Raphael Semiat

Effect of solvent properties on permeate flow through nanofiltration membranes. Part I: investigation of parameters affecting solvent flux

Abstract The objective of the present study was to characterize transport properties of solvents permeating through solvent resistant nanofiltration membranes that have only recently become available. Permeation flows of a number of solvents of different chemical families were measured in a batch cell. The solvents studied were alcohols, paraffins, ketones, acetates, and water, as well as their binary mixtures. The experimental data revealed a marked variation in the level of permeate flux among the various solvents. For instance, permeation flow of pentane was about 60-times faster compared to that of water while permeation flow of ethanol was about 10-times faster than that of water. The dependence of solution flux on the fractional composition of the solvents in the mixture was found to be highly non-linear. The flux of either pure or mixed solvents was mainly affected by surface tension and viscosity of the solvents. The flux of paraffins mixtures in particular was affected by the dielectric constant also.

UTILIZATION OF THE DONNAN EFFECT FOR IMPROVING ELECTROLYTE SEPARATION WITH NANOFILTRATION MEMBRANES

The possibility of utilizing the Donnan effect to enhance NaCl removal from a solution of a water-soluble organic dye (Procion Red H-E7B, ICI) using a nanofiltration membrane was examined. The increased salt removal is obtained by the addition of an ionized polyelectrolyte (Na salt of polyacrylic acid, PNa) having a molecular weight of 60 000 dalton. Salt rejection data and permeate flux were measured in three solution system: NaClH2O, NaClH2O-dye and NaClH2O-dye-PNa. The solutions were fed through a parallel plate osmotic cell having an area of 200 cm2, fitted with an MPF-44 nanofiltration membrane. Polyelectrolyte addition was found to provide a substantial enhancement of salt removal. For example, chloride rejection in a binary system with a salt weight concentration of 1% was around 50% (i.e. permeate salt concentration is about half that of the feed). Addition of 4.4% PNa decreased the salt rejection to a negative value of −68%, signifying that permeate salt concentration was almost 1.7 times higher than the feed concentration. However, polyelectrolyte addition reduced permeate flux and induced a flux limiting phenomenon, similar to that commonly encountered in ultrafiltration. A simple two parameter model is presented which successfully correlates the salt rejection data measured in this work, as well as results reported in the literature. The model enables prediction of salt rejection in multi-ionic systems from data measured easily in a single salt system and can be usefully applied for quantifying NF processes utilizing the Donnan effect.

Effect of solvent properties on permeate flow through nanofiltration membranes: Part II. Transport model

In Part I of this work, permeation flows of a large number of solvents were measured and found to exhibit a wide spread in permeate flux levels. The flux of both pure and mixed solvents was mainly affected by surface tension and viscosity. This paper presents a transport model describing solvent–membrane interactions, governed by viscous and surface forces. The model relates the flux of a solvent mixture with easily measurable solvent and membrane properties (surface tension, viscosity and membrane hydrophobicity). Extensive flux measurements of mixed solvents belonging to several chemical families were well correlated by the model using the following experimentally determined parameters. Membrane properties were characterized by two solvent independent coefficients f1 and f2, while, with minor exceptions, each solvent mixture was characterized by a specific coefficient φ.

Simple technique for measuring the concentration polarization level in a reverse osmosis system

This paper describes the application of a simple technique for determining the mass transfer coefficient and the concentration polarization level in a reverse osmosis (RO) system. The technique is based on evaluation of the permeate flux decline induced by the addition of a salt solution to an initially salt-free water feed. Since the net pressure driving force is influenced by the level of the osmotic pressure prevailing on the membrane surface, the magnitude of flux decline enables the evaluation of membrane surface concentration, and hence the determination of the mass transfer coefficient k. The mass transfer coefficient is given by: k=(J v ) salt /log((ΔP/(π b -π p ))(1-(J v ) salt /(J v ) H2O )). Hence, the value of k can be simply determined from the osmotic pressures π b and π p of the saline feed and of the permeate respectively and by measuring (J v ) H2O , the permeate flux of the salt-free water, and (J v ) salt , the permeate flux of the saline solution. The proposed technique was verified by experiments performed in a tubular RO system under turbulent flow conditions. Experiments covering the Reynolds number range of 2600-10000 yielded the following mass transfer correlation: Sh=(k.d)/D=0.020Re 0.91 S c 0.25 . This expression is practically identical with the theoretically anticipated Deissler correlation, thus lending strong support to the proposed mass transfer measurement technique.

Inception of CaSO4 scaling on RO membranes at various water recovery levels

This paper describes laboratory techniques for characterizing the permissible water recovery fraction in RO desalination, which is limited by the need to avoid scale deposition on the membrane. The CaSO4 scaling system was adopted for the development of these techniques. An upper water recovery bound is determined by operating an RO laboratory module such that the feed concentration level is gradually increased through continuous withdrawal of permeate while the concentrate is recycled to the feed vessel. The maximum possible water recovery is indicated by the solution concentration at which the supersaturation level triggers an immediate copious precipitation process. The procedure for evaluating a lower scaling threshold limit, at which precipitation is prevented or at least delayed for a sufficiently long period of time, is based on induction time experiments. These are carried out by operating the RO system in a full recycle mode (both permeate and concentrate recycled to the feed vessel) and measuring the time required for the onset of scaling. Results of preliminary tests show that induction time data are well correlated with the supersaturation level according to a relationship based on nucleation theory. Surface energies for CaSO4 nucleation on RO membranes are found to conform with literature data, thus lending support to the proposed induction time technique.

