Grantee Research Project Results
Final Report: Environmentally Benign Synthesis Of Sodium Hydroxide Without Chlorine Using Ion Exchange Fibers
EPA Grant Number: R831433Title: Environmentally Benign Synthesis Of Sodium Hydroxide Without Chlorine Using Ion Exchange Fibers
Investigators: Sengupta, Arup K. , Warner, Steven B. , Munley, Vincent G. , Sengupta, Sukalyan
Institution: Lehigh University , University of Massachusetts - Dartmouth
EPA Project Officer: Richards, April
Project Period: October 15, 2003 through December 4, 2007
Project Amount: $319,998
RFA: Technology for a Sustainable Environment (2003) RFA Text | Recipients Lists
Research Category: Nanotechnology , Pollution Prevention/Sustainable Development , Sustainable and Healthy Communities
Objective:
Currently, the production of sodium hydroxide (NaOH) and chlorine (Cl2) are closely linked, and they are produced universally as co-products of electrolysis processes. As long as chlorine production remains coupled with the production of NaOH, it will be nearly impossible to promulgate regulations banning or reducing productions of various chlorinated compounds and enforcing them globally. The general objective of this research project is to synthesize NaOH without co-production of chlorine through an ecologically clean route. The specific objective of this research is to synthesize 4-6 percent NaOH using ion exchange (IX) fibers from sea water without producing chlorine.
Summary/Accomplishments (Outputs/Outcomes):
Characterization of the Fiber Material
Commercially available IX fibers have been procured from three separate sources (Actilex Fibers from Focus Polymers, United Kingdom; Smopex fibers from Johnston Matthey, Illinois; and Fiban Fibers from Unitechprom, Belarus). We procured adequate amounts of IX fibers in our laboratory and could purchase more as necessary. Initial investigations focused on a characterization of the fiber material. To this end, a sample of the weak acid IX fiber material was titrated with the strong base NaOH in an attempt to verify the overall capacity of the fiber. These findings were used to validate the manufacturer-suggested capacity of the material (4.5 meq/g) and are comparable to commercially available polymeric resins. Microscopic analysis was also used to determine the approximate dimensions of the material. Fiber materials were found to have a cylindrical shape with a diameter ranging from 1 to 50 μm and an average length of 3.7 cm. It should be noted that commercially available polymeric resins are spherical in form and have a significantly larger diameter (500 to 1000 μm).
Development of Binary Isotherm Data
Binary isotherm data for ions most significant to this process were developed through two basic methods. First, batch isotherm tests were conducted using strong-base IX fibers. These fibers are the key component in generating sodium bicarbonate from seawater. Binary data were collected regarding chloride (Cl-) and bicarbonate (HCO3-) ions. Batch tests were conducted using 200 mL containers at total aqueous phase bicarbonate concentrations of 5, 10, and 100 meq/L. Varying masses of Cl--loaded fiber (0.1-12.0 g) were placed in the batch containers and stirred for a period of 72 hours, until equilibrium had been achieved. The Cl-/HCO3- separation factor under the experimental condition is:
This separation factor in favor of chloride indicates that the regeneration of spent strong-base fibers using a bicarbonate solution is likely to be the least efficient portion of the described process. High concentrations of bicarbonate are likely to be required during for the regeneration phase to be most efficient.
The second set of binary selectivity data assessed the relative affinity of calcium (Ca2+) and hydrogen ion (H+) toward IX fibers with carboxylate functional groups, and their exchange kinetics are centrally responsible for the success of the proposed process. At near-neutral or under slightly alkaline conditions, calcium or other divalent ions are very selectively removed by IX fibers in the presence of sodium ions. Conversely, calcium ions can be efficiently eluted by increasing the hydrogen ion concentration in the aqueous phase (i.e., reducing pH). To determine relative Ca2+/H+ affinity for IX fibers, an equilibrium desorption test was carried out by sparging carbon dioxide in a batch reactor containing 10.0 gm of calcium-loaded IX fiber in 1.0 L of distilled water. The desorption process was quite rapid, and nearly 26 percent of the calcium was eluted from the fibers at an equilibrium pH of 5.6. The H+/Ca2+ separation factor under the experimental condition is:
A very high separation factor value in favor of hydrogen ion explains why a weak-acid gas, such as carbon dioxide, is an effective regenerant.
