes. This thesis consists in developing a methodological framework for the eco-design of food processes, which combines simulation and multiobjective optimisation. This development is supported by the example of the evaporation of milk. The process is modelled with a commercial process simulator, to which property models of the product and heat transfer coefficient models, selected from the literature, are integrated. Cleaning is taken into account with a simplified tool for the calculation of the inventories of a cleaning-in-place sequence.Economic and environmental criteria are computed, so as to analyse the different design solutions. Multiobjective optimisation is carried out with a genetic algorithm, and a multicriteria decision-making method identifies the solutions which offer the best compromise. The potential of the framework is demonstrated through three eco-design studies. This work paves the way for the eco-design of the milk powder production including membrane operations and drying, the use of energy integration methods, and the integration of product quality.

Bovine milk fat globules naturally vary from less than 0.2 µm to 15 µm in diameter. Milk has at least two distinct distributions of fat globules. While the majority (~90%) of globules in milk are of the smaller distribution (average diameter of 0.4 µm), virtually all the fat is carried in the larger globules (average diameter 3.5 µm). This distribution suggests some compositional and/or functional significance might exist between the two populations of fat globules, which may be related to origin of these globules in the lactating cell. Milk fat globules have a unique structure, composed of a core droplet of non polar lipids (triacylglycerol) surrounded by a lipid bilayer membrane known as milk fat globule membrane (MFGM). Other than MFGM, there is another source of membrane that has been identified in skim milk. It has been hypothesized that this skim milk membrane (SMM) is derived from MFGM, but little data are available to support this idea, and the membrane may also have alternate origins. In this study, different aggregates of lipids (small and large fat globules, SMM, skim milk) from milk were isolated and characterized for their lipid contents. Isolation of small and large fat globules fractions was verified by laser diffraction particle size analysis. The lipids were extracted from isolated different lipid aggregates and individual classes were separated using thin layer chromatography. Lipids were transesterified to fatty acid methyl esters and analyzed by gas chromatography-mass spectrometry. The results indicate that there are some compositional differences between native milk fat globule membranes of different sizes. For example, the total phospholipid fraction of small fat globules (SFG) contained significantly more unsaturated C18:1n9 and C18:2n6 than large fat globules (LFG). Conversely, sphingomyelin composition of SFG contained less C18:1n9 and C18:2n6cc, but more long chain fatty acids C22:0, C23:0, and C24:0. Phosphatidylethanolamine composition of SMM contained more C17:1 than SFG and LFG. The composition of C18:1n9 in triacylglycerol increased with fat globule size. Clear differences were also found in lipid profile of SMM and small and large fat globules from milk. Composition differences between SMM and native milk fat globules of different sizes suggest that origin of this membrane material in skim milk might have some different source than that of MFGM.

Pasteurized skimmilk at 4°C was acidified to pH 5.8 with 85.5% phosphoric acid (136g H3Po4;100 kg skimmilk), then warmed to 54°C and ultrafiltered to a protein concentration 9.1 ± 0.2%. The retentate was heated to 76.5°C for 16 s then cooled to 2°C. Phosphoric acid (85.5%) was added at a rate of 3.41g per kg retentate. The acidified retentate was slowly warmed to 29.5 °C (3°C/5 min) when the pH was checked. The pH at this point was no lower than 5.4. Heating was continued until a temperature of 32.2°C was reached. Glucono delta lactone was added to the retentate (17.6 g/kg retentate) and left undisturbed for approximately 80 min. The curd was cut at pH 4.7 with 0.64 cm curd knives and allowed 10 min for syneresis. Permeate obtained from the same lot of milk was acidified to pH 4.8 (66 g H3Po4;100 kg permeate), then added to the curd at 32.2°C (three parts permeate to four parts retentate) and used as a cooking vehicle. The curd was cooked to 59°C in 90 min. The curd was held at 59°C for 10 min, drained and washed once with ice water. Cream dressing containing 12.5% fat and 3% salt was used at the rate of two parts curd to one part dressing. Control cottage cheese was produced by a direct acid method from the same skimmilk used to produce ultrafiltered curd. Use of ultrafiltered skimmilk retentate for cottage cheese making resulted in 2.24% more curd (corrected to 20% solids) and 2.24% more curd per kg original milk protein than the control. However, satisfactory firmness in UF curd required slightly more than 20% solids in the final product. Sensory evaluations indicated that creamed cottage cheese was not significantly different (p

