WO2010033817A2

WO2010033817A2 - Crude glycerol purification process

Info

Publication number: WO2010033817A2
Application number: PCT/US2009/057497
Authority: WO
Inventors: John Haselow
Original assignee: Redox Tech, Llc
Priority date: 2008-09-18
Filing date: 2009-09-18
Publication date: 2010-03-25
Also published as: WO2010033817A3

Abstract

Processes for producing relatively salt-free glycerol from the crude glycerol by-product from biodiesel production are disclosed. The processes involve diluting crude glycerol to a concentration between about 30% and 70% water by volume, subjecting the resulting glycerol/water solution to ultrafiltration conditions, and subjecting the ultrafiltered glycerol/water solution to electrodeionization conditions. The resulting de-salted glycerol/water solution can optionally be passed through adsorbent charcoal to remove colored impurities. The water can then be removed from the de-salted glycerol/water solution. The resulting de-salted glycerol is typically greater than 90% pure, ideally, greater than 95% pure, and. preferably, greater than 99% pure.

Description

CRUDE GLYCEROL PURIFICATION PROCESS

FIELD OF THE INVENTION

The present invention relates to process for producing purified glycerol from the crude glycerol fraction produced during biodiesel production.

BACKGROUND OF THE INVENTION

Biodiesel fuel technology is being developed throughout the world as a way to decrease reliance on crude oil, to boost local economies, and to reduce carbon dioxide emissions by using a renewable fuel source. Biodiesel fuel is comprised of methyl and/or ethyl esters of fatty acids, and is typically derived by the transesterification of vegetable oils and/or animal fats with methanol or ethanol, catalyzed by a base such as potassium hydroxide. The resulting by-product, crude glycerol, includes the potassium hydroxide as well as any water present in the potassium hydroxide.

Any fatty acids present in the triglycerides must be neutralized with the basic catalyst. As a result, if the triglycerides include even one percent fatty acids, the amount of salt and water in the glycerol increases significantly. Because glycerol is only about ten percent of the reaction volume, an additional one percent of potassium hydroxide significantly increases the amount of salt and water in the crude glycerol product, relative to the use of one percent by volume of the basic catalyst.

The crude glycerol by-product has a very low value, and biodiesel producers have actually paid to have the by-product removed from their facilities. However, pure glycerol has significant market value, particularly in the pharmaceutical industry. It would be advantageous to provide a process for purifying the crude glycerol. The present invention provides such a process.

SUMMARY OF THE INVENTION

Processes for producing relatively pure glycerol from the crude glycerol fraction resulting from biodiesel production are disclosed. Purified glycerol resulting from this process, and pharmaceutical compositions comprising this purified glycerol, are also disclosed. Biodiesel production typically begins with triglycerides as the starting material.

Any triglyceride can be used that provides a biodiesel fuel composition with desired properties, including vegetable oils and fats and animal oils and fats. Any vegetable oil or animal fat can be used, and in one embodiment, the triglycerides comprise between approximately 0.1 and 10 percent fatty acids by weight.

The triglycerides are reacted with an alcohol, ideally methanol and/or ethanol, to form fatty acid methyl and/or ethyl esters. The transesterification reaction is catalyzed with a soluble basic catalyst, such as sodium hydroxide, sodium methoxide, and the like The transesterification reaction is typically run at a temperature between around 7O⁰C and the reflux temperature of the alcohol.

After the transesterification reaction is complete, the reaction mixture separates into two separate layers. The top layer is the biodiesel product, and the bottom layer is the crude glycerol fraction, which includes fatty acid carboxylate salts, alkoxide salts, water, and glycerol. If an alkoxide catalyst, such as potassium or sodium methoxide, ethoxide, or t- butoxide is used instead of hydroxide, the glycerol will include significantly less water.

The crude glycerol fraction can be reacted with acid, such as hydrochloric, sulfuric, or phosphoric acid, to convert the alkoxide salts to the corresponding alcohol, and the fatty acid carboxylate salts to the corresponding free fatty acids. The free fatty acids can be removed, for example, by extraction or decantation, and the methanol, ethanol, or other alcohol removed by distillation. The resulting glycerol is relatively pure, including predominantly glycerol and salt. The salt includes the cation of the base used to conduct the transesterification reaction, and the anion of the acid used to neutralize the basic catalyst. If the neutralization is carried out carefully, there will not be a large excess of acid present.

If hydrochloric acid is used to neutralize the basic catalyst, the remaining chloride salt might cause damage to certain plant equipment (i.e., stainless steel reactors).

Accordingly, it can be advantageous to use other acids, such as sulfuric, phosphoric or organic acids.

The process uses electrodialysis to remove the salt from the glycerol. The resulting glycerol is relatively pure, and can be used, for example, as an excipient in the pharmaceutical industry.

The process involves taking a glycerol composition that includes a salt impurity, such as that obtained from biodiesel production, and diluting the glycerol with water to a concentration of between around 10 and 70% glycerol by volume, more preferably between around 40 and 60% glycerol by volume, most preferably around 50% glycerol by volume.

The dilution of the glycerol reduces the viscosity and improves the mobility of ions such that it is now able to be processed at the desalination step. The desalination step comprises electrodionization (EDI), typically through an EDI stack.