Otherwise identical experiments were carried out for ions that may enter the process stream in addition to calcium. These ions included sodium, magnesium, zinc, nickel, and copper. Selectivity data gathered during these experiments suggest the following selectivity sequence:
H+ > Cu2+ > Ni2+ > Zn2+ > Ca2+ > Mg2+ > Na+
For all cases, the hydrogen ion selectivity with respect to each of the above-mentioned ionic species was greater than unity. This indicates that hydrogen ion in the form of carbon dioxide-sparged snowmelt may serve as an efficient regenerant.
Regeneration of the Fiber Material Using Carbon Dioxide
This research concurrently examined the properties of both traditional and fibrous IX materials as exhibited during the regeneration cycle. During experimental regeneration cycles, the fiber and resin materials were placed in an epoxy coated 8 mm x 250 mm glass column. Carbon dioxide-sparged snowmelt served as the regenerant solution and was pumped from a pressurized reactor through the column using a liquid chromatography pump. Components were connected to the ion exchange column by 2.0 mm (inside diameter) Teflon tubing. For each regeneration cycle column run, 1 g of the fibers were weighed and packed into the column to a uniform height of 3.5 cm, thus ensuring a uniform bed resistance for each individual column run. The regeneration cycle was examined for the fiber material and compared to a similar column run using 0.45 g of the commercially available resin. The quantities of each material used ensured a similar overall exchange capacity for both materials. It should be noted that although physically dissimilar, chemically, the two ion exchange materials used were identical (e.g., they both contained carboxylic functional groups). Prior to use in the regeneration experiment, fiber and resin materials were placed in an ionic form containing only calcium. The effluent concentration was analyzed for calcium using a Perkin Elmer Atomic Absorption Spectrophotometer (Model AAnalyst 200) and a hollow cathode lamp. The effect of an increase in carbon dioxide partial pressure (above the regenerant solution) was examined for both materials. Partial pressures ranging from 1 psi to 8 psi were employed. Experimental data suggest that fiber materials, unlike resin materials, could be efficiently regenerated (> 95% recovery) using an insignificant amount of regenerant solution (< 10% of recovered product water). Additionally, this regeneration efficiency was seen to increase with corresponding increases in carbon dioxide partial pressure. The identical experimental configurations also suggest that fiber materials present a suitable alternative to traditional resin ion exchange materials without significantly altering or compromising the operational characteristics associated with these traditional ion exchange processes.
Evidence of a Difference in the Kinetic Properties of Fiber and Resin Ion Exchange Materials
Evidence supporting a difference in desorption kinetics could be seen in two distinct ways during the preceding experiments. First, kinetic differences are suggested in the high regeneration efficiency of fiber materials as compared to their resin counterparts. Second, although ion exchange materials were being placed in the initial ionic form (calcium) prior to regeneration, evidence of kinetic differences were suggested by the resulting effluent histories. The two materials were placed in the desired initial ionic form by passing a calcium solution (25 mg/L) through the glass column containing the ion exchange materials. The column effluent was examined to determine the extent of conversion. During this process involving each of the two materials, as the calcium gradually started exiting the column, the influent flow was deliberately discontinued for 24 hours. When the flow was resumed, the effluent of each column was examined for changes in the calcium concentration. Little change could be seen in the calcium concentration exiting the fiber column. For the resin, however, a significant drop (> 20%) in the calcium effluent concentration could be observed. Following the passage of several hundred-bed volumes of influent solution after this restart, calcium eventually reached the concentration prior to interruption. For intraparticle diffusion controlled processes, the concentration gradient within the sorbent particle serves as the driving force and governs the overall rate. With the progress of any column run, this concentration gradient attenuates. The interruption allows the sorbed calcium to spread out evenly within the spherical bead. As a result, the concentration gradient, and thus the uptake rate immediately after the column restart, is greater than the uptake rate prior to the interruption. In other words, a faster uptake and a consequent drop in the aqueous-phase exit concentration of the solute is (as evidenced by experimental data) confirmatory evidence in support of intraparticle diffusion as the primary rate-limiting step for resin materials. For the fiber materials, this effect was absent, indicating that intraparticle diffusion is not the primary rate-limiting step.
