Introduction During a time of foreign fuel dependency and high green house gas emissions, it is

University of Tennessee at Chattanooga MOLAR RATIO STUDY FOR THE REACTION OF FREE FATTY ACIDS WITH METHANOL TO FORM FATTY ACID METHYL ESTERS OR BIODIESEL FUEL by Trip Dacus ENCH 435 Course: Ench435 Section: 001 Date: 11/17/2009 Instructors: Jim Henry, Tricia Thomas, Frank Jones

Introduction During a time of foreign fuel dependency and high green house gas emissions, it is imperative for the US to develop a biofuel that can alleviate economic and environmental pressures. Biodiesel, a fuel made from plant and animal oils, uses domestic feedstock for production and recycles CO 2 previously released during combustion. Currently biodiesel is produced from food oils using homogenous base catalysts in a batch reaction process. The creation of significant byproducts, the displacement of a food source, and the batch nature of this process make biodiesel an expensive fuel alternative. In other areas of the biodiesel project, pure metal and metal oxide catalysts were tested in order to develop a heterogeneous, continuous flow process. It was found that several vegetable oils can be successfully converted to biodiesel using these metal and metal oxides at low temperatures (see Appendix A). Use of the novel catalysts did not create unwanted byproducts that require purification. Successful catalysts also create biodiesel at low temperatures for a variety of vegetable oils containing traditional impurities such as free fatty acids. Free fatty acids traditionally cannot be used as a feedstock for commercial biodiesel production since it almost exclusively converts to soap in the presence of a base catalyst. Since our system does not use base catalysts, free fatty acids were able to be successfully converted to biodiesel without the formation of soap byproducts. In order for further experimentation on free fatty acids to continue, it was necessary to determine the proper molar ratio of methanol and free fatty acids to use in order to optimize conversion to biodiesel.

Objective The purpose of this experiment is to determine the proper molar ratio of methanol and free fatty acids to use in order to optimize conversion of these reactants into biodiesel. The methanol and free fatty acid are to be reacted in a 4:1, 12:1, and 24:1 ratio at 120 o C in the presence of nickel (II) oxide in 1.5mL stainless steel vials for a 1.5 hour residence time. A nuclear magnetic resonance machine will be used to analyze the product mixture and determine percent conversion to biodiesel. Background and Theory Reaction Chemistry Free fatty acids or FFA, such as oleic acid which can be seen below in Figure 1, are saturated or unsaturated hydrocarbon chains linked to a carboxylic acid. They are considered an impurity in various food and fuel industries since they form from the degradation of triglycerides or pure food oils. A typical triglyceride molecule can be seen below in Figure 2. R 1, R 2, and R 3 represent the three hydrocarbon chains similar to that of oleic acid.

Figure 1 Free Fatty Acid 1 Figure 2 Triglyceride 1 Both free fatty acids and triglycerides can be used as a feedstock for biodiesel production. However, in the presence of a base catalyst, free fatty acid converts to soap instead of the biodiesel molecule. As a result, feedstock oil streams must be purified of free fatty acids in order to be used as a feedstock for biodiesel in current industrial processes. A soap molecule can be seen in Figure 3. When the positive ion from the base catalyst, in this case Na, dissociates, it reacts with FFA to form soap. Since soap is soluble in both polar and non polar liquids, it is a difficult impurity to separate from product streams and increases the processing time and costs for biodiesel production.

Figure 3 - Carboxylate Salt or Soap 2 Even when triglycerides are used as a feedstock, there is a possibility of the formation of soap as the base catalyst removes the hydrocarbon chains from the glycerol backbone. The reaction to create soap competes with the reaction to create fatty acid methyl esters (FAME), or biodiesel. This is called a transesterification reaction. Figures 4 and 5 show the transesterification reaction using a tryclyceride and a free fatty as a reactant. R COO C H R 1 2 R 3 COO C H COOCH 2 2 2 + 3CH 3 HOC H K1 OH HOC H K 2 HOCH 2 2 + R R COOCH R 1 2 3 COOCH COOCH 3 3 3 Triglyceride Methanol Glycerol Methyl Esters (Oil/fat) (Biodiesel fuel) Figure 4 - Transesterification reaction with a triglyceride

