Iron control during acidizing plays a key role in the success of matrix treatment. Ferric ion precipitates in the formation once the acid is spent and the pH exceeds 1-2. Precipitation of iron (III) within the formation can cause formation damage. Chelating agents such as EDTA and NTA are usually added to acids to minimize iron precipitation. Drawbacks of these chelating agents include limited solubility in strong acids and poor environmental profile. Hydroxy EDTA was introduced because of its higher solubility in 15 wt% HCl. However, its solubility in 28 wt% HCl is low and it is not readily biodegradable. In this study we studied the formation damage caused by iron precipitation in acidizing operations and tested the chelate L-glutamic acid, N,N-diacetic acid (GLDA). This chelant is soluble in higher concentrations of HCl. It is readily biodegradable, and is an effective iron control agent. A study was conducted to study the concentration of iron at different pHs ranging from 1-4 without the presence of any chelating agent at room temperature. A similar study was conducted in the presence of a chelating agent. To simulate field conditions, coreflood tests were conducted on Indiana Limestone, Austin Chalk and Pink Desert. Tests were conducted with and without the chelant. Samples of core effluent were collected and iron and calcium concentrations were measured using atomic absorption spectroscopy (AA). The cores were scanned using X-ray before and after acid injection. Results indicated that precipitation of iron can cause serious reduction in core permeability. The chelate was found to be very effective in chelating iron upto 300 degrees F. No permeability reduction was noted when GLDA was added to the acid. Material balance calculations show that significant amount of the iron that was added to the injected acid was produced when GLDA was used. This chelant is effective, environmentally friendly and can used up to 300 degrees F.

The preparation and characterisation of chelating sorbents suitable for the high efficiency separation of trace metals in complex samples, using a single column and isocratic elution, is described. Hydrophobic, neutral polystyrene divinylbenzene resins were either impregnated with chelating dyes or dynamically modified with heterocyclic organic acids, using physical adsorption and chemisorption processes respectively. A hydrophilic silica substrate was covalently bonded with a chelating aminomethylphosphonic acid group, to assess the chelating potential of this molecule. These substrates were characterised in terms of metal retention capability (selectivity coefficients and capacity factors), separation performance, column efficiency and suitability for analytical applications. Chelating molecules with different ligand groups were found to have unique selectivity patterns dependant upon the conditional stability constants of the chelate. Other factors, including mobile phase constituents - complexing agents, ionic strength and pH, column length and column capacity were additionally investigated to examine their effect upon the separation profiles achieved. The promising metal separation abilities illustrated by a number of these chelating columns were exploited for the determination of trace toxic metals in complex sample matrices using High Performance Chelation Ion Chromatography (HPCIC). This included the determination of beryllium in a certified stream sediment, uranium in seawater and a certified stream sediment, and cadmium, lead and copper in a certified rice flour. The results for each analysis fell within the certified limits, and reproducibility was good. The optimisation of post column detection systems using chromogenic ligands additionally gave good detection limits for the metals in each separation system.

The thermal processing of milk changes the composition and surface properties of the colloidal particles present and alters the physical properties of the milk. Whilst some changes such as those used to improve the texture of products such as yoghurt and are desirable, others such as gel formation during the manufacture of Ultra-High Temperature milk are highly undesirable. This work aims to characterize the effects of milk composition and pH on the chemical and physical changes that occur when milk is heated in order to understand and control the effect of thermal processing on the functional properties of the milk. Particularly important are: (i) the changes to the integrity of the casein micelles and the extent to which they are reversible on cooling of the heated milk, (ii) the changes to the speciation of the components of the serum as they re-equilibrate in response to the changed environment during heating and on cooling, (iii) the heat-induced denaturation of the whey proteins, (iv) the interaction between the components of the micelles and those in the milk serum, particularly those interactions that lead to aggregation or other changes that affect the functional properties of the milk on heating. This project includes thermal treatment (90°C/10 min) of control skim milk solutions (9% Milk Solids Non Fat) with or without addition of calcium chelating agents (orthophosphate (Pin) & Ethylenediaminetetraacetic acid (EDTA)) and concentrated skim milk solutions (up to 21% MSNF). The pH range chosen was 6.2 to 7.2. Almost all of the studies on heat stability to date have been carried out by heating the milk and determining the changes that have occurred after the milk is cooled. This project is focussed on the direct measurements in real time of the changes that occur at the exact temperature. The experimental techniques included pH, calcium activity and 31P NMR measurements at high temperatures to investigate the consequences to the change in mineral speciation, Size Exclusion Chromatography in combination with SDS-PAGE analysis for protein speciation during heating and Diffusing Wave Spectroscopy and viscosity measurements to determine the heat stability of milk systems. pH and calcium activity decreased with increase in temperature for all the milk systems studied. These changes were largely reversible as enough time was given for equilibration. pH and calcium activity changes during heating are a function of milk composition. The quantity, size and the composition of the protein aggregates present in the serum phase after mild centrifugation (~33,000g) of heated (90°C/10min) milk solutions were found to be a function of pH and milk composition (including the consequent differences in speciation of the components of milk). DWS and the viscosity measurements showed that pH at the temperature of heating is one of the primary determinants in influencing the aggregation of the proteins, which led to thermal stability of milk systems. Hence, changing the milk composition resulted in differences in pH at the temperature of heating, which led to different behaviours of heat stability of milk systems. Careful control of the composition of milk and thereby the pH at the temperature of heating allows a greater control of thermal stability of milk systems.