Skim milk was ultrafiltered to three lactose/protein ratios and spray dried. The skim milk powder was extruded with pearled corn starch at different moisture contents, protein contents, lactose/protein ratios, and feed rates (control variables). Response surface methodology and a central composite in cube experimental design were used. This design required 30 experimental runs with 16 factorial points, 8 axial points, and 6 center points for replication. The physical and functional properties evaluated were expansion ratio, product temperature, bulk density, color, shear stress, viscosity, and water absorption index (response variables). Scanning electron microscopy was done to evaluate the microstructural attributes of the extrudates. A quadratic model was used to express the response variables in terms of the control variables. Response surfaces were generated by assigning center point values to 2 of the 4 control variables and then solving the fitted equations as a quadratic in the remaining 2 control variables. An increase in moisture content decreased expansion ratio, product temperature, color, and water absorption index, and increased bulk density, shear stress, and viscosity. An increase in protein content decreased product temperature, shear stress, viscosity, and water absorption index, increased color, and had no effect on expansion ratio and bulk density. An increase in lactose/protein ratio decreased product temperature, viscosity, and water absorption index, and had no effect on expansion ratio, bulk density, color, and shear stress. Feed rate did not have significant individual effect on any response variable. Expansion ratio had a negative correlation with bulk density (r = -0.61) and shear stress (r = -0.62) and a positive correlation with product temperature (r = 0.52). Bulk density and shear stress were positively correlated (r = 0.69), and product temperature and water absorption index were positively correlated too (r = 0.81). Expansion ratio, bulk density, color, and shear stress were chosen to determine a combined set of extrusion conditions most likely to produce an extrudate with properties suitable for a snack-type product.

The effect of ultra-high temperature processing by direct steam injection and room temperature storage of pH-adjusted and unadjusted 3X (volume reduction) skim milk retentate and the effect of storage at various temperatures of 3X skim milk retentate without pH adjustment on their ζ-potential, viscosity, and pH were determined. Pasteurized skim milk was concentrated to 3X by ultrafiltration. In the pH study portions of skim milk retentate were adjusted to pH 6.38 ± .02 with HCl and 6.85 ± .01 and 7.32 ± .02 with NaOH between ultrafiltration and ultra-high temperature processing. In the storage temperature study, storage temperatures used for pH-unadjusted retentate samples were 11°C, 23°C (room temperature), and 37°C. Although pH 6.38 samples had the lowest viscosity in the pH study before ultra-high temperature processing, these samples precipitated during ultra-high temperature processing. For non-acidified samples, increasing pH of retentate resulted in higher viscosities and quicker age gelation times. Destabilization occurred more rapidly at 37°C than at 23°C or 11°C. The pH drop tended to be greater for samples stored at a higher temperature or adjusted to a higher pH. During 28 weeks of 37°C storage the pH decreased from 6.54 to 6.06. During 32 weeks of 23°C storage pH of samples initially adjusted to pH 7.32 dropped to 7.06. ζ-Potentials of casein micelles were calculated from electrophoretic mobility obtained by measuring Doppler frequency shifts of scattered laser light in samples that had been diluted 300 fold with their own ultra.filtrate. Absolute values of ζ-potential of samples stored at 37°C decreased from -23.4 millivolts immediately after ultra-high temperature processing to -18.5 millivolts after 28 weeks of storage. For samples stored at 11°C and 23°C in the storage temperature study and control samples in the pH study, absolute values of ζ-potential decreased approximately 1.5 to 2.0 millivolts during 28 or 32 weeks of storage.