An EDI stack has the basic structure of a deionization chamber. The chamber contains a ion exchange resin, packed between a cationic exchange membrane and an anionic exchange membrane. Only the ions can pass through the membrane, and the glycerol/water mixture is blocked. The resulting glycerol/water mixture is now desalted and ready for further processing steps

Before the water is removed, colored impurities can optionally be removed by passing the glycerol/water fraction through a adsorbent charcoal, whether in the form of a packed column, packed bed, or other suitable means for removing colored impurities.

For example, one can use a plurality of upstanding adsorption columns, each of which includes an upstanding tubular housing that supports a vertical adsorption bed containing clay granular material and activated charcoal in granular form. The glycerol/water mixture can be flowed downard through passages formed through the intergranular spaces. After this optional step of removing colored impurities, or directly after the desalination, the glycerol/water mixture can be treated to remove the bulk of the water, arriving at a glycerol composition that is greater than 90% pure, preferably greater than 95% pure, more preferably, greater than 98 % pure, and most preferably, greater than 99% pure.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is a process flow diagram of one embodiment of the process described herein.

DETAILED DESCRIPTION OF THE INVENTION Processes for purifying the crude glycerol fraction from biodiesel production are disclosed.

In some embodiments, the processes described herein are integrated processes. As used herein, the term "integrated process" refers to a process which involves a sequence of steps, some of which may be parallel to other steps in the process, but which are interrelated or somehow dependent upon either earlier or later steps in the total process.

The invention will be better understood with reference to the following detailed description.

I. Biodiesel Fuel Production Biodiesel fuel is typically prepared by the transesterification of a triglyceride and an alcohol, using a hydroxide or alkoxide catalyst. If the triglyceride has a relatively high free fatty acid content, one can optionally perform an acid-catalyzed esterification step to react the free fatty acids with an alcohol, neutralize the acid, and then perform the base-catalysed transesterification reaction. This additional process step allows one to use relatively inexpensive feedstocks, such as trap grease, yellow grease, used cooking oil, and off-spec vegetable oil and animal fat, but can produce additional salt as the acid catalyst is neutralized. However, the process described herein removes this salt, thus not only providing a relatively pure glycerol fraction, but also allowing for the use of less expensive raw materials. A more detailed description of the components used to produce a biodiesel fuel is provided below.

A. Triglycerides

Vegetable oils are mostly comprised of triglycerides, which are triesters of glycerol, CH₂ (OH)CH(OH)CH₂ (OH), and fatty acids. Fatty acids are, in turn, aliphatic compounds containing 4 to 24 carbon atoms and having a terminal carboxyl group. Diglycerides are esters of glycerol and two fatty acids, and monoglycerides are esters of glycerol and one fatty acid. Naturally occurring fatty acids, with only minor exceptions, have an even number of carbon atoms and, if any unsaturation is present, the first double bond is generally located between the ninth and tenth carbon atoms. The characteristics of the triglyceride are influenced by the nature of their fatty acid residues.

Any source of triglycerides can be used to prepare the biodiesel product that provides a biodiesel product with the desired properties, including vegetable oils and fats and animal oils and fats. Examples of suitable vegetable oils include, but are not limited to, crude or refined soybean, corn, coconut (including copra), palm, rapeseed, cotton and oils. Examples of suitable animal fats include, but are not limited to, chicken fat, pork fat, beef fat, tallow, lard, butter, bacon grease and yellow grease. Naturally-occurring fats and oils are the preferred source of triglycerides because of their abundance and renewability.

The processes described herein can tolerate a relatively higher FFA concentration than traditional biodiesel processes. Accordingly, used cooking oils, rancid oils, or any triglyceride feedstock with up to about 15% FFA content or more can be used.

B. Alcohols Any alcohol that provides a biodiesel product with the desired properties can be used to prepare the fatty acid alkyl esters. Typically, ethanol and/or methanol are used, although any saturated straight, branched, or cyclic C_1-6 alcohol can be used. Examples include methanol, ethanol, propanol, isopropanol, butanol, isobutanol, t-butanol, pentanol, cyclopentanol, isopentanol, neopentanol, hexanol, isohexanol, cyclohexanol, 3- methylpentanol, 2,2-dimethylbutanol, and 2,3-dimethylbutanol.

It is preferred that any alcohol used in the present invention contains less than five percent water, preferably less than approximately one percent water, to avoid saponification or hydrolysis of the triglycerides and to minimize the amount of water present in the glycerol by-product.

The fatty acid alkyl esters are preferably methyl esters, ethyl esters, or combinations thereof. Blends of ethyl and methyl esters are slightly less expensive and can perform nearly as well in biodiesel fuel as pure ethyl esters, and have lower melting points, albeit with the limitation of additional toxicity.

