Concentration of Dilute Industrial Wastes by Direct Osmosis

The purpose of this research was to study the feasibility of using direct osmosis with sea water to concentrate dilute industrial wastes. several continuous flow laboratory size osmosis units were designed, constructed, and operated successfully. Dilute waste solutions were concentrated by direct osmosis using simulated sea water on the other side of the membrane. With the reverse osmosis membranes currently available, permeation rates were much lower than expected based upon their reported reverse osmosis rates. Another problem was that the diffusion rate of sodium chloride from the sea water to the waste solution and of the metallic ions from the waste solution to the sea water. were greater than could be tolerated in most applications. This method of concentrating waste solutions does not appear to be practical until more selective high flux membranes than are cu r rently available are developed. This method would be feasible if a suitable membrane were available. Membrane development was not within the scope of thi s i nvestigation.

There is a current need for more economical methods of treating industrial waste. If valuable products or raw materials can be recovered from the waste, the cost of the treatment will be partially offset by the value of these recovered materials (41,33). In some cases, the value of the materials recovered may even be greater than the treatment cost.
Often industrial wastes are in the form of very dilute aqueous solutions and large volumes must be handled.
These wastes would contain a relatively small amount of pollutant. For example, the rinse water used for washing nickel plated parts might contain only 500 milligrams per liter of nickel salts (41) . Wash waters from a photographic processing laboratory may contain 10 to 100 milligrams per liter of silver salts (41). Recovery of valuable salts from these very dilute solutions would be expensive and might not be practical. An inexpensive method of concentrating dilute solutions of industrial wastes would be very useful in that it would make the recovery of many valuable dissolved materials economically practical. Even if the polluting material is not to be recovered, concen- -2-trating the solution will greatly reduce the volume to be handled in other treatment methods and may result in a reduction in the total treatment cost.
In the past decade, there has been considerable interest in the reverse osmosis process as a method of concentrating wastes and in the recovery of relatively pure water (l, 16,17,20,26,27,33,38,41). Much recent work has been devoted to developing better reverse osmosis membranes and to reducing fouling of reverse osmosis membranes. In the reverse osmosis process, the solution is subject to a high pressure (100 to 600 psig) and relatively pure water flows through a semi-permeable membrane.
In the direct osmosis process, when two solutions are separated by a suitable semi-permeable membrane, nearly pure water flows from the less concentrated to the more concentrated solution. No pressure differential is needed across the membrane. The need for a large pressure differential across the membrane in reverse osmosis requires that the equipment be constructed to withstand this high pressure. Also, the membranes, which are usually thin plastic film, must be supported by some strong porous backing material. This backing material often reduces the flow rate through the membrane. Another problem encountered in the reverse osmosis process is the gradual reduction in the permeation rate through the membrane. This reduction in flow rate is attributed to the compaction of the membrane due to the high pressures.
Dilute solutions can be concentrated by direct osmosis at atmospheric pressure without the need for a pressure differential across the membrane.
If the waste solution is separated by a suitable membrane from another more concentrated solution whose water osmotic pressure is less than that of the waste solution, water will flow from the dilute waste solution to the concentrated solution. Actually the factor governing the direction of flow is not the concentration of the solution but its osmotic pressure.
