- Email: [email protected]
Physico-chemical effect of simple alkaline and acid solutions in cleaning sequences of spiral ultrafiltration membranes fouled by skim milk
Desalination 200 (2006) 192–194
Physico-chemical effect of simple alkaline and acid solutions in cleaning sequences of spiral ultrafiltration membranes fouled by skim milk Lydie Paugam*, Murielle Rabiller-Baudry, David Delaunay Chimie et Ingénierie des Procédés, UMR 6226 CNRS “Sciences Chimiques de Rennes” Université Rennes 1– ENSCR, 263 avenue du général Leclerc, Campus Beaulieu Batiment 10 A, CS 74205, 35042 Rennes Cedex, France email: [email protected] Received 27 October 2005; accepted 3 March 2006
1. Introduction Ultrafiltration (UF) is a membrane process widely used in dairy industry in order to adjust the protein content before the cheese making process. The high fouling resulting from milk filtration induces 2–3 h daily cleaning step to restore membrane performances, limiting thus the productivity. The chemical cleaning (apart from disinfection), carried out (at 50° C) on empirical and non optimised bases, is as follows: alkaline cleaning (pH 11.5, to remove organic matter) followed by acid cleaning (pH 1.6, to remove mineral matter). But this cleaning mode cannot prevent the decrease of 50% of the production flux over several months, often observed at industrial scale. Optimisation of the cleaning step can be performed according to various approaches, among them: • Looking for a more efficient cleaning solution [1] • Limiting the number of chemicals used [2]
*Corresponding author.
• Optimising the alkaline/acid sequence with single solutions (synergy). Our previous works have shown that the irreversible fouling is mainly due to proteins. Moreover quantification of proteins by FTIRATR on a plane PES membrane, after UF of skim milk in similar conditions as in the spiral membrane, showed that 24 (±2) µg of protein per cm2 remained after water rinsing. These proteins are the target of the cleaning.
2. Experimental In this study several single solutions are used: nitric acid at pH = 1.6, hydrochloric acid at pH = 1.6, sodium hydroxide pH = 11.5 and a chlorinated alkaline one at pH = 11.5 with 200 ppm of active Cl2. The efficiency of these solutions is determined for the cleaning (TMP = 2 bar, 50° C, VRR = 1, v = 0.5 m s–1, t = 1 h) of spiral PES membrane (HFK-131, Koch, 5–10 kg mol–1, 6.5 m2) fouled by skim milk during 150 min in the same filtering conditions.
Presented at EUROMEMBRANE 2006, 24–28 September 2006, Giardini Naxos, Italy. 0011-9164/06/$– See front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.desal.2006.03.292
L. Paugam et al. / Desalination 200 (2006) 192–194
3. Results and discussion 3.1. Efficiency of simple alkaline solutions Neither of these solutions allows to reach the membrane hydraulic property (0.90 of its initial flux) in only one step. For all of them, the flux values are nearby (0.57 < J/J0 < 0.62, with J0 the starting water flux before milk filtration and J the water flux after cleaning and rinsing). The sodium hydroxide sole does not hydrolyse proteins (2 bonds maximum per hour at 50° C and pH = 11.5) [3]; it is therefore the modification of the charge state of the proteins imposed by the pH change that might explain the action of the sodium hydroxide. The chlorinated alkaline solution efficiency is slightly better, probably due to the oxidation by Cl2. The latter flux result is consistent with the analysis of the residual protein deposit on the surface of a plane PES model membrane (cleaned and fouled in similar conditions). 3.2. Efficiency of simple acid solutions — role of nitrate ions The use of hydrochloric acid does not improve the water permeability of the membrane which is consistent with the lack of inorganic fouling deposit on the membrane. On the other hand, the cleaning with nitric acid leads to an increase of the water flux without any reduction of the residual protein deposit on the membrane surface. The flux alone cannot be the only criteria to evaluate the effective cleaning of the membrane. The water flux enhance after the nitric acid cleaning is discussed with respect to specific interactions of nitrate towards residual proteins on the membrane. 3.3. Sequences Various sequences are studied: nitric acid + alkaline, alkaline +nitric acid and alkaline +alkaline (the alkaline solution being chlorinated or not).
