Reverse osmosis experiments

Dandavati, M S., Doshi, M. R., and Gill, W. N. (1975). Hollow fiber reverse osmosis Experiments and analysis of radial flow systems. Chem. Eng. Sci., 30, 877-886. 

Figures 2 and 3 show typical test results for flux decline in laminar flow where the pressure and temperature are varied and the Reynolds number is held fixed. Similar behaviors are found with variations in Reynolds number and for turbulent flow. The important feature of the data is that the flux decline is exponential with time and an asymptotic equilibrium value is reached. Each solid curve drawn through the experimental points is a least-square fit exponential curve defined by Eq. (19). It is interesting to note that Merten et al ( ) in 1966 had observed an exponential flux decay in their reverse osmosis experiments. However, Thomas and his co-workers in their later experiments reported an algebraic flux decay with time (4,5).

The reverse osmosis membranes were tested in the standard experimental set-up (10). The experiments were carried out at three different pressures 17.4, 40.8 and 102 bars the corresponding sodium chloride concentrations were 3500 ppm, 5000 ppm and 29000 ppm. Before the reverse osmosis runs, membranes were thermally shrunk for 10 minutes in water and subsequently pressurized at 15-20% higher pressures than those used during the reverse osmosis experiments. A feed flow rate of 400 ml/mln was used giving a mass transfer coefficient k = 40 x 10 cm/s on the high pressure side of the membrane. 

If an asymmetric membrane is reversed, and the reverse osmosis experiments are carried out, the degree of salt rejection will be quite different from the results obtained for the normal experiments. In this case, the salt rejection is given by ... 

In reverse osmosis both solvent and solute diffuse because of gradients in their chemical potentials. For the solvent there is no gradient of chemical potential at an osmotic pressure of x at applied pressures p greater than 7r, there is such a gradient that is proportional to the difference p — ir. To a first approximation, the gradient of the solute chemical potential is independent of p and depends on the difference between concentrations on opposite sides of the membrane. This leads to the result that the fraction of solute retained varies as [1 + const./(p — 7r)] 1. Verify that the following data for a reverse osmosis experiment with 0.1 M NaCl and a cellulose acetate membrane follow this relationship ... 

Reverse-Osmosis Experiments. All reverse-osmosis experiments were performed with continuous-flow cells. Each membrane was subjected to an initial pure water pressure of 2068 kPag (300 psig) for 2 h pure water was used as feed to minimize the compaction effect. The specifications of all the membranes in terms of the solute transport parameter [(Dam/ 6)Naci]> the pure water permeability constant (A), the separation, and the product rate (PR) are given in Table I. These were determined by Kimura-Sourirajan analysis (7) of experimental reverse-osmosis data with sodium chloride solution at a feed concentration of 0.06 m unless otherwise stated. All other reverse-osmosis experiments were carried out at laboratory temperature (23-25 °C), an operating pressure of 1724 kPag (250 psig), a feed concentration of 100 ppm, and a feed flow rate >400 cmVmin. The fraction solute separation (/) is defined as follows ... 

By inserting a hypodermic needle through the O-ring seal, this equipment was adapted to reverse osmosis experiments by using a cellophane doublet separated by a powder gap in place of the single cellophane used by Richards [14). After some difficulty with the reliability of the pressure system, fresh water was produced at the rate of 0.06 gallon per square foot per day. 

Paul and Paciotti [19] took this work a step further by measuring the flux of a liquid (hexane) through a membrane both in pervaporation experiments with atmospheric pressure on the feed side of the membrane and a vacuum on the permeate side, and in reverse osmosis experiments with liquid at elevated pressures on the feed side and at atmospheric pressure on the permeate side. The hexane flux obtained in these two sets of experiments is plotted in Figure 2.17 against the hexane concentration difference in the membrane (c o(m) — c,eimi). The concentrations, cio(m) and Cie(m), were calculated from Equations (2.26), (2.36) and (2.72). 

The aspect of hole filling by plasma deposition can be demonstrated by the transport characteristics of LCVD-prepared membranes. First, the porosity as porous media calculated from the gas permeability dependence on the applied pressure can be correlated to the salt rejection of the composite membrane as shown in Figure 34.13. The effective porosity s/, where s is the porosity and q is the tortuosity factor, is measured in dry state and may not directly correlate to the porosity of the membranes in wet state. The effective porosity of LCVD-prepared membranes was measured before the reverse osmosis experiment. The decrease of porosity (as porous media) is clearly reflected in the increase in salt rejection in reverse osmosis. 

The influence of different cross-linking reagents on the properties of the membranes was Investigated by reverse osmosis experiments. A procedure for preparing the membranes was devised that yielded membranes of medium retention of phenol against an aqueous phenol solution of 2 g/litre at pH 13. 

Several membranes with good values for the retention of organic and inorganic compounds were investigated by reverse osmosis experiments with the effluents from the memufacture of organic intermediate products. The composition of the effluents were complex with analysis as follows ... 

Reverse Osmosis Experiments. This work makes further use of reverse osmosis data already reported with respect to membranes made from cellulose acetate Eastman E-398 polymers ( >, 5, ). 