Intraparticle Versus Film Diffusion as the Rate-Limiting Step for Fiber and Resin Ion Exchange Materials
Experimental data were collected using a shallow bed column with effluent recycle to evaluate the kinetic properties of the two materials. A solution of dilute hydrochloric acid (0.01 M) was passed through fixed bed containing 0.2 g of each material in a calcium-loaded form. The effluent resivor was analyzed for calcium at various time intervals over a 30-minute time period. The resulting experimental data plot was given as the fractional attainment of equilibrium as a function of time. The predicted desorption profiles for both materials were then obtained using a film diffusion control (FDC) model and plotted for comparison. Pertinent hydrodynamic conditions used for both of the materials were identical. Comparison of predicted and experimental data indicates that the FDC model can accurately predict the desorption kinetics of the fiber material given different hydrodynamic conditions. Alternatively, it was seen that the FDC model cannot be used to predict the desorption kinetics of the resin materials. These results provide confirmatory evidence illustrating the differences between the rate-limiting steps governing each material. Because it is well recognized that sorption/desorption kinetics of spherical ion exchanger resin (beads) are often limited by intraparticle diffusion, a second model was developed to explain the desorption kinetics of the resin material. The predicted desorption profile for the resin material was obtained a second time using a Intraparticle Diffusion model and plotted. Comparison of predicted and experimental data indicates that the Intraparticle Diffusion model can accurately predict the desorption kinetics of the resin material given different hydrodynamic conditions.
Desorption Kinetics of Carbon Dioxide Regeneration
Because the preferred regenerant for the proposed processes is carbon dioxide as opposed to dilute hydrochloric acid, kinetic tests were next performed using carbon dioxide (100 psi) as the regenerant under similar experimental conditions. Desorption experiments were conducted using both fiber and resin materials under various hydrodynamic conditions. Experimentally, this was nearly identical to the previous kinetic experiments with the use of carbon dioxide being the only significant difference. Experimental results were recorded and modeled according to the proposed kinetic mechanisms. Comparison of predicted and experimental data indicates again that intraparticle diffusion is the rate-limiting step for resin materials, whereas film diffusion is the rate-limiting step for fiber materials.
Evidence of shrinking Beads and the Subsequent Effect on Efficiency of Carbon Dioxide Regeneration
One of the striking findings of previous kinetic investigations was that although the IX fibers were amenable to efficient regeneration with carbon dioxide-sparged snowmelt, commercial weak-acid ion exchange resins (C-104) responded poorly to the same regeneration process. Spherical resin beads and IX fibers are chemically similar; both have weak-acid carboxylate functional groups covalently attached to a polymer substrate. Their equilibrium properties are thus identical, and they exhibit high calcium removal capacity in the presence of competing sodium ions at near-neutral pH. However, because of the spherical geometry of the resin beads with sizes in the range of 400-1200 μm, the sorption kinetics is intraparticle diffusion controlled.
Light microscopy was used to evaluate shrinking tests for fiber and resin samples. Light microscopy was performed on a Westover Scientific Micromaster I binocular light microscope. Objective lenses of 10X and 40X magnifications were used for resin and fibrous materials, respectively. Fiber and resin samples in calcium forms were placed on a glass slide and covered with a cover slip. A small amount of water was added at the edges. To observe the effects of swelling and shrinkage, a few drops of 0.25 M HCl was added at the edges of the cover slip. All pictures of particles and measurements of particle diameters were made using a magnetic circular dichroism digital microscope head attached to the light microscope head and a laptop computer. Digital camera measurements were calibrated using calibration slides prior to data acquisition. All calibrations and measurements were made in the units of micrometers.
During the regeneration process, a spherical bead with a 490 μm diameter shrank fairly rapidly to 402 μm under the experimental conditions in less than 10 minutes. The rate of shrinking, however, gradually slowed down, and no significant shrinking was observed after 10 minutes. For an enhanced sensitivity, the weak-acid IX fiber chosen for the test under identical conditions had a relatively high cylindrical diameter (65 μm). Note that no significant change or trend in shrinking was observed for the fiber material during the course of the experiment.