+ CH K 3 3OH H 4 2O + RCOOCH K 3 Oleic Acid Methanol Water Methyl Ester (FFA) Figure 5 Transesterification reaction with a free fatty acid In previous studies using soybean oil as a triglyceride feedstock, it was found that metal and metal oxide catalysts were able to successfully create FAME without the formation of soap byproducts since there were no positive ions in the reactant feed. Further studies showed that free fatty acids could also be used as a reactant without the formation of soap. Therefore, it can be concluded that free fatty acids do not have to be purified out of a reactant feed for a process using solid, heterogeneous catalysts, nor do product streams have to be purified of soap byproducts. This is a great advantage for future biodiesel production since free fatty acids are a major component of many waste oils such as restaurant waste grease and trap grease from water processing plants that were previously unused in the industrial production of FAME. In the United States, 2.75 billion lbs of waste recyclable restaurant oil grease and 11.64 billion lbs of animal fat (such as poultry fat) are produced annually. Yellow grease or restaurant waste grease has a 15wt% FFA composition and costs between $0.01 to $0.07

per pound. Brown grease or trap grease from water processing plants has a 6 to 15wt% FFA composition and costs between $0.09 to $0.20 per pound. 3 Due to their high FFA content, these waste oils are either purified or not used in industrial biodiesel production. Use of heterogeneous catalysts allows for FFA to be used in feedstocks without purification making these inexpensive feedstocks a feasible alternative for biodiesel production. Tall oil was used as the primary source of FFA for the molar ratio study. It is comprised of mostly oleic acid which is the hydrocarbon shown in Figure 5. Nuclear Magnetic Resonance Spectroscopy NMR spectroscopy utilizes the magnetic properties of nuclei in order to determine the composition and properties of organic chemical samples. Proton or hydrogen nuclei exhibit either an α or a β spin state that will align with or against a strong, external magnetic field. When a radiation energy is introduced to the aligned nuclei that are equal to the energy difference between the two spin states, the spin of the nuclei will flip. This radiation energy is in the radio frequency of the electromagnetic spectrum, and nuclei that absorb this radiation are said to be in resonance with the applied energy. Detectors within the NMR instrument can sense when this electromagnetic energy is being absorbed, and the resulting plot of absorption for different electromagnetic frequencies is what is used to analyze the chemical sample.

The greater the magnetic field that the sample is exposed to is, the greater the resolution of the instrument. In UTC s chemistry department, they use a 10 tesla NMR to perform their resonance analysis. To put in perspective how strong of a field this is, a typical household bar magnet usually creates a 0.001 tesla magnetic field. UTC s NMR utilizes helium cooled (between 4 and 2 K) superconductors to generate its powerful magnetic field. Further use of shimming coils ensures the proper shape of the magnetic field that the chemical sample is exposed to. The electron cloud in surrounding nuclei acts as a shield from the applied magnetic field. This shielding effect causes different protons to absorb different levels of radiation depending on the electron density of the surrounding nuclei. As a result, the frequency of absorption can be linked with different atomic structures. Samples are solvated in a variety of liquids for analysis such as D 2 O (water with deuterium nuclei) and deuterochloroform (CDCl 3 ). The deuterium present in these solvents respond to a known radio frequency and can be used as a reference point for the rest of the resonance peaks in the chemical sample. The samples are then spun along their axis so that the position of molecules within the sample can be averaged which increases the overall resolution of the spectrum. The sample is then exposed to an electromagnetic pulse containing a wide range of frequencies, and the protons absorb a specific frequency according to their resonance. The intensity of the frequency absorbed by the protons then decay over time, and a computer converts this data into an intensity-verses-frequency plot called a Fourier Transform.