C. Acid- Catalyzed Esterification

Should an initial esterification step be required to reduce the free fatty acid content, any acid catalyst that is suitable for performing esterification reactions can be used, in any effective amount and any effective concentration. Examples of suitable acids include, but are not limited to, hydrochloric acid, hydrobromic acid, sulfuric acid, phosphoric acid, and solid catalysts such as Dowex 50™. Strong acids, such as sulfuric acid, are preferred catalysts. In one embodiment, approximately one cup of concentrated sulfuric acid is added per ten gallons of oil. The reaction mixture is heated, for example, to the reflux temperature of the alcohol, until the fatty acids are esterified. When an acid catalyst is used to esterify FFAs present in the triglycerides, the acid typically must be removed before the transesterification of the triglycerides is performed using a basic catalyst, or additional catalyst must be used. For this reason, it can be preferred to use a solid acid catalyst, as these can be readily separated from the triglycerides (i e., via decantation or filtration).

D. Base- Catalyzed Transesterification

The triglycerides can be transesterified to form fatty acid esters and glycerol. Methods for performing transesterification reactions are well known in the art, as they are the principal method by which biodiesel is formed. The transesterification reactions generally go to completion in approximately six to twenty four hours, and can be run in both batch-type and continuous reactors. Reaction conditions for transesterification reactions are known to those of skill in the art. As discussed above, biodiesel production can include the optional step of esterifying any FFAs that are present before the transesterification step. The transesterification ideally can use an alkoxide or hydroxide catalyst. If sodium hydroxide or another metal hydroxide is used as a catalyst, it will react with any fatty acids to form water and the corresponding carboxylate salt, and with an alcohol to form water and an alkoxide salt. If an alkoxide is used, and the alcohol starting material is relatively dry, one can minimize the water content in the reaction mixture, since the alkoxide will react with any fatty acids which are present to form the alcohol and the corresponding carboxylate salt, but will not form water. The glycerol thus produced will be drier than that produced where the water content of the triglyceride, alcohol, and basic catalyst is not taken into consideration.

E. Product Isolation When the transesterification reaction is completed, the reaction mixture is allowed to separate into two phases. The top phase is the biodiesel product, and the bottom phase is a crude glycerol by-product. The crude glycerol includes various impurities, including fatty acid carboxylate salts and other organic matter (commonly referred to as MONG, sometimes translated as '"matter organic, not glycerol."). Other impurities include water, alkoxide salts, and unreacted alcohol. In a typical crude glycerol waste stream, the glycerol is between about 40 and about 60 percent pure at this stage.

Biodiesel is separated from the glycerol/alcohol/alkoxide mixture, for example, by decantation or by dropping the waste glycerol fraction from the bottom of the reactor. Excess alcohol, used to ensure that the transesterification goes to completion, can be isolated by distillation and can be re-used, either before or after the alkoxide catalyst is neutralized (to form the alcohol).

The glycerol/alkoxide mixture can be treated with acid, preferably a dry acid, such as hydrogen chloride gas, phosphoric acid, or sulfuric acid. Ideally, no more acid is used than is necessary to neutralize the alkoxide and to protonate the fatty acid carboxylate salts.

The resulting free fatty acid can be isolated, for example, by decantation or extraction. Once the fatty acid has been removed, the glycerol fraction typically includes water, methanol (or other alcohols), and salts resulting from the neutralization of the fatty acid carboxylate salts and the alkoxide and/or hydroxide salts. Ideally, the amount of acid added is carefully titrated so that there is not a significant excess of acid. The alcohol (and any water that is present) can be removed by distillation, and the alcohols recycled to the biodiesel reaction.

Since the resulting partially-purified glycerol is essentially free of the alcohol and MONG, all that remains is to efficiently remove the salts that are present. In some embodiments, a portion of the salt can be removed by filtration or decantation. However, given the viscosity of glycerol, and the fact that the typical salts that are present are somewhat soluble in glycerol, further processing steps are required to produce an essentially salt-free glycerol product. The partially-purified crude glycerol fraction is then desalinated using the process described herein to produce a glycerol fraction with a lower salt concentration, typically less than 100 ppm, preferably less than 70 ppm, more preferably, less than 50 ppm, and, most preferably, less than 10 ppm.

II. Removal of Salt from the Salt- Containing Glycerol Fraction

Salts can be substantially removed from the glycerol fraction by a process known as electrodeionization, which represents an improvement over prior attempts to purify glycerol using electrodialysis.

A. Dilution of the Glycerol Fraction

Because glycerol is very viscous, the first step in the process is to dilute the glycerol fraction to a concentration between about 70 and about 30% glycerol in water, preferably between about 60 and about 40%, and, most preferably, about 50%. At these concentrations, the glycerol can flow relatively freely through an EDI apparatus.

B. Ultra- Filtration of the Diluted Glycerol Fraction

The main difference between electrodeionization ("EDI"), as used in the process described herein, and electrodialysis, as used in the past, is that EDI has ion exchange resin between the anion and cation exchange resins. This is important because it improves removal of salts at low salt concentrations.

Electrodialysis becomes very slow and inefficient at salt concentrations less than -1000 ppm or so. Although conventional wisdom is that EDI will not work well with organic-laden process streams, the real problem appears to be organic ions rather than organics. The organic ions can foul the ion exchange resm. Accordingly, it can be important to remove fatty acids and fatty acid carboxylate salts using ultrafiltration. Ultrafiltration is not fundamentally different from reverse osmosis, microfiltration or nanofiltration, except in terms of the size of the molecules it retains. Ultrafiltration (UF) is a variety of membrane filtration in which hydrostatic pressure forces a liquid against a semipermeable membrane. Suspended solids and solutes of high molecular weight, in this case, free fatty acids or fatty acid carboxylate salts, are retained, while water and low molecular weight solutes pass through the membrane. Ultrafiltration processes are well known to those of skill in the art.