Master of Science

Orientador: Mirna Lúcia Gigante / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia de Alimentos / Made available in DSpace on 2018-08-22T11:26:38Z (GMT). No. of bitstreams: 1 Maciel_GuilhermedeMoura_M.pdf: 1635237 bytes, checksum: 8aa0d57dc3b96eae51dcddcb4e837442 (MD5) Previous issue date: 2013 / Resumo: O objetivo deste trabalho foi avaliar o efeito da matriz encapsulante no rendimento de encapsulação do Lactobacillus acidophilus, na sua viabilidade durante o armazenamento a 4 ± 1°C e 25 ± 1°C por 90 dias, e na sua resistência à simulação gastrointestinal in vitro. As micropartículas foram produzidas a partir de soluções de soro doce e leite desnatado reconstituídos (30% sólidos totais), adicionadas de suspensão de L. acidophilus (1% v/v), e submetido a spray-drying a temperatura de entrada e saída de 180°C e 85-95°C, respectivamente. Amostras de 5g de micropartículas foram acondicionadas em embalagens de poliéster metalizado e polietileno seladas a vácuo, e armazenadas em dessecadores a 4 ± 1°C e 25 ± 1°C. As micropartículas produzidas foram avaliadas quanto às umidade, atividade de água, diâmetro médio e rendimento de encapsulação do L. acidophilus no tempo zero, e quanto a viabilidade e atividade de água após 7, 15, 30, 45, 60 e 90 dias de armazenamento. A sobrevivência do probiótico livre e microencapsulado foi avaliada durante a exposição às condições gastrointestinais in vitro (pH 2,0 e pH 7,0). Os experimentos foram repetidos 3 vezes e os dados analisados por ANOVA e teste de Tukey para comparação entre as médias (p < 0,05). A matriz encapsulante não afetou significativamente o rendimento de encapsulação, o diâmetro médio, a umidade e a atividade de água das partículas, que foi em média de 81,1 ± 5,0%, 12,94 ± 0,1?m, 4,53 ± 0,4% e 0,18 ± 0,04, respectivamente. A viabilidade do L. acidophilus nas microcápsulas reduziu em média 0,43 log UFC/g ao final dos 90 dias de armazenamento, e apresentaram contagens superiores a 106 UFC/g ao final deste período. Ambas as micropartículas se mostraram eficientes na proteção do probiótico à simulação gastrointestinal, sendo que as micropartículas de leite favoreceram o aumento da viabilidade do L. acidophilus / Abstract: The aim of this study was to evaluate the effect of the encapsulating matrix in the encapsulation yield, resistance to in vitro gastrointestinal simulation and viability of Lactobacillus acidophilus La-5 during storage at 4 ± 1 °C and 25 ± 1 °C for 90 days. The microparticles were produced from solutions of reconstituted sweet whey and skim milk (30% total solids), inoculated with a suspension of L. acidophilus (1% v / v) and subjected to spray-drying at inlet and outlet temperature of 180 °C and 85-95 °C, respectively. Five gram samples of microparticles were packed in metalized polyester and polyethylene packing, vacuum sealed, and stored in desiccators at 4 ± 1 ° C and 25 ± 1 ° C. The microparticles produced were evaluated for moisture, water activity, average diameter and encapsulation yield of L. acidophilus at time zero, and the viability and water activity after 7, 15, 30, 45, 60 and 90 days of storage. The survival of the free and microencapsulated probiotic was evaluated during exposure to in vitro gastrointestinal conditions (pH 2.0 and pH 7.0). The experiments were repeated 3 times and the data were analyzed by ANOVA and Tukey's test for the comparison between means (p <0.05). The encapsulating matrix did not significantly affect the encapsulation yield, the average diameter, the moisture and water activity of the particles, which were on average 81.1 ± 5.0%, 12.94 ± 0.1 m, 4.53 ± 0.4% and 0.18 ± 0.04, respectively. The viability of the microencapsulated L. acidophilus reduced on average 0.43 log CFU/g at the end of 90 days of storage, and remained higher than 106 CFU/g at the end of this period. Both microparticles were effective in protecting the probiotic during gastrointestinal simulation, and the skim milk microparticles favored an increasing in the viability of L. acidophilus / Mestrado / Tecnologia de Alimentos / Mestre em Tecnologia de Alimentos