To develop a mechanistic understanding of the poor regenerability of calcium-loaded spherical resin beads with carbon dioxide, let us consider a single bead as a polyelectrolyte gel with carboxylate functional groups. The affinity sequence for weak-acid carboxylate functional groups stands as follows: H+ >> Ca2+ > Na+. Uptake of H+ during regeneration by a weak-acid carboxylate group is essentially an association reaction leading to a major decrease in its osmotic pressure, thus causing expulsion of water from the gel phase. A spherical IX resin bead, therefore, gradually shrinks with the progress of regeneration through the uptake of hydrogen ions that involves counter-transport of H+ and Ca2+. At the onset, hydrogen ions would initially displace the outermost (i.e., peripheral) calcium ions. Such an exchange would, however, dramatically decrease the water content of the regenerated portion, thus decreasing the overall intraparticle diffusivity near the outer periphery of the resin bead. The progress of the regeneration process increases the depth of the relatively impervious skin, thus further slowing down the counter-transport of H+ and Ca2+. Scientifically, this hypothesis is in agreement with the sequence of photographs taken during the light microscopy test. Previous studies with weak-acid cation exchange resins also provided optical confirmation of shrunk cores during acid regeneration. For carbon dioxide regeneration, hydrogen ion concentration in the bulk phase cannot be as high as it is normally with mineral acid regeneration. Thus, the concentration gradient across the shrunk core is too small to overcome the diffusional resistance. The poor regenerability of resin beads with carbon dioxide is thus attributed to enhanced diffusional resistance offered by the shrunk peripheral layers with very low water content.
To the contrary, the IX sites for fibers reside primarily on the surface and the phenomenon of intraparticle diffusion, as demonstrated earlier, is of lesser significance. Protonation of weak-acid functional groups has only marginal impact on diffusional resistance and, hence, the carbon dioxide regeneration is efficient for IX fibers.
Laboratory Scale Generation of Sodium Hydroxide
A laboratory-scale unit was assembled and used in the preparation of NaOH. The prepared materials was then analyzed for purity regarding specific ionic constituents.
In order to better evaluate both the NaOH synthesis process and the difficulties encountered while using a fixed bed column, laboratory scale units of varying configurations (batch, fixed bed continuous, and fluidized bed) were assembled to produce sodium hydroxide. Batch reactors consisted of a 2.5L glass reagent container while fixed bed units consisted of a 300 mm or larger epoxy coated glass column that contained the fiber materials. For fixed bed processes, there were also several vessels containing the necessary reagents used as influent solutions. High purity sodium hydroxide could be produced using all three different configurations. It was also observed that pressure drops across a particular column due to swelling and slurry loading could be effectively avoided. This was accomplished in two ways: first, allowing the thickest portions of the slurry to settle prior to being passed through the column and second, using a configuration that provided a lower packing density of fiber materials.
Reverse Osmosis and Concentration of Product Stream to 2% NaOH
To this point, concentrations up to 1 percent have been successfully and efficiently generated using ion exchange fibers. Subsequently, higher concentrations (up to 2%) were generated using reverse osmosis to concentrate the original 1% solution. A experimental scale flat-leaf reverse osmosis (RO) unit was used to concentrate the NaOH solution.
RO membranes were obtained from the Dow Chemical Corporation. For sodium hydroxide concentration, a Sea-Water High salt Rejection membrane (SWHR) was utilized (model SW30HR). These membranes are typical to a single pass seawater desalination processes, and are commercially available. Maximum suggested operating pressure is 1000 psi. Membranes were delivered in a rolled sheet and then cut to the dimensions specified for the flat leaf RO unit used during the concentration process.
It is likely that concentrations of 4 to 6 percent NaOH product stream are possible given different RO process configurations. In particular, the use of a spirally wound membrane as opposed to the flat plate (laboratory scale) apparatus used for experimental purposes may significantly increase product concentration.