Fertilizer phosphorus (P) can become unavailable to crops due to immobilization of P in acidic soils through forming chemical bonds with iron (Fe) and aluminum (Al) amorphous oxides.  Organic chelating agents form strong bonds with metals in soil and may reduce P binding with Fe and Al.  Ethylenediamine tetraacetic acid (EDTA), hydroxyethyl ethylenediamine triacetic acid (HEEDTA), gluconic acid (GA), and citric acid (CA) were tested to determine their influence on water-soluble P (WSP), Mehlich-1 P and Mehlich-3 P in Loam and Sand soils fertilized with P and incubated for 49 days.  Soil P sorption capacity (PSC) was estimated from an oxalate extraction of Fe and Al, and chelates were applied at rates of 90 percent of the PSC.  The EDTA, HEEDTA, and CA significantly (P<0.05) reduced P sorption in the Loam and Sand when measured by WSP.  In soils without P fertilizer added, EDTA and HEEDTA significantly increased WSP, Mehlich-1, and Mehlich-3 P concentrations.  EDTA and HEEDTA were also applied at 0, 30, 60, 90, 120, and 150 percent PSC to produce a rate response curve for WSP in a second soil incubation.  With increasing chelating rate, there was a linear increase in WSP for both soils, thus indicating higher rates of chelating agents were most efficient at decreasing P sorption. EDTA and HEEDTA were also tested in a 4-week greenhouse study for efficiency at increasing plant available P to corn (Zea mays L.) in two soils.  Phosphorus was added with and without the addition of chelating agents to the center of the pot, simulating a starter band of P.  After 4weeks, soils were analyzed for WSP, Mehlich-1, and Mehlich-3 P and corn above- and below-ground biomass was quantified and analyzed for total P concentration.  Without the presence of chelating agents, concentrations of WSP, Mehlich-1 P, Mehlich-3 P, above- and below-ground biomass, and TKP increased linearly as P fertilizer rates increased at 0, 9.6, 19.3, 28.9, and 38.5 kg P ha-1.  Decreased P sorption using chelating agents was not observed in this experiment.  However, with the results from the soil incubation, chelating agents do show potential for increasing plant available P, but the application and incorporation method needs to be further studied. / Master of Science

Thesis (Ph. D.)--University of Missouri--Columbia, 1998. / "December 1998." Typescript. Vita. Includes bibliographical references (leaves 157-166). Also available on the Internet.