Once the glycerol fraction has been treated to remove fatty acid salts, alkoxide. and alcohol impurities, diluted with water to reduce its viscosity, and ultrafiltered to remove particulates such as residual fatty acid carboxylate salts and/or fatty acid particles, it is ready for electrodeionization.

C. Electrodeionization

Electrodeionization (EDI) is a process that removes ionizable species from liquids using ionically active media and an electrical potential to influence ion transport. EDI devices combine ion exchange resins, ion exchange membranes and electrodes. An EDI cell combines the benefits of ion exchange and electrodialysis while minimizing the problems associated with each of these separate technologies. Traditional ion exchange requires regeneration of the exchange media with acids and bases, which generates waste and takes significant time Electrodialysis, such as systems patented by Perry et al (US Patent 5,057,197) and more recently by Schmidt et al (US Patent 7,501,064) become less efficient as salt concentrations decrease because the fluid is becomes increasing less conductive as ions are removed. An EDI cell uses the ion exchange resin to provide high ionic conductivity to decrease the high resistance normally found in the diluate compartments of an electrodialysis cell. The resin's high ionic capacity increases the residence time of the ionic contaminants inside the cell allowing more time for the current to transport these ions into the appropriate compartments. The electrodes generate a potential gradient for ionic movement within the cell. At cation/anion (resin/resin and resin/membrane) interfaces water is dissociated into its constituent ions, H⁺ and OH^", which regenerate the resins on-line, so there is no down time or need for regenerative chemicals as in ion exchange.

The concept of electrodeionization has been extensively investigated since the mid- 1950's, for various purposes. Walters, et al. investigated a batch electrodeionization process for the concentration of radioactive aqueous wastes, based on electrolytic regeneration of ion exchange resins.

In the late 1950's and early 1960's, the theory, design, and operating conditions of the CEDI process were investigated by Glueckauf (Glueckauf, E., "Electro-deionisation Through a Packed Bed", British Chemical Engineering, December 1959, pp. 646-651). He proposed a theoretical model based on a two-stage removal of ions; diffusive transfer of ions from flowing solution to ion exchange resin beads and the transfer of ions along the chain of ion exchange beads. Sammons and Watts (Sammon, D. C, and Watts, R.E., "An Experimental Study of Electrodeionisation and its Application to the Treatment of Radioactive Wastes", AERE-R3137, Chemistry Division. U.K.A E.A. Research Group, Atomic Energy Research Establishment, Harwell, June 1960) studied the deionization of sodium salt solutions using multi-cell electrodeionization modules, quantifying the relationships between solution concentration, flow rates, and applied current. They experimentally demonstrated CEDI, but they did not define in great detail the effects of parameters such as liquid flow velocity, cell width, particle size, and type of resin filling.

