La microfiltration tangentielle sur membrane en céramique est étudiée dans ce travail comme une méthode de fractionnement de suspensions hétérogènes obtenues après hydrolyse acide d'amidon de maïs cireux pour la récupération en continu de nanocristaux d'amidon. Le but est d'évaluer la possibilité de coupler la microfiltration au procédé d'hydrolyse pour augmenter le rendement de production des nanocristaux d'amidon. Une caractérisation des suspensions (taille et charge) a été réalisée à différents stades de la production afin de comprendre l'évolution de la taille des particules au cours de l'hydrolyse et du post-traitement et de pouvoir choisir la bonne membrane pour un fractionnement efficace. Deux pilotes de filtration ont été conçus pour permettre l'étude du fractionnement des suspensions de nanocristaux d'une part à l'échelle laboratoire (membranes planes) et d'autre part à l'échelle semi-industrielle (membrane tubulaire). L'analyse de la suspension produite par le procédé d'hydrolyse classique a montré qu'elle était constituée majoritairement d'agrégats de nanoparticules et de résidu d'amidon partiellement hydrolysé, la quantité de nanocristaux individualisés ne représentant que 5 % de l'amidon initial. L'opération de microfiltration tangentielle a été optimisée (transmission maximale des SNC et colmatage minimum des membranes) en fonction des conditions opératoires grâce à la réalisation d'un plan d'expériences. Dans les conditions optimisées, il était possible de récupérer dans le perméat, 25 % des particules initialement introduites en gardant des flux de perméat important. Ces particules ont une taille inférieure à 300 nm. Dans ces mêmes conditions, il était possible de séparer des nanocristaux d'amidon directement à partir de la suspension acide après hydrolyse. La modélisation du colmatage à partir des essais de filtration frontale a montré que le colmatage se fait essentiellement par formation d'un gâteau à la surface. Nous avons abordé également dans ce travail la piste de la purification en continu des suspensions acides à travers un procédé de diafiltration sur membranes d'ultrafiltration permettant une industrialisation de la production des SNC. / The current work investigates the use of cross flow microfiltration using ceramic membrane to fractionate the heterogeneous suspension obtained after starch hydrolysis in order to isolate starch nanocrystals. The final aim is to evaluate the possibility of coupling the filtration step to the hydrolysis step in a single production loop in order to enhance the starch nanocrystal production yield. The characterizations of the suspension (particle size and charge) obtained with the classic production process indicates that it was a mixture of starch nanocrystal aggregates and starch residues, individualized starch nanocrystals represent only 5 % of the initial starch. The fractionation study was done using two pilot plans, in a dead end configuration at laboratory scale (plate membrane) and in a cross flow configuration at semi-industrial scale (tubular membrane). Design of experiments methodology was used to optimize the fractionation efficiency when filtering a neutral suspension considering the effect of filtration parameter on the transmission yield and membrane fouling. In the optimized condition, it was possible to recover 25 % of starch nanocrystals while keeping the permeate flux at its highest value. Mean diameter of the recovered particle was less than 300 nm. At this condition, it was also possible to recover the starch nanocrystals directly from the acidic mixture obtained at the end of the hydrolysis step. The analysis of fouling mechanism using dead end filtration experiments highlights that membrane fouling occurs because a cake bult up at the membrane surface. This study investigates also the use of ceramic ultrafiltration membrane in a diafiltration process in order to purify the acidic suspension from soluble molecules without modifying particle size distribution which is a promising technique for a large scale production.