Influence of the flow system on the inhibitory action of CaCO3 scale prevention additives

Abstract The influence of hydrodynamics on the degree of CaCO3 scale suppression by anti-scalants was studied by comparing scale formation in the presence of inhibitors in two different flow configurations. One system consisted of a supersaturated solution flowing as a freely falling film outside a vertical stainless steel tube. In the other system, the supersaturated solution was in full flow through a stainless steel pipe. Tests characterizing scale formation and scale inhibition were performed with the two systems operated under comparable conditions with respect to water chemistry, wall shear stress and temperature. The experimental data showed that suppression of scale deposition by an anti-scalant in the falling film system was considerably more effective than that achieved in the pipe flow system. This rather unexpected effect is rationalized by analysis of the fundamental difference in the mechanisms of scale formation and scale inhibition in the two systems. Results of this study highlight the possibility that system hydrodynamics could be a parameter of no lesser importance than chemical factors in scale suppression by additive treatment.

Suppression of CaCO3 scale deposition by anti-scalants

Abstract The objective of this study was to compare the degree of CaCO 3 scale suppression provided by various antiscalants belonging to different chemical categories. The scaling system examined consisted of a freely falling film of a hot aqueous solution of Ca(HCO 3 ) 2 flowing down the outer surface of a vertical pipe. Evaporative air-cooling of the falling film released gaseous CO 2 from the water, thereby creating a substantial CaCO 3 scaling potential. Scale suppression effectiveness was determined by comparing the amount of scale deposited on a test pipe fed with water dosed with an anti-scalant with the amount deposited on an identical reference pipe, fed with the same water but without anti-scalant dosage. In the absence of an inhibitor, the scale layer was found to grow linearly with time, past an induction period. In the presence of an anti-scalant, the length of the induction period was extended and scale growth was arrested at an asymptotic limiting thickness. Scale suppression was more effective in the case of feed waters having a relatively higher temperature and a somewhat lower scaling potential, as compared to feed waters at a lower temperature and a somewhat higher scaling potential. No major difference could be discerned among the tested anti-scalants in the falling film system; all showed a substantially similar scale suppression effectiveness, within a narrow range of concentrations (0.2 to 0.5 ppm). Preliminary results of tests, aiming to check the applicability of the above conclusions to membrane systems, are also presented.

Synthesis, performance, and modeling of immobilized nano-sized magnetite layer for phosphate removal

A homogeneous layer of nano-sized magnetite particles (<4 nm) was synthesized by impregnation of modified granular activated carbon (GAC) with ferric chloride, for effective removal of phosphate. A proposed mechanism for the modification and formation of magnetite onto the GAC is specified. BET results showed a significant increase in the surface area of the matrix following iron loading, implying that a porous nanomagnetite layer was formed. Batch adsorption experiments revealed high efficiency of phosphate removal, by the newly developed adsorbent, attaining maximum adsorption capacity of 435 mg PO(4)/g Fe (corresponding to 1.1 mol PO(4)/mol Fe(3)O(4)). It was concluded that initially phosphate was adsorbed by the active sites on the magnetite surface, and then it diffused into the interior pores of the nanomagnetite layer. It was demonstrated that the latter is the rate-determining step for the process. Innovative correlation of the diffusion mechanism with the unique adsorption properties of the synthesized adsorbent is presented.

Laboratory technique for predicting the scaling propensity of RO feed waters

Abstract This paper describes the further development of a novel laboratory technique for characterizing the scaling potential of RO feed waters. The work is focused on the CaCO3 scaling system, which is of general interest and is encountered in RO purification of Seine River water. Scaling intensity at various water recovery levels is characterized by measuring permeate flux decline and changes in pH, hardness and alkalinity of a solution recycling in the system for a certain period of time, at the tested water compositions. The objective of the present study was to test the reliability of the above intermittent recycle technique for assessing scaling propensity and anti-scalant inhibitory effectiveness by comparing speedily determined laboratory results with long duration field data, measured with Seine River water. Scale deposition thresholds in the presence of four different anti-scalants were determined in the laboratory using simulated Seine River concentrates of compositions corresponding to water recovery levels of 60–90% and of scaling potentials ranging from Langelier saturation index values (LSI) of −0.1 to 2.5. Both field and laboratory results indicated that all four tested anti-scalants enabled a water recovery level of at least 88% at LSI levels exceeding 2.0. The laboratory measurements predicted correctly the field results in three out of four cases in which a comparison was possible. This agreement lends support to the usefulness of the proposed laboratory technique for convenient characterization of scaling propensity of RO feed waters.

Induction times induced in an RO system by antiscalants delaying CaSO4 precipitation

Abstract The aim of the present work is to consolidate laboratory techniques we have developed for determining the scaling threshold limits of RO feed waters and for assessing the relative inhibitory effectiveness of various antiscalants. The maximum permissible water recovery may be determined from induction time measurements of supersaturated solutions recycling through a laboratory RO module with permeate recycling so as to maintain a desired supersaturation level. A systematic series of experiments was carried out so as to characterize the extent of the induction period induced by various antiscalants. Five antiscalants were tested at two dosage levels (6 and 12 ppm) and at three CaSO4 supersaturation levels in the range of 3.7–5.5. Induction times data, extracted from plots of permeate flux vs. time, were successfully correlated with the supersaturation level according to a relationship based on nucleation theory. The results obtained confirm the capability of the proposed recycle technique to discriminate clearly among various antiscalants with respect to their scale suppression effectiveness. Further consolidation of the proposed techniques was obtained by demonstrating that induction times measured in recycle tests are virtually identical to those in once-through runs and that antiscalant molecules retained their full activity under prolonged exposure to pump shear action emanating from the recycling. Finally, surface energies for CaSO4 nucleation measured in this study with and without the presence of an antiscalant were found to be in excellent agreement with literature data. The results of this study lend confidence to the techniques we have developed for characterizing scaling threshold limits and for evaluating rationally the performance of anti-scalants.