193
A first step with nitric acid does not increase the efficiency of the further alkaline step (with or without oxidant). This confirms the ineffective role of acid solutions in the decrease of the protein fouling. So the difference of flux, higher with an alkaline +acid sequence compared with a single alkaline step, is only due to the nitrate adsorption on the membrane and is not a reflection of cleaning efficiency. Two successive alkaline steps (without oxidant) allow reaching the hydraulic property (J/J0 = 0.90) and two chlorinated alkaline lead to the recovery of the initial flux (0.99). The need of two successive steps to obtain the target flux might be explain in terms of a membrane surface to active species in cleaning solution ratio. Limitation of efficiency could be due to the increase of protein/peptide concentration in solution, as the experiments of this work are carried out in batch circulation mode. 4. Conclusions The use of alkaline simple solutions allows reaching the hydraulic property of the membrane in two steps (VRR = 1) of 40 min each one (optimised duration). If the nitric acid is usually used as a second step in the cleaning sequence, its inefficiency in the cleaning of PES UF membranes fouled by skim milk is highlighted whereas it leads to a misleading increase of the flux. Specific interactions between the residual proteins on the membrane and the nitrate ions occur leading to a decrease of the membrane hydrophobicity. This study is a part of a more global work which aims at optimising the cleaning step in terms of cleaning efficiency, energy and water consumption and at limiting required posttreatment. In this context, this work turns the way to follow towards the formulation of an alkaline detergent efficient in a single step (with an optimised duration) and integrating if possible the disinfection step.
194
L. Paugam et al. / Desalination 200 (2006) 192–194
Another point highlighted here is the necessity of analysing the surface of the membrane as some interactions between the cleaning solutions and the residual protein deposit are often suspected. This methodology allows to confirm, when reliable, flux results and to use them without any ambiguity and allows highlighting physico-chemical phenomenon responsible for the cleaning efficiency.
[2]
[3]
References [1]
M. Rabiller-Baudry, L. Paugam, L. Bégoin, D. Delaunay, M. Fernandez-Cruz, C. Phina-Zieben,
C. Laviades-Garcia de Guadiana and B. Chaufer, Alkaline cleaning of PES membrane of skimmed milk ultrafiltration from reactor to spiral wound module via plate and frame module, Desalination, 191(1–3) (2006) 334–343. M. Rabiller-Baudry, L. Bégoin, D. Delaunay, L. Paugam and B. Chaufer, A dual approach of membrane cleaning based on physico-chemistry and hydrodynamics. Application to PES membrane of dairy industry, Chem. Eng. Process, submitted. L. Bégoin, Analyse de modules spirales industriels d’ultrafiltration de produits laitiers. Physicochimie du nettoyage de membranes en polyéthersulfone d’ultrafiltration de lait écrémé, Thèse de l’Université de Rennes 1 (2004).
Physico-chemical effect of simple alkaline and acid solutions in cleaning sequences of spiral ultrafiltration membranes fouled by skim milk Lydie Paugam*, Murielle Rabiller-Baudry, David Delaunay Chimie et Ingénierie des Procédés, UMR 6226 CNRS “Sciences Chimiques de Rennes” Université Rennes 1– ENSCR, 263 avenue du général Leclerc, Campus Beaulieu Batiment 10 A, CS 74205, 35042 Rennes Cedex, France email: [email protected] Received 27 October 2005; accepted 3 March 2006
1. Introduction Ultrafiltration (UF) is a membrane process widely used in dairy industry in order to adjust the protein content before the cheese making process. The high fouling resulting from milk filtration induces 2–3 h daily cleaning step to restore membrane performances, limiting thus the productivity. The chemical cleaning (apart from disinfection), carried out (at 50° C) on empirical and non optimised bases, is as follows: alkaline cleaning (pH 11.5, to remove organic matter) followed by acid cleaning (pH 1.6, to remove mineral matter). But this cleaning mode cannot prevent the decrease of 50% of the production flux over several months, often observed at industrial scale. Optimisation of the cleaning step can be performed according to various approaches, among them: • Looking for a more efficient cleaning solution [1] • Limiting the number of chemicals used [2]
*Corresponding author.
• Optimising the alkaline/acid sequence with single solutions (synergy). Our previous works have shown that the irreversible fouling is mainly due to proteins. Moreover quantification of proteins by FTIRATR on a plane PES membrane, after UF of skim milk in similar conditions as in the spiral membrane, showed that 24 (±2) µg of protein per cm2 remained after water rinsing. These proteins are the target of the cleaning.
2. Experimental In this study several single solutions are used: nitric acid at pH = 1.6, hydrochloric acid at pH = 1.6, sodium hydroxide pH = 11.5 and a chlorinated alkaline one at pH = 11.5 with 200 ppm of active Cl2. The efficiency of these solutions is determined for the cleaning (TMP = 2 bar, 50° C, VRR = 1, v = 0.5 m s–1, t = 1 h) of spiral PES membrane (HFK-131, Koch, 5–10 kg mol–1, 6.5 m2) fouled by skim milk during 150 min in the same filtering conditions.