The NMR resonance signals are measured according to a chemical shift from a reference compound such as tetramethylsilane or TMS. This chemical shift is calculated by dividing the distance in hertz downfield from the reference compound by the operating frequency of the spectrometer and is measured in parts per million or ppm. Dueterochlorophorm was used as the analytical solvent for the FFA molar ratio study. The acyl group present in both FFA and FAME shields surrounding protons and causes the chemical shift used to determine percent conversion to products. The two protons that bond to the carbon that is at the beginning of the hydrocarbon chain in FFA exhibit a chemical shift of 2.2ppm in the presence of dueterochlorphorm. The three protons on the methyl group of FAME exhibit a chemical shift of 3.6ppm in the presence of dueterochlorophorm. As FFA converts to FAME, the 2.2ppm protons next to the acyl group do not react and are present in both reactant and product molecules. This peak in the NMR spectra will remain unchanged regardless of the percent conversion to FAME. The 3.6ppm peak corresponding to the methyl ester group only exists in the FAME molecule. The Jeol software that performs the Fourier transform on the NMR spectra is able to integrate the peaks that correspond to the various chemical shifts. When the area under the 3.6ppm peak is compared to that of the 2.2ppm peak, percent conversion to FAME can be calculated. For 100% conversion, the area under the 3.6ppm peak will be in a 3:2 ratio to that of the 2.2ppm peak since the protons are at a 3:2 ratio in the FAME molecule.

Procedure 1. Place a 15x44mm plastic vial on scale and tare the balance. 2. Measure tall oil L1 in a plastic vial using a 5 ¾ in disposable glass pipette and rubber squeeze bulb. Ensure that no tall oil contacts rubber bulb. 3. Measure out methanol (CH 3 OH) using a graduated cylinder and add to the plastic vial containing tall oil. 4. Using a clean disposable glass pipette, mix reactants thoroughly. Transfer the reactants from the plastic vial into a 5mL stainless steel Gilmont reactor. 5. Place weighing paper onto the scale and tare the balance. Measure nickel II oxide using a spatula. Add the nickel II oxide powder to the Gilmont reactor containing tall oil and methanol. Appropriate amounts of tall oil, methanol, and nickel II oxide can be seen below in Table I. Table I Amounts of tall oil, methanol, and nickel II oxide for corresponding molar ratios Molar Ratio Tall Oil (g) Methanol (ml) Nickel II Oxide (mg) 4:1 1.795 1.0 103 12:1 0.898 1.5 83 24:1 0.449 1.5 65 6. Insert a silicone gasket into the screw cap of the Gilmont reactor. Ensure the gasket is flush against the surface of the screw cap. Rotate the cap into place and

ensure that the gasket has made a seal by tightening the cap firmly onto the reactor. NOTE: initially silicone was used as a gasket material due to its behavior at high temperature. However, silicone was found to be susceptible to chemical attack, and after several reactions, the gaskets began to degrade. It was found that the degraded silicone was inhibiting the NMR spectroscopy (please refer to Results NMR Sample Contamination for further analysis). Teflon gaskets were then used in place of silicone so that they would not degrade during experimentation. 7. Place the reactor into an oven that has been preheated to 120 o C. After 1.5 hours, remove reactor from oven using appropriate gloves. 8. Repeat steps 1-8 so that three reactions have been performed at all three molar ratios, a total of nine reactions. 9. Upon completion of the reaction, allow the reactors to cool to room temperature. It is recommended that natural convection only is used for cooling since other transesterfication reactions were cooled by natural convection. 10. Place a small wad of paper or chem wipe into a 5 ¾ in glass pipette. Use a 9 in glass pipette to push the wad into the nozzle of the 5 ¾ in pipette. 11. Push a 5 ¾ in pipette about half an inch into Celite filtering agent. Flick the side of the 5 ¾ in pipette until the filtering agent has fallen against the wad of paper. 12. Place the pipette into a standard NMR tube.