Este estudo teve como objetivo avaliar a capacidade de remoção da smear layer das paredes de canais radiculares instrumentados pelas soluções de soro fisiológico, EDTA a 17%, vinagre de maçã e SmearClear®, nos três níveis radiculares, associadas ou não ao ultra-som. Para tal propósito, foram utilizados 70 dentes caninos humanos superiores e inferiores extraídos os quais tiveram suas coroas removidas. O comprimento de trabalho foi determinado pela visualização da ponta de uma lima tipo K com diâmetro 10 no forame apical, subtraindo-se 1 milímetro dessa medida. Para o preparo biomecânico, instrumentou-se os dentes com o sistema rotatório ProTaper®, pela técnica coroa-ápice, até o instrumento F5, irrigando os canais com hipoclorito de sódio a 2,5% a cada troca de instrumento. Os dentes foram divididos aleatoriamente em 6 grupos de 10 dentes e 2 grupos controle de 5 dentes de acordo com o método de irrigação final a serem testados: Grupo 1 (controle) Soro fisiológico por 3 minutos sem ultra-som; Grupo 2 - (controle) Soro fisiológico por 3 minutos, sendo o primeiro minuto ativado pelo ultra-som; Grupo 3 - EDTA a 17% por 3 minutos sem ultra-som; Grupo 4 - EDTA a 17% por 3 minutos, sendo o primeiro minuto ativado pelo ultra-som; Grupo 5 - Vinagre de maçã por 3 minutos sem ultra-som; Grupo 6 - Vinagre de maçã por 3 minutos, sendo o primeiro minuto ativado pelo ultra-som; Grupo 7 - SmearClear® por 1 minuto sem ultra-som; Grupo 8 - SmearClear® por 1 minuto ativado pelo ultra-som. Em seguida, os dentes foram irrigados com hipoclorito de sódio a 2,5% e soro fisiológico e secos com cones de papel absorvente. As raízes foram seccionadas longitudinalmente e preparadas para a microscopia eletrônica de varredura com o aumento de 750 vezes. Fotomicrografias foram obtidas dos terços cervical, médio e apical e analisadas por três examinadores para a atribuição de escores para posterior análise estatística. Concluiu-se que as soluções de EDTA e SmearClear® foram efetivas na remoção da smear layer do canal radicular, com pequena vantagem numérica do EDTA, entretanto, sem diferença estatisticamente significante entre elas. O soro fisiológico apresentou resultado significativamente inferior comparado às demais soluções irrigadoras. O vinagre de maçã se mostrou ligeiramente inferior ao SmearClear® e ao EDTA, apresentando diferença estatisticamente significante ao EDTA no terço cervical. O uso do ultra-som não promoveu uma melhora significante na capacidade de limpeza de todas as soluções irrigadoras testadas. Em todos os grupos, o resultado menos favorável em relação à remoção da camada de smear layer foi observado no terço apical do canal radicular, com diferença estatisticamente significante em relação ao terço cervical. / The aim of this study was to evaluate the smear layer removal from instrumented root canal walls by saline solution, 17% EDTA, apple vinegar and SmearClear®, at three radicular levels, with or without ultrasonics. For that purpose, 70 upper and lower human canines were used, which had their crowns removed. The working length was established by visualizing the tip of a 10 K-type file at the apical foramen, by reducing 1 mm from this measurement. For preparation of root canal, the teeth were instrumented with rotatory files ProTaper®, using the crown-down technique, up to F5 instrument, irrigating with 2,5% sodium hypochlorite between each instrument change. The teeth were randomly divided into 6 groups of 10 teeth each and 2 control groups of 5 teeth according to the final irrigation protocol: Group 1 (control) Saline solution for 3 minutes without ultrasonics; Group 2 (control) Saline solution for 3 minutes, with ultrasonic activation for the first minute; Group 3 17% EDTA for 3 minutes without ultrasonics; Group 4 17% EDTA for 3 minutes, with ultrasonic activation for the first minute; Group 5 Apple vinegar for 3 minutes without ultrasonics; Group 6 Apple vinegar for 3 minutes, with ultrasonic activation for the first minute; Group 7 - SmearClear® for 1 minute without ultrasonics; Group 8 - SmearClear® for 1 minute with ultrasonic activation. Subsequently, the teeth were irrigated with 2,5% sodium hypochlorite and dried with paper points. The teeth were split open longitudinally and prepared for scanning electron microscopy at magnification of 750X. Photomicrographs were taken from coronal, middle and apical thirds and analyzed by three examiners for scores attribution for statistical analysis. It was concluded that EDTA and SmearClear® were efficient for smear layer removal from root canals, with a discrete superiority of EDTA, however, without statistical significance. The saline solution showed significant inferior results compared with the others solutions. Apple vinegar was slightly inferior to SmearClear® and EDTA, with statistical difference observed between apple vinegar and EDTA in coronal third. The ultrasonics did not enhance the cleaning ability of all the irrigating solutions tested. In all cases, the least favorable result regarding smear layer removal was observed in the apical third of the root canal, with statistically significant differences in the coronal third.