Submitted by Gabriela Lopes (gmachadolopesufpel@gmail.com) on 2016-10-10T18:10:31Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Tese Vania Zanela Pinto.pdf: 3876720 bytes, checksum: a2458e7baefb3806f5f42a1803a8a8b1 (MD5) / Approved for entry into archive by Aline Batista (alinehb.ufpel@gmail.com) on 2016-10-11T20:56:35Z (GMT) No. of bitstreams: 2 Tese Vania Zanela Pinto.pdf: 3876720 bytes, checksum: a2458e7baefb3806f5f42a1803a8a8b1 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2016-10-11T20:56:35Z (GMT). No. of bitstreams: 2 Tese Vania Zanela Pinto.pdf: 3876720 bytes, checksum: a2458e7baefb3806f5f42a1803a8a8b1 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2014-08-04 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Os amidos mais comumente utilizados industrialmente, são o amido de milho, o de trigo, o de batata e o de mandioca. No entanto, amidos oriundos de fontes não convencionais têm despertado grande interesse de indústrias e cientistas. Isso, associado a possibilidade de expandir e modificar as propriedades do amido por tratamentos físicos, conferem aos amidos de pinhão e de feijão vantagens para o uso na indústria de alimentos. Outra possibilidade para a aplicação de amidos é a utilização destes como matéria-prima alternativa na produção de embalagens e compósitos poliméricos, bem como a produção de nanocristais (SNCs). O estudo foi composto damodificação (pré-tratamentos) e caracterização dos amidos de pinhão e feijão Carioca e subsequente hidrólise dos mesmos para a produção de nanocristais. Os pré-tratamentos com tratamento térmico de baixa umidade (TTBU), annealing (ANN)ultrassom (SNT) e hidrólise enzimática (ENZ) empregados objetivaram diminuir o tempo e/ou aumentar o rendimento de hidrólise, bem como alterar e/ou melhorar as propriedades dos nanocristais a serem produzidos. Os amidos apresentaram alterações nas propriedades de pasta, térmicas e mudança no padrão de difração de raio-X, variando conforme a modificação aplicada. Os nanocristais foram produzidos através de hidrólise ácida (H2SO4 3,0 M) por 5 e 7 dias à 40°C. O rendimento após a hidrólise variou entre 10,2-14,7 % para o amido de pinhão e 35,2-42,3 % para o amido de feijão o que aumentou a cristalinidade relativa. O TTBU e o SNT aumentam o rendimento dos SNCs mas promovem perdas na cristalinidade relativa dos mesmos, enquanto que, o ANN promove melhor estabilidade térmica dos SNCs. O amido de feijão se mostrou promissor para a produção de SNCs, principalmente pelo elevado rendimentoapós 5 dias de hidrólise com ácido sulfúrico. / The most common industrially used starches come from corn, wheat, potato, and cassava. Nevertheless, starches from non-conventional botanical origins are receiving more attention of industries and researchers. Associated with this interest, the possibility of expand and modify the starch properties through physical modifications provide for pinhão and Carioca beans advantages in food industry purposes. The use of theses starches as an alternative to raw materials derived from petroleum for packaging and polymer composites is another possible application. Also included in this possibility, there is the production of starch nanoparticles and nanocrystals (SNCs). This study consisted in modifications (pretreatment), characterization and subsequent hydrolysis of pinhão and Carioca bean starches to produce starch nanocrystals. The starch pretreatments of heat-moisture-treatment (HMT), annealing (ANN), sonication (SNT) and mild enzymatic hydrolysis (ENZ) were used to improve the starch nanocrystals characteristics, as well as decrease the acid hydrolysis time and/or increase the hydrolysis final yield. The pretreatments changed the paste, thermal, and crystallinity properties of the starches, which changed according to the modification applied (HTM, ANN, SNT and ENZ). The SNCs were made by 3 M H2SO4 at 5 and 7 days at 40°C and obtained from the resulted removal of the amorphous lamella from the starch granules, what increased the relative crystallinity. The achieved hydrolysis yields were between 10.2 to 14.7% for pinhão starch and 35.2 to 42.3% for Carioca bean. HMT and SNT pretreatments increased the SNCs yield, but they promote a decrease on the SNCs relative crystallinity, whereas the ANN promotes a better thermal stability, mainly due to the high yield after 5 days of sulfuric acid hydrolysis.

Cette thèse étudie le procédé d’ultrafiltration tangentiel assisté par ultrasons aux échellesmacro et nanométriques. Différentes dispersions colloïdales ont été filtrées (argiles, micelle decaséine, nanocristaux d’amidon et de cellulose). Les propriétés d'écoulement et les changementsinduits par les ultrasons (US) ont été caractérisés. Les organisations structurelles à proximité de lamembrane ont été mises en évidence pour la première fois aux échelles nanométriques, lors de lafiltration par diffusion de rayons X aux petits angles in-situ. L’application des US a permis uneaugmentation significative des flux de perméation d’un facteur 1,6 à 13,5, selon l'organisationstructurale des colloïdes. Trois mécanismes induits par les US ont été identifiés : une érosioncomplète, une rupture partielle ou pas de changement (nanométrique) des couches de particulesaccumulées. Grâce aux profils de concentration obtenus, une approche de modélisation a permisune prévision du flux perméation. / This thesis studies an ultrasonic assisted cross-flow ultrafiltration process from macro tonano scales. Different types of colloids were investigated: synthetic and natural clay dispersions,casein micelles (skim milk) and starch or cellulose nanocrystal suspensions. Firstly, flowproperties and the changes due to ultrasound (US) were investigated. Secondly, structuralorganizations at nanometer length scales in the vicinity of the membrane during filtration havebeen revealed for the first time by real-time in-situ Small Angle X-ray Scattering. The applied USincreased significantly the permeate flux of ultrafiltration by an enhancement factor of 1.6 to13.5, depending on the structural organization of the colloids. The applied US has led to threemain effects: a removal of accumulated particle layer, a partial disruption or no change of thenano-organization. Thirdly, thanks to the obtained concentration profiles, a modeling approachhas allowed a prediction of the permeate flux.