Presented at EUROMEMBRANE 2006, 24–28 September 2006, Giardini Naxos, Italy. 0011-9164/06/$– See front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.desal.2006.03.292
L. Paugam et al. / Desalination 200 (2006) 192–194
3. Results and discussion 3.1. Efficiency of simple alkaline solutions Neither of these solutions allows to reach the membrane hydraulic property (0.90 of its initial flux) in only one step. For all of them, the flux values are nearby (0.57 < J/J0 < 0.62, with J0 the starting water flux before milk filtration and J the water flux after cleaning and rinsing). The sodium hydroxide sole does not hydrolyse proteins (2 bonds maximum per hour at 50° C and pH = 11.5) [3]; it is therefore the modification of the charge state of the proteins imposed by the pH change that might explain the action of the sodium hydroxide. The chlorinated alkaline solution efficiency is slightly better, probably due to the oxidation by Cl2. The latter flux result is consistent with the analysis of the residual protein deposit on the surface of a plane PES model membrane (cleaned and fouled in similar conditions). 3.2. Efficiency of simple acid solutions — role of nitrate ions The use of hydrochloric acid does not improve the water permeability of the membrane which is consistent with the lack of inorganic fouling deposit on the membrane. On the other hand, the cleaning with nitric acid leads to an increase of the water flux without any reduction of the residual protein deposit on the membrane surface. The flux alone cannot be the only criteria to evaluate the effective cleaning of the membrane. The water flux enhance after the nitric acid cleaning is discussed with respect to specific interactions of nitrate towards residual proteins on the membrane. 3.3. Sequences Various sequences are studied: nitric acid + alkaline, alkaline +nitric acid and alkaline +alkaline (the alkaline solution being chlorinated or not).
193
A first step with nitric acid does not increase the efficiency of the further alkaline step (with or without oxidant). This confirms the ineffective role of acid solutions in the decrease of the protein fouling. So the difference of flux, higher with an alkaline +acid sequence compared with a single alkaline step, is only due to the nitrate adsorption on the membrane and is not a reflection of cleaning efficiency. Two successive alkaline steps (without oxidant) allow reaching the hydraulic property (J/J0 = 0.90) and two chlorinated alkaline lead to the recovery of the initial flux (0.99). The need of two successive steps to obtain the target flux might be explain in terms of a membrane surface to active species in cleaning solution ratio. Limitation of efficiency could be due to the increase of protein/peptide concentration in solution, as the experiments of this work are carried out in batch circulation mode. 4. Conclusions The use of alkaline simple solutions allows reaching the hydraulic property of the membrane in two steps (VRR = 1) of 40 min each one (optimised duration). If the nitric acid is usually used as a second step in the cleaning sequence, its inefficiency in the cleaning of PES UF membranes fouled by skim milk is highlighted whereas it leads to a misleading increase of the flux. Specific interactions between the residual proteins on the membrane and the nitrate ions occur leading to a decrease of the membrane hydrophobicity. This study is a part of a more global work which aims at optimising the cleaning step in terms of cleaning efficiency, energy and water consumption and at limiting required posttreatment. In this context, this work turns the way to follow towards the formulation of an alkaline detergent efficient in a single step (with an optimised duration) and integrating if possible the disinfection step.
194
L. Paugam et al. / Desalination 200 (2006) 192–194
Another point highlighted here is the necessity of analysing the surface of the membrane as some interactions between the cleaning solutions and the residual protein deposit are often suspected. This methodology allows to confirm, when reliable, flux results and to use them without any ambiguity and allows highlighting physico-chemical phenomenon responsible for the cleaning efficiency.
[2]
[3]
References [1]
M. Rabiller-Baudry, L. Paugam, L. Bégoin, D. Delaunay, M. Fernandez-Cruz, C. Phina-Zieben,
C. Laviades-Garcia de Guadiana and B. Chaufer, Alkaline cleaning of PES membrane of skimmed milk ultrafiltration from reactor to spiral wound module via plate and frame module, Desalination, 191(1–3) (2006) 334–343. M. Rabiller-Baudry, L. Bégoin, D. Delaunay, L. Paugam and B. Chaufer, A dual approach of membrane cleaning based on physico-chemistry and hydrodynamics. Application to PES membrane of dairy industry, Chem. Eng. Process, submitted. L. Bégoin, Analyse de modules spirales industriels d’ultrafiltration de produits laitiers. Physicochimie du nettoyage de membranes en polyéthersulfone d’ultrafiltration de lait écrémé, Thèse de l’Université de Rennes 1 (2004).