No description available.

The present work is divided into two sections. The first section deals with the synthesis of regenerable adsorbents for the removal of arsenic from contaminated waters. An adsorbent based on carboxymethylated polyethylenimine grafted agarose gels was synthesized and characterized as a regenerable synthetic ferric oxide adsorbent with high capacity for arsenate ions at pH 3.0. Similarly, four metal ion chelating adsorbents based on dipicolylamine were synthesized and characterized with respect to their Cu(II), Fe(III) and As(V) adsorption capacities. The most efficient adsorbents were Nov-PEI-DPA and Nov-TREN-DPA. Additionally, a commercial ion exchange resin was modified with permanganate to oxidize arsenite into arsenate. A complete oxidation-adsorption system was proposed in which a column packed with the oxidation resin was connected in series with an adsorbent column composed of the polyethylenimine grafted agarose gels.The second section involved work with magnetic nanoparticles. First, composite adsorbents consisting of magnetic particles encapsulated within agarose beads with and without grafted iminodiacetic acid (IDA) chelating groups were synthesized. The adsorption capacity of the adsorbents for Cu(II), Fe(III) and As(V) at different concentrations was investigated. Batch experiments were carried out to determine the Fe(III) and As(V) adsorption isotherms for the magnetic Novarose-IDA. Regenerability of the adsorbent was achieved with a pH change of the inlet solution, without affecting its magnetic or adsorption properties.Magnetic composite particles were synthesized for biomedical applications. First, magnetic nanoparticles were coated with silica and then used for gold nanoshell production. These nanoshells were functionalized with a Brij S10 derivative, containing carboxylic groups, using dodecanethiol as a bridging agent to incorporate a fluorescent biomolecule.Finally, magnetic and gold particles were encapsulated in PLGA nanoparticles. Docetaxel was loaded on these multifunctional nanoparticles and released studies were performed at 37°C. The presence of magnetite, colloidal gold and gold nanoshells in the PLGA nanoparticles was revealed by the coloration acquired by the polymeric nanoparticles. The release of drug from the polymeric nanoparticles showed a biphasic behavior with an initial burst followed by a prolonged slow release. There was no effect of the presence of magnetic or metallic particles on docetaxel release.

No description available.

Níquel laterítico compõe 70% das reservas disponíveis do metal. A produção de níquel a partir dessas reservas representa 40%, dos quais é possível extrair também cobre e cobalto. Isso ocorre devido ao alto teor de impurezas, principalmente ferro. Com a crescente demanda desses metais, o uso das reservas de lateritas de níquel passou a ser mais profundamente investigado, assim como o desenvolvimento de processos hidrometalúrgicos. Nesse caso, o íon férrico prejudica a recuperação do níquel e do cobalto, uma vez que em trocas iônicas esse metal compete na ocupação dos sítios catiônicos. Em processos de extração por solvente um efeito semelhante é observado. Além disso, durante a eventual precipitação do ferro ocorre a coprecipitação. Assim, para a troca iônica, o íon ferroso pode ser menos prejudicial do que o férrico. O presente trabalho teve por objetivo o estudo do processo de redução do íon férrico em solução aquosa, e também a posterior obtenção de cobalto e níquel por meio de resinas quelantes. Estudou-se o processo de redução do íon férrico e o efeito na adsorção de metais por troca iônica. Os ensaios de redução química do íon férrico em solução foram estudados utilizando ditionito de sódio, metabissulfito de sódio e sulfito de sódio. Os ensaios com os agentes redutores ditionito e metabissulfito de sódio foram feitos em São Paulo, e os ensaios com o sulfito de sódio foram feitos na The University of British Columbia. Os agentes redutores foram adicionados na solução monoelementar de ferro para redução do potencial redox. As variáveis potencial redox, entre 860mV e 240mV; pH, entre 0,5 e 3,5; temperatura, entre 25°C e 60°C; e tempo, entre 30min e 96 horas, foram estudadas em frascos erlenmeyer sob agitação constante. Analisou-se, então, a redução do íon férrico em solução multielementar de níquel laterítico. Nos ensaios de troca iônica, realizados em batelada e em coluna, utilizou-se a resina quelante Lewatit TP 207, de grupo funcional iminodiacetato, e a