13. Open the reactors and then use a 9 in glass pipette to thoroughly mix the products. Add enough product mixture to the 5 ¾ in pipette containing the filtering agent to moisten the top most layer. 14. Using a clean 5 ¾ in pipette, add 4 to 5 ml of chloroform-d to the 5 ¾ in pipette containing the product mixture. 15. Allow the chloroform product mixture to filter through Celite into the NMR tube. 16. Submit the sample for NMR analysis. To clean reactors 1. Rinse thoroughly with tap water 2. Rinse thoroughly with acetone 3. Use a Chem wipe to remove excess catalyst and to dry reactor, reactor caps, and gaskets. Equipment The following is a list of equipment that used to perform the experiment Glassware and accessories - 5 ¾ in disposable pipettes - 9 in disposable pipettes - Rubber squeeze bulbs - 15x44mm plastic vials with screw threads

- NMR sample tubes - NMR sample tube caps - NMR tube stand - 10mL graduated cylinder Reactants and chemicals - Celite filtering agent - Nickel II oxide catalyst powder - Tall Oil 1% rosen composition - Laboratory grace methanol reagent - Acetone and acetone bottle Miscellaneous equipment - 15mL Gilmont stainless steel reactors - Silicone gaskets - Teflon gaskets - Chem wipes - Balance with 0.1 mg accuracy - Oven with 200 o C heating capability - Insulated gloves for handling hot reactors - Chemical sample labels Results

NMR Sample Contamination In order to better seal the stainless steel reactors used for chemical analysis, silicone gaskets were used in place of the original plastic gaskets supplied by Gilmont. Silicone was used since it behaves well at high temperatures and its behavior as an elastomer creates a seal for the reactors. After several reactions, it was found that the silicone gaskets began to degrade due to chemical attack from the reactants. Eventually they were replaced with Teflon gaskets that were able to resist chemical attack. Figure 6 below shows a picture of a degraded silicone gasket and an intact Teflon gasket. Initially this did not pose a problem, since the reactions with triglycerides still had a proper seal and the NMR spectroscopy did not show any contamination in the product mixture. When FFA was used as a reactant, the results from the NMR spectroscopy showed some unusual peaks that could not be explained by known chemicals in the sample. Unknown peaks are usually not a problem for spectroscopy analysis, but these peaks covered the 3.6ppm peak that was used to analyze the percent conversion to FAME. A picture of the NMR spectroscopy analysis can be seen below in Figure 6. A study was then performed to determine the source of contamination. Although silicone was the hypothesized source, contaminants in the solvent, methanol, or tall oil could also be why an unknown peak was covering the 3.6ppm range.

Figure 6 Degraded silicone gasket (left) and intact Teflon gasket (right) Unknown Peaks Covered 3.6ppm Peak Figure 7 Contaminated NMR spectroscopy Methanol, tall oil, and pure deuterochlorophorm were analyzed to see if these chemicals had been accidently contaminated during experimentation. The NMR resulting

spectroscopy of these materials showed that there were no unknown materials present. NMR spectra for pure methanol and tall oil can be found in Appendix B. An NMR was also run on an un-reacted methanol, tall oil mixture that to see if mixing these products resulted in any spectroscopy abnormalities. Figure 7 below shows the NMR spectroscopy for the un-reacted methanol, tall oil mixture. Figure 8 NMR spectroscopy of un-reacted methanol, tall oil mixture Figure 7 shows no unknown peaks around the 3.6ppm shift. The small peak that does exist at the 3.6ppm shift is a side band from the 3.4ppm methanol peak, and is the result of improper magnetic coil shimming in the NMR. These side bands were unable to be removed throughout the experimentation, but they did not inhibit the ability to determine percent conversion to FAME. The same un-reacted mixture was then washed over one of the degraded silicone gaskets to see if exposure to the gasket resulted in an unknown peak. Figure 8 below shows the NMR spectroscopy after the mixture was washed over the degraded silicone gasket.