Existe um número muito limitado de estudos referentes à produção de componentes antimicrobianos ou bacteriocinas produzidas por espécies de bifidobactérias. Nesse âmbito, o objetivo deste trabalho foi avaliar a produção de bifidobacteriocina em leite desnatado (LD). Para tanto, o estudo foi dividido em três etapas. A primeira etapa constituiu na preparação dos meios de cultura Man, Rogosa e Sharpe (MRS), Bifidus Selective Medium (BSM) e LD suplementado com 1% (p/v) de Tween 80 (T80), Inulina (I) ou Extrato de levedura (YE). Nesta etapa, os processos fermentativos foram conduzidos em shaker, nas condições: 50 rpm/37ºC/48h. Foram realizadas análises de pH, concentração de açúcares e ácidos, crescimento celular e determinação da atividade da bifidobacteriocina pelo método de difusão em ágar contra L. monocytogenes. Na segunda etapa, e baseado nos resultados obtidos, foi desenhado um delineamento composto central (CCD) construído a partir dos seguintes parâmetros: temperatura (34, 37, 40 °C) e concentração de YE (0,5; 1,0; 1,5 g/L). Na terceira etapa do trabalho, foram realizados os cultivos em biorreator de 2 L, contendo 10% de leite desnatado, nas seguintes condições: 200 rpm, 36°C, 2,0 g/L de YE, 48h de incubação em anaerobiose. Obteve-se em LD suplementado com YE (1%), combinado ao método de difusão em placa modificado (prévia refrigeração das placas por 12h), contra L. monocytogenes (2130 AU/mL), com uma fase exponencial de 24h, &#181;m de 0,604/h. A otimização feita através do CCD permitiu atingir níveis de atividade de 3.000 AU/mL a 3.100 AU/mL (ensaios 7, 11 e 14, blocos 3 e 1) contra L. monocytogenes, em condições ótimas de crescimento de YE: 2,0 g/L1 e T°C: 36°C. A análise de regressão mostrou ser estatisticamente significativa a relação entre as variáveis: \"concentração de \"YE e temperatura\". Os resultados indicaram que o leite desnatado é um meio adequado para produção de bifidobacteriocina. / There are few publications that have been reported about bacteriocin production by Bifidobacterium species. Therefore, the aim of this work was measure bacteriocin production in skim milk by B. lactis. Consequently, this work was divided in three stages. First, MRS, BSM and LD medium were tested with additives (Tween 80 (T80), Inuline (I) or Yeast extract (YE)) for bacteriocin production and cellular growth. Fermentation processes were conducted in shaker under specific conditions: 50 rpm/37ºC/48h. pH; sugars; acids; biomass, and bacteriocin activity against L. monocytogenes, L. plantarum, E. coli, L. sakei e S. aureus strains were analyzed . In the second stage, based on the obtained results, a central composite design (CCD) was created using the parameters: temperature (34, 37, 40 ºC), and concentration of YE (0.5, 1.0, 1.5 g/L). After, the activity was measured by two methods of plates pre-diffusion (cooling and addition of Tween 20). Third step consisted of 2 L bioreactor cultivations containing 10% skim milk diluted in 1.5 L of water (6.5 pH), under 200 rpm, 36 ºC, 2.0 g/L of YE, 48h, under anaerobic condition. Finally, the cultures supplemented with LD and YE (1%) with a modified plate diffusion method (cooling plates for 12 h) showed bacteriocin activity against L. monocytogenes (2130 AU/mL) with an exponential phase of 24 h, &#181;m of 0.604/h. The optimization performed using CCD resulted in a higher level of activity 3000 AU/mL to 3100 AU/mL mL (Run 7, 11 and 14, blocks 3 and 1) against L. monocytogenes, also with ideal growth conditions of YE: 2,0 g/L1 and T °C: 36 °C. The pH value varied between 6.4 and 4.0. Concentration of produced acid lactic varied from 3.03 to 4.72 g/L and biomass concentration from 3.4 to 11.1 Lg UFC/mL. Regression analysis was significant to the variables: YE concentration and temperature. Results indicated that skim milk is a proper medium for \"Bifidobacteriocin\" production.