resina Lewatit TP220, de grupo funcional bis-picolilamina. Três soluções foram estudadas: uma preparada com Fe(III), outra com Fe(II) e a terceira com Fe(III) após o processo de pré-redução. Os ensaios em batelada foram realizados com a utilização de frascos erlenmeyer, sob agitação constante, com 100mL de solução para 1mL de resina. Estudou-se o efeito do pH, entre 0,5 e 3,5; tempo, entre 30min e 480min; e temperatura, entre 25°C e 60°C. Nos ensaios em sistema contínuo, as três soluções foram alimentadas em colunas de vidro preenchidas com resina. A solução foi alimentada com bombas peristálticas a vazão constante. Para eluição das colunas, ácido sulfúrico 1mol.L-1 foi alimentado na coluna com utilização de bomba peristáltica. A segunda parte do trabalho, realizado na The University of British Columbia, foi o estudo do uso de sulfito de sódio no processo de pré-redução. Foram estudadas duas resinas: a Lewatit TP 207, seletiva para cobre; e a Lewatit TP 220, seletiva para níquel e cobalto. O sulfito de sódio foi adicionado na solução para redução do potencial, em frascos, e colocado sob agitação constante. Após reação, as soluções foram colocadas em contato com a resina quelante, e ficaram em agitação. Os ensaios em batelada foram realizados e o efeito do pH estudado entre 0,5 e 3,5. No processo em coluna, a Coluna 1, preenchida com a resina Lewatit TP 207, foi utilizada para remoção do cobre; e a solução de saída foi alimentada na Coluna 2, preenchida com a resina Lewatit TP 220. Para a eluição, foram estudados os ácidos clorídricos e sulfúricos em duas diferentes concentrações, 1mol.L-1 e 2mol.L-1. Hidróxido de sódio foi utilizado para remover o ferro na solução obtida na saída da Coluna 2. A separação do cobalto da solução foi feita utilizando a técnica de extração por solventes (Cyanex 272 20%), estudando o efeito do pH, 4,0 e 5,0, e da temperatura, 25°C e 65°C. Os resultados mostraram que a redução do íon férrico utilizando ditionito de sódio foi de 100% na solução monoelementar e de 70% na multielementar contendo os outros metais. Nos ensaios de troca iônica em batelada, utilizando a resina TP 207, 62% do cobre foi adsorvido na solução após processo de pré-redução. Para solução com Fe(II), a adsorção de cobre foi de 61%; e para solução com Fe(III), 49%. Nos ensaios de troca iônica após pré-redução do ferro com sulfito de sódio, a adsorção do cobre foi de 69% em pH 2,0 pela resina TP 207. A resina TP 220 foi mais seletiva para níquel e cobalto em pH 2,0, em que as adsorções destes metais foram 32,5% e 69%, respectivamente. Nos ensaios em coluna, a Coluna 1 foi utilizada para remoção de cobre, porém houve perda de 17% de níquel e 7% de cobalto. Na alimentação da Coluna 2, verificou-se que 98% do níquel e 84% do cobalto foram adsorvidos. A solução obtida da Coluna 2 teve concentração de 618mg.L-1 de ferro, 13231mg.L-1 de níquel e 179mg.L-1 de cobalto. A remoção de 100% do ferro foi possível em pH 4,0. Para separação do cobalto da solução rica em níquel, utilizou-se a extração por solventes com o Cyanex 272 20% em querosene, no qual 99% do cobalto foi separado da solução a 65°C e pH 5,0, sem perda de níquel. Para estudos futuros, a remoção do cobre no início do processo pode ser explorada com a utilização de outras técnicas, a fim de se evitarem perdas de níquel e cobalto. Outro ponto que pode vir a ser explorado é a máxima utilização das colunas de troca iônica nas mesmas condições deste trabalho - solução com Fe(III), com Fe(II) e após processo de pré-redução - sobretudo em escala piloto, para estudar o efeito do estado de oxidação do ferro em um possível envenenamento da resina. / Nickel laterite ores represent 70% of the available metal reserves. The nickel production from these reserves represents 40%, where it is also possible extract copper and cobalt. It occurs due to the high impurities content, mainly iron. With the growing demand of these metals, the use of nickel laterite reserves became more deeply investigated, as well as hydrometallurgical process development. In this case, the ferric iron difficult the nickel and cobalt recovery, once in ion exchange processes this metal competes in the occupation of the cationic sites. In solvent extraction processes the same effect is observed. Besides that, during the eventual iron precipitation there is a co-precipitation. Therefore, for ion exchange, ferrous iron may be less damaging than ferric iron. The aim of this work was to study the reducing process of ferric iron in aqueous solution, and also the