Limiting Conditions and Product Purity
During some fixed bed column runs, difficulties were encountered on account of two phenomena. Initial sodium loading of the fiber material resulted in significant swelling effects within the column. These observations during operation are consistent with experimental findings discussed earlier in this report. These swelling effects resulted in a pressure drop across the ion exchange column. The subsequent pumping of a lime slurry through this column during the NaOH generation phase further exacerbated the pressure drop across the column. Additional clogging was observed within the tubing and valve apparatus. These effects were not seen when employing a batch or fluidized bed configuration.
For all three configurations, concentrations up to 1% wt/vol NaOH were generated with relative ease. Concentrations were verified using both a sodium ion selective electrode (ISE) and titration. Experimental data suggests that concentrations beyond 1% cannot be produced in an efficient or feasible manner using ion exchange processes. This 1% limitation is on account of the low solubility of lime at exceedingly high pH as opposed to any ion exchange phenomena. It should be noted however; the NaOH solution can be further concentrated using technologies such as reverse osmosis (RO). To this effect, 1% NaOH was easily concentrated (on a laboratory scale) to 2% using a flat-leaf RO unit. Such concentrating processes had no effect on the overall purity of the NaOH solution.
In each instance the purity of the sodium hydroxide solution was extremely high. Ion chromatography and atomic absorption spectroscopy was used to measure the concentration of various ions including calcium, magnesium, chloride, sulfate, and others. For all significant ions, concentrations were exceedingly low or below analytical detection limits. For comparison, an identical analytical analysis was conducted using high-purity (1%) laboratory grade sodium hydroxide. No significant differences were seen between the two NaOH samples in terms of purity.
Rate Limiting Effects Associated With NaOH Production
For the caustic synthesis, the two competing rate-limiting steps are:
- Dissolution of milk of lime;
- Ion Exchange uptake of Ca2+
Maximum dissolution rate of lime is
(rD)max = kSAaSCSCa (1)
In which kSA is the specific rate constant (L×min-1×m-2), as is the specific surface area (m2/cm3) and CSCa is the saturation concentration of calcium (mole/L). Under a given hydrodynamic condition, the maximum rate of Ca2+ uptake on IX-fibers is liquid film diffusion controlled and given by
(rIX)max = kfafCCa (2)
In which kf is the liquid phase mass-transfer coefficient, af is the specific surface area of IX-fibers and CCa is the liquid-phase concentration of calcium. For a mass balance on the liquid phase calcium within a system, an overall reaction rate, ro can be written as a combination of equations (1 and 2).
r0 = (rD) - (rIX) (3)
Batch kinetic studies were performed using both lime and ion exchange fibers. Rates were independently determined for each of the potentially limiting steps. Further experiments studied the behavior when both rates were directly competing within a typical synthesis process. It was determined that under optimal conditions, the dissolution of lime was the rate limiting step and a pure (calcium-free) product could be produced. Process optimization was accomplished by changing the ratio of the slurry and ion exchange volumes (Vslurry/VIX). This optimal ratio was determined to be Vslurry/VIX ~ 1.0. It should also be noted that this condition is pertinent to both fixed bed and batch systems. Additionally, this ratio is within viable range of packing density and porosity associated with fiber materials
Economic Evaluation of the Process
Based on experimental results presented in the previous sections, the necessary process inputs for a full scale sodium hydroxide generation system were estimated. For economic comparison, the proposed process was considered both on this larger scale and at more modest scales. Important process parameters and inputs (flow rates, fiber quantities, chemical needs) were estimated in order to define the pertinent operating conditions capacity. Once the operating conditions were defined, costs could be assigned to the individual process components.
From these calculations, two points are noteworthy. First, the quantity of lime required for this process is substantial and will have a significant cost and operational impact on this process. Second, the volume of regenerant (alkalinity-free water) required is substantial with respect to any water treatment (or unit operation) process. For large-scale processes (~500 ton/day), these factors may be limiting. While, chemical-free regeneration will spare this process significant cost in terms of regeneration, the operational considerations may be unacceptable. For instance, trying to procure a production site where this volume of low-alkalinity water (rainwater or snowmelt) is readily available may be difficult.
