Biodiesel reaction and separation methods range from the time-honored large batch tanks with long residence time reactions and water washes for product separation to intensification, enzymatic, and supercritical reactions coupled with distillation and mechanical separation methods. Selecting the best reaction and separation method for the process depends upon the feedstock characterization, a process we covered in the August issue of Biodiesel Magazine. This installment highlights the specific drivers behind selecting the best reaction and separation techniques while considering process throughput, overall conversion efficiency and plant economics. 

The biodiesel reactions have three elements for driving the conversion of the feedstock to the finished product: mixing, molar ratio, and residence time. In simple terms, you have to present the reagent molecules with the opportunity, enough energy and enough time to react. It’s kind of like a seventh-grade school dance.

Mixing is the first operation to consider. Mixing drives the reagent interface surface area. The interface surface area is increased by decreasing through shear the dispersed phase liquid droplet size to the smallest size possible. In the usual batch reactor, the mixing is motivated by an agitator. Many high-shear flow strategies have been successfully employed to intensify phase interaction. The number of molecules motivated to react is driven by the surface area of the two immiscible phases.

The smaller liquid droplet sizes will create a greater surface area for the phase interface. The proper application goes back to the feedstock and what is being carried forward through the reaction. For example, clean, dry feedstock (the classic 1 percent MIU, or moisture, impurities and unsaponifiables, and 1 percent free fatty acid, or FFA, feedstock) reacts easily in a batch given the accepted residence time. Feedstocks containing phospholipids, high moisture, high FFA, and other contaminants may be more difficult to separate than clean and dry feedstock. It has been demonstrated, when employing high-shear and external energy methods, that over shearing can form too small a liquid droplet size and create an emulsion that is very difficult to separate. The real money is in the ability shear the phases to the optimum liquid droplet size that maximizes reaction conversion but minimizes the post reaction product separation effort. In-line high-shear mixers, high-shear orifices, high-pressure pumps, ultrasonic cavitation, and ultrasonic intensification methods have all been employed in this service to varying degrees of success.

Mixing energy is not enough to provide economic conversion efficiencies. Catalytic energy is required. The classic way to add this energy is through the addition of caustic. Recently, heat and pressure energy has been demonstrated in a supercritical catalytic reactor. The supercritical reactor has the added advantage of having the reagents in a single miscible phase as the polarity of the alcohol is minimized in the supercritical state allowing the oil and water to mix with infinite surface area, and separated in later distillation steps.

Molar ratio is the amount of one reagent compared to another, sort of like the boy-girl ratio at the dance. Increasing the molar ratio of the alcohol to the triglyceride/FFA in the reaction increases the probability of the alcohol’s availability for reaction. Only the stoichiometric limit of alcohol will be consumed in the reaction. The excess alcohol drives the reaction towards the product side of the reaction equation as per LeChatlier’s Principle. The trick is to minimize the excess alcohol needed to attain the desired conversion. Due to economic and environmental drivers, alcohols should be recovered and the post reaction alcohol separation required is an energy hog. Most alcohols have boiling points in the mid-100 to 200 degrees Fahrenheit range. The energy to boil them is high and the recovery of this heat in the cooling condensation is not favorable. Alcohol will distribute itself in both the water and the glycerol byproduct phases. The glycerol will also have a higher solubility in the oil phase in the presence of the alcohol and higher concentrations of alcohol at this step will leave a greater amount of glycerol to be removed from the fuel in the downstream refining operations. 

The residence time is dependent upon the reaction kinetics and is one of the key driving economic factors in any chemical process. The slower the reaction, the larger the vessels or the slower the flow rate of the system. In batch reactors, residence time is defined as the time it takes the reagents to complete conversion into products. In flow reactors, this is described as the space time or space velocity. In any case, the flow rate, reaction time and reactor volumes are the related physical principles. Backyard biodiesel systems can have residence times on the order of days. Common commercial system residence times are on the order of hours, whereas a supercritical plug flow reactor could have a space time on the order of seconds. Determining the reaction rate is critical to the economics, and reaction rate optimization using various catalysts will remain an active research topical area for years to come. 