subsequent obtaining of cobalt and nickel through chelating resins. It was studied the reducing process of ferric iron and the effect of it in metals adsorption by ion exchange. Experiments of chemical reduction of ferric iron in solution were studied using sodium dithionite, sodium metabisulfite and sodium sulfite. Experiments with reducing agents sodium dithionite and metabisulfite were performed in São Paulo, and experiments with sodium sulfite were performed in The University of British Columbia. Reducing agents were added in ferric iron mono-elementary solution to decrease the redox potential. The variables potential redox, between 860mV and 240mV; pH, between 0,5 and 3,5; temperature, between 25°C e 60°C; and time, between 30min and 96 hours were studied in erlenmeyer flasks under constant stirring. Then, analyzed ferric iron reduction in multielementary solution of nickel laterite. In ion exchange experiments, performed in batch and column, it was used chelating resin Lewatit TP 207, with iminodiacetate functional group, and resin Lewatit TP 220, with bis-picolylamine functional group. Three solutions were studied: prepared with Fe(III), other with Fe(III) and the third with Fe(III) after pre-reducing process. Experiments in batch system were performed using erlenmeyer flasks, under constant stirring, with 100mL of solution to 1mL of resin. It was studied the effect of pH, between 0,5 and 3,5, time, between 30min and 480min, and temperature, between 25°C e 60°C. In experiments in continuous system, the three solutions were fed in glass columns filled with resin. The solution was fed using peristaltic pumps at constant flow rate. For column elution, sulfuric acid 1mol.L-1 was fed to the column using peristaltic pumps. The second part of this work, performed at The University of British Columbia, was the study of sodium sulfite application at prereducing process. It was studied two resins: Lewatit TP 207, selective for copper, and Lewatit TP 220, selective for nickel and cobalt. Sodium sulfite was added to the solution to decrease the potential, in flasks and it was placed under constant stirring. After reaction, the solution was placed in contact to the chelating resin, which was placed under stirring. Batch experiments were performed, and the effect of pH was studied between 0,5 and 3,5. In column process, the Column 1, filled with Lewatit TP 207, was used for copper removal, and output solution was feed in Column 2, filled with Lewatit TP 220. For the elution, it was studied sulfuric and hydrochloric acids in two different concentrations, 1mol.L-1 e 2mol.L-1. Sodium hydroxide was used for iron removal from solution obtained in Column 2 output. Cobalt separation was performed using solvent extraction technique (Cyanex 272 20%), studying the effect of pH, 4,0 and 5,0, and temperature, 25°C e 65°C. Results shows that ferric iron reduction using sodium dithionite was 100% in mono-elementary solution and 70% in multi-elementary solution with other metals. In ion exchange experiments performed in batch using resin TP 207, 62% of copper was adsorbed by the resin after pre-reducing process. For solution with Fe(II), the copper adsorption was 61%, and for solution with Fe(III), 49%. In ion exchange experiments after pre-reducing process using sodium sulfite, the copper adsorption was 69% at pH 2,0 by the resin TP 207. The resin TP 220 was more selective for nickel and cobalt at pH 2,0, where these metals adsorptions were 32,5% and 69%, respectively. In experiments performed in column, the Column 1 was used for copper removal, however there were losses of nickel (17%) and cobalt (7%). In the feeding of Column 2, it was found that 98% of nickel and 84% of cobalt were adsorbed. Solution obtained in Column 2 had concentration of iron 618mg.L-1, nickel was 13231mg.L-1 and cobalt 179mg.L-1. The iron removal was 100% at pH 4,0. For cobalt separation in nickel-rich solution, it was used the solvent extraction with Cyanex 272 20% with kerosene, where 99% of cobalt was separated from solution at 65°C and pH 5,0, without nickel loss. For future studies, the copper removal in the beginning of the process can be explored using other techniques, in order to avoid nickel and cobalt losses. Another point that can be explored is the maximum use of ion exchange columns in the same conditions of this work - solution with Fe(III), with Fe(II) and after the pre-reducing process - mostly on a pilot scale, to study the effect of iron oxidation state on possible resin poisoning.