When the annual cost is compared to the tons of caustic produced, it can be seen that the cost per ton is within the range of recent caustic soda market prices ($300-330 per ton). These costs show the relative feasibility of this process in terms of the major cost components. It should also be noted however, that moderate increases in the cost of production may be at least partially offset by the increased demand for high-purity sodium hydroxide.
When comparing this proposed process to existing electro-chemical processes, it can be seen that both are relatively capital intensive. Regarding production costs, for traditional electro-chemical processes, electricity is the most significant factor. For the proposed process, the electrical component is less, while the raw material/chemical component (lime) is significantly more. Therefore, the future viability of this process will most likely be dictated by the relative costs of electricity and lime with regard to their respective processes. In the near term, relatively inexpensive electricity favors the traditional process although future environmental regulations regarding chlorinated derivatives and chlorine demand in general may play a significant role.
Conclusions:
High purity sodium hydroxide was successfully generated using ion exchange fibers. In the production of NaOH no chlorine by-product was generated. Product concentration using the fixed bed process was limited to 1% wt./vol. NaOH on account of solubility constraints with respect to calcium hydroxide. Further increases in product concentrations were achieved through the use of reverse osmosis. Concentrating to 2% using reverse osmosis did not adversely affect product purity. It is also important to note that commercially available ion exchange resins were ineffective in generating sodium hydroxide due to the poor regenerability of resin materials using carbon dioxide. The process indicates economic feasibility as compared to existing processes. Because the process is not energy intensive, this feasibility may increase as electricity costs continue to rise. In the short-term, an optimal production scenario may be in the generation of small-volume, ultra-high purity, laboratory grade NaOH.
John Greenleaf, one of the researchers involved in this project, completed his Ph.D. in September 2007 from Lehigh University under the guidance of Professor Arup K. SenGupta. The dissertation was titled “Environmentally Benign Synthesis of Sodium Hydroxide and Hardness Removal Using Ion Exchange Fibers”. The project has also lead to an innovative approach to sequestering carbon dioxide while producing useful products including sodium hydroxide and hardness-free water.
Award and Presentation:
John Greenleaf presented the findings of this research at the 232nd national meeting of the American Chemical Society in San Francisco, CA. At this meeting he received the graduate student research paper award from the ACS Division of Environmental Chemistry for the above paper entitled “Environmentally Benign Hardness Removal Using Ion-Exchange Fibers and Snowmelt”.
Journal Articles on this Report : 3 Displayed | Download in RIS Format
Other project views: | All 6 publications | 3 publications in selected types | All 3 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Greenleaf JE, Lin J-C, SenGupta AK. Two novel applications of ion exchange fibers: arsenic removal and chemical-free softening of hard water. Environmental Progress 2006;25(4):300-311. |
R831433 (Final) |
Exit Exit Exit |
|
Greenleaf JE, SenGupta AK. Environmentally benign hardness removal using ion-exchange fibers and snowmelt. Environmental Science & Technology 2006;40(1):370-376. |
R831433 (Final) |
Exit Exit |
|
Greenleaf JE, SenGupta AK. Flue gas carbon dioxide sequestration during water softening with ion-exchange fibers. Journal of Environmental Engineering 2009;135(6):386-396. |
R831433 (Final) |
Exit |
Supplemental Keywords:
chlorine, sodium hydroxide, caustic soda, environmental risk reduction, ion exchange fibers, green house gas, carbon dioxide extraction, alternative chemical synthesis, INTERNATIONAL COOPERATION, Sustainable Industry/Business, Scientific Discipline, RFA, Technology for Sustainable Environment, Sustainable Environment, Chemical Engineering, Chemicals Management, Environmental Chemistry, chlorinated solvent reduction, environmentally-friendly chemical synthesis, green chemistry, alternative solvents, alternative chemical synthesis, alternative materials, ion exchange, carbon dioxide extraction,, RFA, Scientific Discipline, Sustainable Industry/Business, Sustainable Environment, Environmental Chemistry, Technology for Sustainable Environment, alternative materials, alternative solvents, carbon dioxide extraction, chlorinated solvent reduction, environmentally benign alternative, alternative chemical synthesis, environmentally-friendly chemical synthesis, green chemistryProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.