The first patent for an electrodeionization device was granted in 1953. Verkeer et al described an apparatus for the deionization of salt-containing liquids using alternating layers of anion and cation resins. The patent was granted in 1957 (Verkeer, H.E. et al., "Process and Apparatus for the Electrolytic Deionization of Salt-Containing Liquids", UK Patent No 776,469, June 5, 1957). A patent was also granted to Kollsman in 1957 (Kollsman, P., "Method and Apparatus for treating ionic fluids by dialysis". US Patent No. 2,815,320, December 3, 1957), describing an apparatus for the purification of acetone. In this time period and in the 1960's, numerous patents were granted for various types of electrodeionization devices, including those to Tye (Tye, F.L., "Improvements relating to Electrodialysis Process", UK Patent No. 815,154, June 17, 1959.), Pearson (Pearson, R.G., "Electrolytic Deionization". US Patent # 2,794,777, June 4, 1957), Kressman (Kressman, T.R.E., "Process for the Removal of Dissolved Solids from Liquids", U.S. Patent #2,923,674, Feb. 2, 1960), and Parsi (Kressman, T.R.E., "Process for the Removal of Dissolved Solids from Liquids", U.S. Patent #2,923,674, Feb. 2, 1960). Matejka (Matejka, Z., "Continuous Production of High- Purity Water by Electro-deionisation", J. Appl. Chem. Biotechnol., 1971, Vol. 21, April, pp. 117-120) and Shaposhnik (Shaposhnik, V.A.. Reshetnikova, A.K., Zolotareva, R.I., LV. Drobysheva, and N.I. Isaev, "Demineralization of Water by Electrodialysis with Ion-exchanger Packing Between the Membranes", Zhumal Prikladnoi Khimii, Vol. 46, #12, pp. 2659-2663, December, 1973, translated by Consultants Bureau, Plenum Publishing Corp., 227 W. 17th St., New York) extended the investigation into the operating conditions and performance of the CEDI process in the 1970's and 1980's. Matejka investigated EDI for the deionization of brackish or tap water to produce high-purity water. Several researchers in Israel were also very actively studying electrodeionization, including Selegny (Selegny, E. and E. Korngold, "Method of Separation of Ions From a Solution", U.S. Patent # 3,686,089), Korngold (Korngold, E., "Electrodialysis processes using ion exchange resins between membranes", Desalination, 1975, Vol. 16, No. 2, pp. 223-233) and Kedem (Kedem, O.. "Reduction of polarization in electrodialysis by ion-conducting spacers", Desalination, 1975, Vol. 16, No. 1, pp. 105-118) During this time, new devices were being proposed and patents being granted to a number of others, including those to Davis, Tejeda, Kunz, and Giuffrida et al. ((Davis, T. A., "Electrically Regenerated Ion Exchange System" U.S. Patent # 4,032,452, June 28, 1977; Tejeda, R.A., "System For Demineralizing Water By Electrodialysis", US Patent # 3,869,376, Mar. 4, 1975; Kunz, G., "Process and Apparatus For Treatment of Fluids, Particularly Desalination of Aqueous Solutions", US Patent # 4,636,296, Jan 13, 1987; and Giuffrida, AJ. , Jha, A.D. and Ganzi, G.C., "Electro-deionization Apparatus", US Patent # 4,632,745, Dec. 30, 1986). The first commercially available continuous electrodeionization modules and systems were introduced in 1987 under the trade name Ionpure - now a U. S. Filter company ("U.S. Filter"). A more comprehensive review of the technical literature on electrodeionization is given by Ganzi et al. (Ganzi, G.C., Wood, J.H. and Griffin, CS. , "Water Purification and Recycling Using the CDI Process", Environmental Progress, Vol. 11, No. 1, February 1992). A recent review of CEDI technology was provided by Henley (Henley, M., "Technology Overview: CEDI Gains Ground as Water Treatment Approach," Ultrapure Water, pg 15, October 1997). It is believed that there are more than 1300 industrial size CEDI installations worldwide, about 95 % of which have been supplied by U.S. Filter. Applications have included pharmaceutical, electronics, power generation, food and beverage, and laboratory.

The EDI process employs anion and cation permeable ion exchange membranes, with ion exchange resins packed between them. Applying a DC electric potential causes ions to move from one compartment to another, effecting a separation. Because the concentration of ions is reduced in one compartment (diluate) and increased in the other (concentrate), the process can be used for either purification or concentration. Under the influence of the electric field, cations will migrate in the direction of the negatively charged cathode, through the cation-exchange resin, cation-permeable membrane and into the concentrating stream. An anion permeable membrane on the opposite side of that stream prevents further migration, effectively trapping the cations in the concentrating stream and allowing them to be flushed to drain. The process for anion removal is analogous, but in the opposite direction, toward the positively charged anode.

In continuous electrodeionization for the preparation of deionized water, the process operates in two regimes (Ganzi, G.C., "The Ionpure(TM) Continuous Deionization Process: Effect of Electrical Current Distribution on Performance", Presented at the 80th Annual AIChE meeting on Nov. 28, 1988, Washington, DC). In the first regime, at higher salinity or at the inlet portion of the resin bed, the resins in the diluting streams remain in the salt forms, and efficiencies are derived from the resin-enhanced electrical conductivity of the ion- depleting compartments. In the second regime, at low salinity or at the outlet portion of the resin bed, the DC electric potential causes water to dissociate into its constituent ions, H and OH , electrochemically converting the resins to the hydrogen and hydroxide forms. This phenomenon, known as electroregeneration, accounts for the ability of CEDI systems to produce multi-megohm water, much like a continuously regenerated mixed bed ion-exchange column.

When compared with conventional resin-based ion exchange, which need to be chemically regenerated, EDI offer a variety of benefits. Most obvious is the elimination of the regeneration process and its associated hazardous regeneration chemicals - acid and caustic. Since EDI operates through a combination of ion-transfer across the resins and membranes, as well as electrochemical regeneration of the outlet portion of the bed, the resins and membranes are never fully exhausted.

For the treatment of glycerin, EDI offers numerous advantages over electrodialysis, ion exchange, and the most commonly applied method of distillation. With EDI treatment of glycerin, the desalting process is completed at ambient temperatures, and with a glycerin concentration of 5 to 60 weight percent, but most preferably 40 weight percent. After salt is removed via EDI, the water is removed with a vacuum evaporation process at roughly 130 degrees Celsius Distillation requires the glycerin to be boiled at temperatures approach 270 C. The glycerin must then be condensed. The higher temperature distillation process is more energy intensive and can create thermal decomposition and polymerization byproducts. Also, a glycerin distillation processes leaves about 5 to 20 percent distillation bottoms that are rich in salt and other organic materials. EDI-processed glycerin does not generate a salt-laden waste glycerin stream. The electrical consumption of EDI desalting is roughly a tenth of the distillation process. In addition, the capital equipment and worker skill required for EDI is significantly less than distillation.

The EDI process described herein produces glycerol/water solutions of very high purity, using significantly less of the chemical products used in conventional ion exchange processes. That is, EDI system membranes and electricity can replace the significant volumes of acid and caustic chemicals that would otherwise be required using ion exchange. Further, the process permits one to desalinate to extremely low salt levels, for example, less than 100 ppm, preferably less than 70 ppm, more preferably, less than 50 ppm, and, most preferably, less than 10 ppm.