In those locations near the ocean, where sea water is available, it would be suitable for use as the concentrated solution. If a desalination plant is close by, the brine from this plant would be an even better source of a concentrated solution as its osmotic pressure would be even lower than sea water. A by~product benefit of using brine would be the dilution of the brine before it is discharged back into the ocean. If a proper membrane is used the only effect on the sea water or brine will be dilution, as most of the pollutant in the waste water should not pass through the membrane. Both solutions were at atmospheric pressure. As expected, water passed through the membrane from the sugar solution to the brine. The sugar solution was concentrated from 1.5 to 2.3% sucrose and the brine was diluted from approximately 3.85% to 3.1% equivalent sodium chloride. The water flow rate through 2 the membrane was 1.23 gal/ft /day. rt was the purpose of this research to make a study of the feasibility of using the direct osmosis process with sea water to concentrate dilute industrial wastes. The membranes used in this investigation were limited to those commercially available reverse osmosis membranes. Preliminary tests were made using distilled water as the waste solution in order to study the membrane's rejection of sodium chloride, but later simulated metallic wastes as well as an actual industrial waste were tested. The concentrating solution was limited to sea water.
The variables studied in addition to the different membranes and waste solutions were: 1. The flow rates of both sea · water and the waste solution through the osmosis unit.
2. Concentration of the was.te solution.
3. The rejection of the solute and of sodium chloride in the sea water by the different membranes.  There has been an interest in the permeation of liquids through membranes as early as 1831 (35 In another investigation by the same men in the same year (7) , the effect of the presence of different concentrations of acids and bases upon the osmosis of chloride solu-t d 1 .ed The obJ'ect of the study was to test the · ns was s u · tiO · that by altering the sign of the charge of the hypothesis membrane (by having acids and bases present) , the osmotic effects may be greatly altered.
The results show that the presence of acid or alkali not only may alter the electrical sign of the capillary wall system, but also may alter, or even reverse the electrical sign of the membrane system.
The direction of the osmosis and its magnitude are closely related to electrical orientation of the cell system. Abnormal osmosis depends on the electrical orientation of the membrane system and the electrical orientation of the capillary wall.
Kahlenberg (24)     It is convenient to picture a membrane as a jumble of polymer chains. The interstitial volume in a polymer through Which transferring species pass is the void spaces between polymer chain. In transfers through polymers with short interchain distances, the transferring species must often push polymer segments apart to slide past them. Highly ta lline or highly crosslinked polymers are of this type. crys other polymers with less interchain attraction have wider spaces between the polymer chains, or longer polymer segments that are more free to move aside.
The resistance to transfer through such polymers is lower than that through polymers with very high interchain attractive forces, or through polymers that are highly crystalline or highly crosslinked.
The selectivity of cellulose acetate reverse osmosis membranes stems from the following mechanism (17). The polymers must also be excellent film formers because even extremely tiny mechanical flaws in the film are enormously larger than the diameter of water like solvent molecules. Transfer of species through such highly organized tight membranes is similar to the previously mentioned transfer in which the moving species pushes aside the polymer strands. Therefore, the resistance to transfer is quite high. However, high fluxes through such materials have been achieved by making the effective thickness of the :::s::: where J 1 is the solvent flux expressed in gallons per square foot per day, K 1 is a membrane constant, A is the cross sectional area of the membrane, x is the membrane skin thickness, ~ P is the applied pressure, and 411' is the osmotic pressure differential across the membrane. The permeate quality is determined by the rate at which solute passes through the membrane, according to the following equation: where J 2 is the solute flux, k 2 is the solute distribution coefficient between the membrane and solution, D is the diffusivity of solute in the membrane, and Ci and c 0 are the concentration of solute in the feed and permeate, respectively ( 17) .
For this direct osmosis work, the following expressions d The P ermeate or flux, J, is defined as: where Q is the amount of liquid passing through the membrane during the time interval t. A is the cross sectional area of the membrane.
The salt flux, F, is given by the equation below: where C is the concentration of sodium chloride in the s dilute waste out of the osmosis cell in milligrams per liter, V is the volume of dilute waste leaving the cell Wo in liters, and t is the time interval of the run in hours.
A is the exposed membrane area in square feet.