Reverse osmosis experiments were performed with 12 salts and four membranes and the results obtained are given in Table V. These results were used to obtain values for unknown quantities l NaCl I >M+ (Ej)y, and (Ai)x ... 

Reverse osmosis experiments were performed with the same membranes that were used previously at concentrations where the formation of ion pairs was significant. The values of (-AAG/RT) for the unassociated ions that were determined previously were used in eq. (10) with the values of ln(PAtl/K6) for the associated ion experiments to obtain the values of (-AAG/RT)jp presented in Table VII. 

These systems were designed for up to 20 MPa in nanofiltration and reverse osmosis experiments. The cell volume is 190 mL. The magnetic stirrer is purchased from Amicon. The drawing is shown in Figure A2.3. 

The detail of the permeation cell for reverse osmosis experiments and the flow system arc shown schematically in Figure 3.9. The permeation cell is made of stainless steel 310 and consists of two detachable parts. The upper part is a high-pressure chamber. A wet membrane is mounted on a stainless steel porous plate embedded in the lower part of the cell such that the active surface layer of the asymmeU-ic membrane faces the feed solution under high pressure. A wet Whatman filter paper is placed between the membrane and the porous plate to protect the membrane from abrasion. The feed solution is supplied to the feed chamber of the permeation cell by a pressure pump, while the permeate... 

Calculate the separation of Na" " (1) Cl (2). and NO (4) ions from the reverse osmosis experiment involving the feed NaCl-NaNO.i. solution of 0.250 (NaCI) and 0.789 (NaNOi) molal at the operating pressure of 10,342 kPa (gauge). Use the following numerical values for the calculation ... 

The above equations arc exactly the same as those for pervaporation. Therefore, according to the above approach, the flux and the separation factor of a reverse osmosis experiment should approach those of pervaporation (with zero permeate pressure) when the operating pressure approaches infinity. This is because the second term in the flux equation (contribution of the permeant activity to the driving force) becomes negligible when the pressure on the feed side becomes infinity. As for pervaporation, the second term in the flux equation becomes zero when the pressure on the permeate side becomes zero. It should be noted that this conclusion is valid only on the basis of the assumption i.e., the pressure is equal to the feed pressure across the membrane. 

In a reverse osmosis experiment using a flat piece of membrane under the condition of no concentration... 

A comparison was made between sulfonated PPO (5 microns thickness, lEC 2.4 cast from 2/1 chloroform/methanol mixture), ultra-thin membrane of cellulose methylsulfonate o-propyl sulfonic acid and asymmetric cellulose acetate membrane in the treatment of alkaline copper cyanide feed. The reverse osmosis experiments were conducted in the following five steps. 

Nurlaila prepared SPPO TFC membranes by coating Desal E-500 substrate membranes with 0.5 or 1.0 wt % of SPPO (intrinsic viscosity of the base PPO polymer, 1.58 dL/g in chloroform lEC value, 1.87 meq/g). e membranes were then heat treated in a hot water bath, before being subjected to reverse osmosis experiments with NaCl or MgS04 solution. The results of RO experiments are given in Table 11. Comparison of the results from several series of experiments such as 1-4, 6-9, and 10-14, leads to a conclusion that the flux increases with heat treatment with little sacrifice in solute rejection. 

Mw of base PPO = 310 000, lEC value = 1.73 meq/g) solutions and coated on Desal E-500 membranes that were used as substrate membranes. Figure 8 shows results of reverse osmosis experiments using aqueous NaCl (8.6 mmol/L) and MgS04 (12.5 mmol/L) feed solutions. 

SPPOH with an lEC value of 1.93 meq/g polymer was dissolved in different solvents to prepare a 1 wt.% solution, and the solution was coated on the top-side of a PES ultrafiltration membrane (HW 18, supplied by Osmonics). The coating was repeated three times. The membranes were maintained in the hydrogen form (SPPOH) and were stored in distilled water until testing. The solvents used were chloroform/methanol mixtures with chloroform contents of 0, 18, 42 and 66 wt.%, respectively. Intrinsic viscosity of SPPOH in different solvent mixtures was measured at 25°C. Reverse osmosis experiments were performed using four composite membranes and three electrolyte solutions. The results for solute separation and flux are shown in Figs. 10 and 11, respectively. The flux of each... 

A solution of 1 wt. % SPPOH in methanol (PPO of intrinsic viscosity = 0.46 in chloroform at 25 °C, lEC = 2.0 meq/g) was coated on top of a commercial polyethersulfone substrate membrane (HW 17, supplied by Osmonics). Reverse osmosis experiments were carried out at 1000 kPa gauge in the presence of different electrolyte solutes. The results are summarized in Table 16. ... 

The pure water flux and solute separation data also depend on the metal cation, Me", when -SO3H is completely ion-exchanged to -S03 Me. To assure the complete exchange of-SO3H to -S03 Me, the membrane was immersed into 0.5N alkali hydroxide and washed with distilled water before use in reverse osmosis experiments. Table 18 shows the results of reverse osmosis experiments. The pure water flux and mixture water flux decreased, whereas the separation increased with an increase in the cationic radius from Li to... 

The observed trends in reverse osmosis experiments can be explained by the superimposition of the nature of a membrane with electric charges and that of a membrane with long alkyl chains. 

---