The EDI process combines semi-impermeable membrane technology with ion- exchange media to provide a high efficiency demineralization process. Electrodialysis uses electrical current and specially-prepared membranes which are semi permeable to ions based on their charge, electrical current, and the ability to reduce the ions based to their charge. Through electrodialysis, an electrical potential transports and segregates charged aqueous species. The electrical current is used to continuously regenerate the resin, minimizing the need for periodic catalyst regeneration. In use, the glycerol/water solution is passed through an EDI stack. An EDI stack has the basic structure of a deiomzation chamber. The chamber contains an ion exchange resin, packed between a cationic exchange membrane and a anionic exchange membrane. Thus, the electrodeionization (EDI) process uses a combination of ion-selective membranes and ion-exchange resins sandwiched ("stacked") between two electrodes (anode (+) and cathode (-)) under a DC voltage potential to remove ions from the glycerol/water solution. Only the ions can pass through the membrane, and the water is blocked.

By spacing alternating layers of anion- and cation- selective membranes within a plate-and- frame module, a "stack" of parallel purifying and concentrating compartments are created. The ion-selective membranes are typically fixed to an inert polymer frame, which is filled with mixed ion-exchange resins to form purifying chambers. Screens can be placed between the purifying chambers to form concentrating chambers.

In one embodiment, the EDI apparatus comprises a tube with an inlet port and an outlet port, and the glycerol/water solution passes through the inlet and out through the outlet. In between the inlet and outlet ports, the tube comprises various membranes and spacers. For example, the "sandwich" or "stack" includes alternating layers of a concentrate spacer and a dilute spacer, with alternating layers of an anion membrane and a cation membrane between the concentrate and dilute spacer layers, can be used. In one embodiment, the configuration of the "sandwich" is as follows: a concentrate spacer, an anion membrane, a dilute spacer, a cation membrane, a concentrate spacer, an anion membrane, a dilute spacer, and a cation membrane.

The glycerol/water solution flows into a resin-filled diluting compartment, setting in motion several different processes. Strong ions are scavenged out of the feed stream by the mixed bed resins. A strong direct current field is applied across the stack of components, pulling charged ions off the resin and drawing them towards the respective, oppositely- charged electrodes. Accordingly, these charged strong-ion species are continuously removed and transferred into the adjacent concentrating compartments.

Ion-selective membranes operate using the same principle and materials as ion- exchange resins, and they are used to transport specific ions away from their counter-ions. Anion- selective membranes are permeable to anions but not to cations; cation-selective membranes are permeable to cations but not to anions. The membranes are not water- permeable (or glycerol permeable).

As the ions travel towards the membrane, they can pass through a concentration chamber, but cannot reach the electrode because they are blocked by a contiguous membrane that contains a resin with the same charge.

As the strong ions are removed from the process stream, the conductivity of the stream becomes quite low. The strong, applied electrical potential splits water at the surface of the resin beads, producing hydrogen and hydroxyl ions. These act as continuous regenerating agents for the ion-exchange resin. The regenerated resins allow one to ionize neutral or weakly-ionized aqueous species, such as carbon dioxide or silica. Ionization is followed by removal through the direct current and the ion exchange membranes.

The ionization reactions occurring in the resin in hydrogen or hydroxide forms for the removal of weakly ionized compounds are listed below:

CO₂ + OH^" ==> HCO₃ ^"

HCO₃ ^" + OH^" ==> CO₃ ^2"

SiO₂ + OH^" ==> HSiO₃ ^"

H₃BO₃ + OH^" ==> B(OH)₄ ^"

NH₃ + H⁺ ==> NH₄ ⁺

EDI offers several advantages over traditional ion-exchange processes, including savings in both energy and operating expenses. A further environmental benefit is achieved by eliminating the need for periodic regeneration of the ion exchange resins, and the associated use of regenerating chemicals. It can be important to treat the glycerol/water stream to a pretreatment step, such as ultrafiltration, for the EDI process to function optimally. The process can be further improved by using a "strip gas" such as nitrogen to remove any carbon dioxide from the glycerol/water solution, since any ions formed from the carbon dioxide can lower the outlet resistivity of the glycerol/water solutions produced by the process described herein. D. De-Coloration of the De-Salted Glycerol Fraction

The resulting product can be relatively colored, and, depending on the end user requirements, it can be important to de-color the purified glycerol. This can be accomplished, for example, by passing the glycerol/water solution through adsorbent charcoal, whether in the form of a packed column, packed bed, or other suitable means for removing colored impurities. The decolorization step can be performed at various stages during the process, for example, before or after electrodeionization.

For example, one can use a plurality of upstanding adsorption columns, each of which includes an upstanding tubular housing that supports a vertical adsorption bed containing clay granular material and activated charcoal in granular form. The glycerol/water mixture can be flowed downard through passages formed through the intergranular spaces.