Effect of Variables
The variables studied in addition to the different membranes and waste solutions were:

1.
The flow rates of both sea water and the waste solution through the osmosis unit. Originally several tests were made using no membrane support. Howev er, the thin films were so fle x ible that they  The rest of the membrane is very porous and its function is to support the dense surface layer of the active side.
Tempering the membranes at a high temperature increases the thickness of the dense surface layer.
In reverse osmosis, the solution to be concentrated is in contact with the active side. In direct osmosis, it was not apparent which solution, the sea water or the dilute waste solution, should be in contact with the active side. Runs were made with the active side toward both the sea water and toward the dilute Waste solution. A slightly higher permeation rate was ob-h the sea water next to the active side. The ·ned wit ta1 JDSjoritY 0 f the test runs were made this way. procedure -There were some preliminary steps required before running the tests. suff i cient amounts of sea wate r had to be prepared. This was done by mixing the correct amount of artificial sea salt with water. The totameters used to monitor the flow rates of the sea water and dilute waste inlet streams had to be calibrated.
If a dilute metallic waste was to be used, sufficient amounts of this waste had to be prepared.
Next, an osmosis cell had to be prepared. A film or membrane was cut to the desired size to fit the cell. The film was then ca re fully placed between the two h a l v es of the cell and, with the backing material and the rubber gaskets in place, the bolts at each corner of the cell were ti ghtened to seal the c e ll. The cell was then conne cted to the r e st Of the experimental equipment.
One burette was fill e d with sea water and anothe r bure tte filled wi t h dilute was te. The p ump s were s tar ted and the apparatus was allowed to run to check the cell for leakage. I f the re was no leakage f rom the cell, the apparatus was read y f or use . The permeation rate was determined from these measurements.
The sodium chloride concentration of the dilute waste streams entering and leaving the osmosis unit were both obtained in order to determine the sodium chloride flux through the membrane.
In runs in which actual dilute metallic wastes were used, the dilute metallic ion concentration of the sea water streams entering and leaving the cell and the dilute waste streams entering and leaving the osmosis unit were obtained.
These data allowed the determination of material balances for the metal ions.
The exposed membrane area was recorded for use in cal-CUlating both water and salt fluxes.
lc ulations ~ 1 of liquid passing through the membrane was The vo ume from the differences in the dilute waste streams determined t of the osmosis unit and the sea water streams in in and ou and out of the osmosis unit.
In order to reach a steady state, data were not recorded until a reasonable time had elapsed after the osmosis unit had began to run.
The permeation rate or flux for the liquid was calculated from Equation 1, where Q is the volume of liquid passing through the membrane in time t and A is the exposed area of the membrane. The flux was expressed in gallons per square foot per day.
The rate of salt permeation, S, from the sea water through the membrane into the dilute waste solution is given by the following equation: ( 4) where C 8 is the concentration of sodium chloride in the dilute waste out of the 11 · · 11' l' t ce in mi igrams per i er, v Wo is the volume of dilute waste leaving the cell in liters, and t is the time interval of the run in hours.
The units are milligrams per hour. -37- The salt flux, F, is given by the equation below: F == S/A ( 5) where s is the salt rate in milligrams per hour and A is the exposed area of the membrane in square feet. The salt flux is expressed as milligrams per hour per square foot.
The relative water to salt flux is given by the following equation: where Q is the volume of liquid passing through the membrane in time t and S is the rate of salt permeation through the membrane. The relative water to salt flux is dimensionless.
Sample calculations of all types appear in the appendix.  In an attempt to limit the passage of sodium chloride h me mbrane the KP-98 membrane was tempered and through t e ' the liquid and salt fluxes were studied. r ed at four different temperatures. tempe were four minutes.       Tests were then run to see if any of the chromium was passing through the membrane into the sea water.  since there was a substantial amount of chromium in the sea water out in the preceding set of runs, four runs were made to calculate the amount of chronium in all four streams entering and leaving the osmosis unit. All of these runs were made with sea water on the active side of the membrane.  membrane runs gave no osmosis rate. Therefore, further tests 'these with this membrane were not conducted. The data for these runs (runs 40 and 41) are found in the appendix.
The effect of tempering temperature on the KP-00 membrane was studied in the next series of runs. Tempering was aone in water for four minutes. In all of these runs, the sea water on the active side of the membrane and distilled water was used as the waste solution.   The effect of osmotic pressure on the permeation rate S tudied in this series of runs. In the first two .,as runs ( 7 5 and 76) , sea water was on the active side and disw ater on the other side of the membrane. In runs tilled 78, brine made of 50% sea water and 50% distilled 77 and water was used on the active side of the membrane.  An actual waste wash water from a fish and shellfish processing plant was concentrated. The KP-90 membrane was used with the active side toward the sea water solution.  Table XIX.   .. The effect of tempering on the KP-00 membrane was tudied. Figure 4 shows the effect of tempering temperature the permeation rate. The permeation rate reached a 'PaXirnum at a tempering temperature of approximately so 0 c.
'the manufacturer (19) found that the permeation rate deincreasing temperature of tempering when the used for reverse osmosis with an 0.5 percent chloride solution and an external applied pressure T he shape of the curve of Figure 4 is the repsi.
h product of two effects. Refer to the equation, 1Glt of t e ......, . h can be rewritten for direct osmosis as, wu1C .
is the water flux in gallons per square foot per Kl is a membrane constant, A is the cross sectional of the membrane, x is the membrane skin thickness, 6'1T' is the effective osmotic pressure differential.
The osmotic pressure reaches its maximum value as the approaches ideal semi-permeability. The effective PlOtic pressure increased with tempering temperature. The The resistance term, K 1 /x, therefore, deas the tempering temperature increases. It is the of these two effects which leads to the results of and Figure 4.

a PP
Based on the results obtained, the reverse osmosis membranes tested do not appear to behave in the same manner dl. ·rect osmosis as they do for reverse osmosis applicafor t ions. The generally accepted mechanism described in Chapter II does not appear to be applicable at the lower pressures used during direct osmosis. Heat treating, or tempering, the membrane did reduce the salt flux but the oamosis rate (water flux) was also reduced. The low perm-1ation rates and high salt fluxes indicate a different chanism for direct osmosis with reverse osmosis membranes.
In summary, these reverse osmosis membranes do not beas expected when used for direct osmosis.
The concenration of industrial wastes by direct osmosis using exist -9 reverse osmosis membranes does not appear to be feasible sed on the results presented here.