Transesterification proceeds in three steps. Breaking up the triglyceride molecule into diglycerides is an easy first step thermodynamically and kinetically. The statistically deterministic reaction intermediate ratios then drive the reaction rate of the second step, which is the diglyceride degradation to monoglycerides. The monoglyceride to FAME is the slowest step in the reaction chain as the linear nature of the monoglyceride contributes to the carbon backbone stability. A free fatty acid converts more readily than a monoglyceride at higher temperature and pressures so the reaction strategy is a balance between FFA and available catalyst versus the use of just temperature and pressure without any homogeneous catalyst. This reaction progression difficulty is also the reason many low-tech reactor systems have such a hard time in the final refinement of the fuel to meet the ASTM standards. Many biodiesel plants have great conversion to 95 percent but getting the final conversion is difficult and the separation techniques needed to remove the unreacted monos have many economic penalties. Many times the monos are hidden in the oil/glycerin separation step and, in that case, it may appear to the producer that the reaction is fine, but, in truth, monoglycerides are emulsifying agents that inhibit the disengagement of the FAME from the unreacted reagents and undesirable byproducts. A strategy of using a two-step reaction to optimize the reaction conditions for the tough third mono-to-FAME step has been employed using many methods. Not to be mistaken with the acid base two-step that is commonly employed to convert high FFA material, this two-step reaction first separates the base catalyzed product from the water and glycerin and then the remaining fuel, glycerin, and triolein is finished in a second series reactor. 

Conversion reactions are affected by the feedstocks and the impurities that are present in the reactor. Water will reverse the transesterification reaction, but water formation is unavoidable in the high-FFA reaction as it is a byproduct of the reaction. It has been demonstrated that rapid removal of any formed moisture will increase yield from the transesterification reaction. In high-FFA reactions, one technique to sequester the water back from reaction is introducing alcohol in excess. The high molar ratio alcohol blocks the water from reacting with the oil due to statistical probabilities. Alcohol-to-oil molar ratios of 50:1 have produced high-yield, short residence time reaction performance, but distillation of the 50:1 alcohol to the purity required for its reuse is usually detrimental to the economics.

Many operations have learned how to sneak in 5 to 6 percent FFA feedstocks for reaction. One key factor in single digit FFA feedstock conversion is ensuring that the reagents are bone dry. A centrifuge separation or another dehydration method could help alleviate the problem of moisture-laden feedstock. In one application, fats, oil and grease (FOG) was separated using common pumper techniques where the feedstock oil was dewatered by conventional low-temperature methods; however, the coagulation polymers and the various associated chemicals in those processes need to be considered prior to introduction to the reaction system. 

Water is completely miscible in most alcohols used with biodiesel production. The water formed in the transesterification will fully associate with the alcohol, requiring distillation of the alcohol for its recovery and reuse.

Pervaporation separation techniques have been developed to reduce the energy requirements of the distillation chemical methods. Pervaporation will not replace the needed distillations, but implementation could reduce the energy required. Gross dehydration of the alcohol can be accomplished by pervaporation with final distillation returning the alcohol to the usable dehydrated state for reuse in the transesterification reaction. This strategy can fit well with smaller scale, community-based decentralized biodiesel efforts.

While water is an unavoidable byproduct of high-FFA feedstocks, sulfur is an unavoidable contaminant. The classic method to remove sulfur from petroleum feedstocks is through the use of hydrogen. In a large-scale refinery, this technique can be supported. On the local-scale, sulfur can be removed but it requires a series of steps and careful processing techniques while tracking the sulfur disposition. Sulfur can exist in multiple species. Usually most sulfur species are soluble in the biodiesel fraction of any separation. Water washes and cold surface absorbent techniques will remove a fraction of the sulfur, but not enough when employing sulfur-laden feedstocks. Sulfur removal is definitely an area for an integrated process system strategy as it may take multiple steps or multiple unit operations to get the sulfur to the needed finished product specification.  Sulfur species are split in a range of light, mid, or heavy compounds. The sulfur is usually most concentrated in light and heavy fractions, working to the advantage in considering distillation. One technique that has been successful in removing sulfur is to have a series of distillation steps for the finished fuel. Careful control of the column pressure and temperature will allow the removal of the lights in the first distillation step and will leave the heavies behind in the second distillation step. Fuel distillation increases energy consumption, but the refining also allows removal of various residues such as unreacted glycerides and minerals that often prevent the fuel from meeting the ASTM D6751 specification.

The sulfur specifications for both ASTM D6751 and ASTM D975 fuel are 15 ppm. A stretch goal to chase for overall favorable biofuels economics might be to produce a specialized diesel fuel that can meet a 1 ppm sulfur level.  As with most specialty chemicals, sulfur that meets a 1 ppm standard commands premium pricing. Oilseed feedstock has lower sulfur levels, so most operations have no problem hitting the 15 ppm specification. Feedstock that has a high-FFA percentage typically also has a high sulfur level.  This is not the case everywhere, but high-FFA material is degraded trigycerides and many of these feedstock sources come from materials that pick up a wide variety of sulfur-containing degradation products. As more high-FFA waste feedstocks have been introduced into systems through blending or complex reaction techniques, sulfur has become a rising concern.

Authors: Christina Borgese, Marc Privitera
Founding Engineers, PreProcess Inc.
(949) 201-6041
christina@preprocessinc.com