After this optional step of removing colored impurities, or directly after the desalination, the glycerol/water mixture can be treated to remove the bulk of the water, arriving at a glycerol composition that is greater than 90% pure, preferably greater than 95% pure, more preferably, greater than 98 % pure, and most preferably, greater than 99% pure.

E. Re-Cycling the Electrodeionization Apparatus

EDI offers another advantage over ED or conventional ion exchange, in that you can reverse the polarity across the system, and get desorption of ions from the exchange media. This permits one to clean in place without adding chemicals. The water removed from the glycerol- water fraction can be re-used in the process, as it tends to have a relatively low ion concentration.

The present invention will be better understood with reference to the following non- limiting example.

Example 1 : Glycerin Purification Process

As shown in Figure 1, during transfer to a storage tank, crude glycerin is coarse- filtered by pumping it (using Pump 10) through a strainer (20) to remove any larger objects (> V. cm) and then through a 10 micron filter (30) to remove suspended solids. The crude glycerin is then transferred to another tank (40) where it is diluted for further processing. In one embodiment, tank 40 includes a site glass (50). The glycerol is pumped (using Pump 60) to a second tank (70) where it is diluted with water to a concentration of between around 10 and 70% glycerol by volume, more preferably between around 40 and 60% glycerol by volume, most preferably around 50% glycerol by volume. The water going into tank 70 is ideally subjected to a purification step, by passing the water through a water softener (80) and a de-chlorination filter (90) The crude glycerin can then, if necessary, be pumped (via Pump 100) through an ultrafiltration unit (110) to remove fatty acids and other organics with a molecular weight higher than glycerin. The rejected material can be sold as a product or discharged as waste and sent to storage (Drum 120). Some potential uses of the rejected material are substrates for anaerobic digesters, among others After ultrafiltration, the dilute crude glycerin can optionally be filtered through activated carbon, for example, through carbon bag filter 130 to further remove organics which impart a yellow or brown color to crude glycerin. The dilute, filtered glycerin is recirculated through an electrodeiomzation process where anions and cations are removed, first passing through the deiomzation tank (140) before being pumped (via Pump 150) through an electro-deionization unit (160). The removed anion and cation salts and/or organic acids are concentrated into salt concentrate stream (stored in storage unit 170), which can optionally be recirculated through the electrodeionization unit (via Pump 180). The ions are forced from the dilute, ionized glycerin by the application of an electric field over the deiomzation cells. When the dilute glycerin reaches asymptotic limits on the withdrawal of glycerin is ready to be concentrated. Excess water can be removed by vacuum evaporation but other processes such as pervaporation or molecular sieves can be employed, or combinations thereof.

For example, the relatively cool glycerin/water mixture can pass through a heat exchanger (190) and travel (via Pump 200) to evaporator tanks (210 and 220) which are supplied heat via a boiler (230). The evaporated glycerol then can travel (via pump 240) to a storage tank (250). Hot air and water vapor can travel (via pump 260) to a heat exchanger (270), and then the cooled, deiomzed water can travel to a condensate tank (280) where it can then be fed (via pump 60) back into the process.

References 'Glueckauf, E., "Electro-deionisation Through a Packed Bed", British Chemical Engineering, December 1959, pp 646-651

¹ Sammon, D. C, and Watts, R.E., "An Experimental Study of Electrodeionisation and its Application to the Treatment of Radioactive Wastes", AERE-R3137, Chemistry Division, U.K A.E.A. Research Group, Atomic Energy Research Establishment, Harwell, June 1960. ¹ Verkeer, H.E. et ah, "Process and Apparatus for the Electrolytic Deionization of Salt- Containing Liquids", UK Patent No 776,469, June 5. 1957

¹ Kollsman, P., "Method and Apparatus for treating ionic fluids by dialysis", US Patent No. 2,815,320, December 3, 1957. ] Tye, F.L., "Improvements relating to Electrodialysis Process", UK Patent No. 815,154, June 17, 1959.

¹ Pearson, R.G., "Electrolytic Deionization", US Patent # 2,794,777, June 4, 1957.

¹ Kressman, T.R.E., "Process for the Removal of Dissolved Solids from Liquids", U.S. Patent

#2,923,674, Feb. 2, 1960. λ Parsi, E.J., "Removal of Dissolved Salts and Silica from Liquids", US Patent # 3,149,061, Sept. 15, 1964.

¹ Matejka, Z., "Continuous Production of High-Purity Water by Electro-deionisation", J. Appl. Chem. Biotechnol., 1971, Vol. 21, April, pp. 117-120.

¹ Shaposhnik, V.A., Reshetnikova. A.K., Zolotareva, R.I., I. V. Drobysheva, and N.I. Isaev, "Demineralization of Water by Electrodialysis with Ion-exchanger Packing Between the Membranes", Zhurnal Prikladnoi Khimii, Vol. 46, #12, pp. 2659-2663, December, 1973, translated by Consultants Bureau, Plenum Publishing Corp., 227 W. 17th St., New York.

¹ Selegny, E. and E. Korngold, "Method of Separation of Ions From a Solution", U.S. Patent # 3,686,089. ] Korngold, E., "Electrodialysis processes using ion exchange resins between membranes", Desalination, 1975, Vol. 16, No. 2, pp. 223-233.

¹ Kedem, O., "Reduction of polarization in electrodialysis by ion-conducting spacers", Desalination, 1975, Vol. 16, No. 1, pp. 105-118.

¹ Davis, T.A., "Electrically Regenerated Ion Exchange System" U.S. Patent # 4.032,452, June 28, 1977.

¹ Tejeda, R.A., "System For Demineralizing Water By Electrodialysis", US Patent # 3,869,376, Mar. 4, 1975.

¹ Kunz, G., "Process and Apparatus For Treatment of Fluids. Particularly Desalination of Aqueous Solutions", US Patent # 4,636,296, Jan 13, 1987. ] Giuffrida, A.J., Jha, A.D. and Ganzi, G.C., "Electro-deionization Apparatus", US Patent # 4,632,745, Dec. 30, 1986.

¹ Ganzi, G.C., Wood, J.H. and Griffin, CS. , "Water Purification and Recycling Using the CDI Process", Environmental Progress, Vol. 11, No. 1, February 1992.

¹ Henley, M., "Technology Overview: CEDI Gains Ground as Water Treatment Approach," Ultrapure Water, pg 15, October 1997.

¹ Ganzi, G.C., "The Ionpure(TM) Continuous Deionization Process: Effect of Electrical

Current Distribution on Performance", Presented at the 80th Annual AIChE meeting on Nov. 28, 1988, Washington, DC. Modifications and variations of the present invention relating to a fuel additive composition and an alternative fuel derived from the composition will be obvious to those skilled in the art from the foregoing detailed description of the invention.

Claims

Claims:

1. A method for producing glycerol with a relatively low salt concentration, from a starting material comprising glycerol with a relatively high salt concentration, comprising the steps of: a) diluting the glycerol with a relatively high salt concentration to a concentration between about 10% and 70% water by volume, and b) subjecting the glycerol/water solution to electrodeionization conditions

2. The method of Claim 1, further comprising the step subjecting the diluted glycerol/water solution to ultrafiltration conditions after dilution and before the electrodeionization step.

3. The method of Claim 1, further comprising the step of passing the de-salted glycerol/water solution through adsorbent charcoal to remove colored impurities.

4. The method of Claim 1, further comprising removing the water from the desalted glycerol/water solution.

5. The method of Claim 4, wherein the de-salted glycerol is greater than 90% pure.

6. The method of Claim 4, wherein the de-salted glycerol is greater than 95% pure.

7. The method of Claim 4, wherein the de-salted glycerol is greater than 99% pure.

8. The method of Claim 1, wherein the glycerol with a relatively high salt concentration is derived from the reaction of crude glycerol from a biodiesel process, wherein the crude glycerol comprises a fatty acid carboxylate impurity and an alkoxide impurity, and the crude glycerol is treated with sufficient acid to convert the fatty acid carboxylate impurity to a free fatty acid, and the alkoxide impurity to an alcohol, resulting in the formation of a salt impurity, and wherein the free fatty acid and alcohol are removed before the resulting glycerol with a relatively high salt concentration is subjected to the method of Claiml.

9. The method of Claim 1, wherein the glycerol/water solution comprises between about 40-60% glycerol and between about 60-40% water, by volume.

10. The method of Claim 1, wherein the glycerol/water solution is approximately 50% glycerol and approximately 50% water by volume.

11. A method for preparing a relatively salt-free glycerol product from the crude glycerol by-product from biodiesel production, comprising the steps of: a) obtaining a triglyceride starting material which contains from about 0.5 to about 10% free fatty acids, b) transesterifying the triglyceride with a Cw alcohol in the presence of a basic hydroxide or alkoxide catalyst to form a product mixture including a biodiesel fraction and a crude glycerol fraction, wherein the crude glycerol fraction includes carboxylate salts of any free fatty acids present in the triglyceride starting material, and hydroxide and/or alkoxide salts of the basic catalyst used in the transesterification reaction, c) acidifying the crude glycerol fraction to convert the fatty acid carboxylate salts to free fatty acids, and the hydroxide and/or alkoxide salts to water and/or alcohol, d) removing the fatty acids and alcohol from the crude glycerol fraction to leave a crude glycerol fraction comprising predominantly glycerol and a salt comprising the anion of acid used in the acifidication step and the cation(s) from the fatty acid carboxylate salt and the alkoxide salt (i.e., "glycerol with a relatively high salt concentration"). e) diluting the glycerol with a relatively high salt concentration to a concentration between about 30% and 70% water by volume. f) subjecting the glycerol/water solution to electrodeionization conditions.

12. The method of Claim 11, further comprising the step of subjecting the resulting glycerol/water solution to ultrafiltration conditions, after the dilution step and before the electrodeionization step.

13. The method of Claim 11, further comprising the step of passing the de-salted glycerol/water solution through adsorbent charcoal to remove colored impurities.

14. The method of Claim 11, further comprising removing the water from the desalted glycerol/water solution.

15. The method of Claim 14, wherein the de-salted glycerol is greater than 90% pure.

16. The method of Claim 14, wherein the de-salted glycerol is greater than 95% pure.

17. The method of Claim 14, wherein the de-salted glycerol is greater than 99% pure.