Figure 9 NMR spectroscopy of un-reacted methanol and tall oil washed over degraded silicone gasket After washing the mixture over the degraded silicone gasket, the unknown peak was discovered in the NMR spectroscopy in Figure 8 at the 3.66ppm shift which matches the peak in Figure 6. Also, the 3.4ppm peak corresponding to methanol has been broadened similar to Figure 6 as well. It was then decided that silicone gaskets were no longer suitable for the reaction studies, and Teflon gaskets which resist chemical attack were acquired and used for further experimentation. Molar Ratio Study Table II contains the percent conversion to FAME data acquired from the NMR spectroscopy analysis for the three different molar ratios. The uncertainty of percent conversion were calculated using Student T test for 95% confidence (n = 3, t = 2.9).

Table II Percent conversion to FAME for Methanol:FFA molar ratios Percent Conversion to FAME 4:1 12:1 24:1 Trial 1 55.4 37.8 48.9 Trial 2 49.8 57.55 37.4 Trial 3 73 42 30.4 Average 59.4 45.8 38.9 Uncertainty 23.2 19.75 18.5 Figure 9 below contains a bar graph with the average percent conversion to biodiesel and uncertainties for each of the tested molar ratios.

Percent Conversion (%) 90 80 70 60 50 40 30 4to1 12to1 24to1 20 10 0 Figure 10 Average Percent Conversion to FAME for 4:1, 12:1, and 24:1 methanol to FFA molar ratios with uncertainties. From Figure 10, it can be seen that the 4 to 1 molar ratio results in the highest average conversion to biodiesel and the 24 to 1 results in the lowest. According to previous research, the 12 to 1 ratio was favored when triglycerides were used as a reactant. It would then be logical that using free fatty acids as a reactant would favor a lower ratio. From Figure 4, a single mole of a triglyceride requires three moles of methanol in order to convert to FAME. However, free fatty acids only require one mole of methanol for their reaction which can be seen in Figure 5.

Although Figure 10 shows that the 4 to 1 average percent conversion is higher, it is important to note that both the 12 to 1 and 24 to 1 average conversions are contained within the uncertainty of the 4 to 1 ratio. From a statistical standpoint, it cannot be concretely confirmed that the 4 to 1 ratio is the most favorable due to the large degree of uncertainty in this ratio s percent conversion to FAME. However, the nature of this study is to determine a favorable ratio with the understanding that other ratios may behave similarly. If future experiments are to optimize conversion to FAME using FFA as a reactant, this study would then recommend using a 4 to 1 ratio with respect to FFA as a reactant. In addition to this, the 4 to 1 ratio has a much higher viscosity than the 12 to 1 and the 24 to 1 mixtures. This intuitively is due to the low viscosity of methanol, which makes mixtures with higher ratios of methanol less viscous. Viscosity can play a significant role in industrial processing since many operations that require pumps to supply a flow rate of fluids have a limited range of viscosity in which they are allowed to operate. In these applications, it can be said that increasing the molar ratio of methanol to FFA can help to reduce the viscosity of the reactant feed without significantly affecting the percent conversion to FAME. Conclusions The 4 to 1 methanol to free fatty acid ratio exhibits the highest percent conversion to fatty acid methyl esters when compared to 12 to 1 and a 24 to 1 ratios. The average percent conversion of both the 12 to 1 and the 24 to 1 ratios fell within the uncertainty of the 4 to

1 ratio. The 4 to 1 mixture was also significantly more viscous than both the 12 to 1 and 24 to 1 ratios. Therefore the 4 to 1 ratio can be diluted if its viscosity does not fall within the limits of the design process without significantly changing the percent conversion to FAME. Recommendations It is recommended that additional reactions be run for the three ratios in order to decrease the level of uncertainty for the average conversion to biodiesel. If the 12 to 1 and 24 to 1 average percent conversions no longer fall within the uncertainty of the 4 to 1 average, then the 4 to 1 ratio can be said to have definite advantages over the other two ratios. However, at this point there exists too much uncertainty to be able to declare this. Sources 1 Chemical and Nutritional Properties of Olive Oil. <www.oliveoilsource.com/olivechemistry.htm> 2 Phase Interactions; Nature of Colloids. <jan.ucc.nau.edu/~doetqpp/courses/env440/env440_2/lectures/lec19/lec19.html> 3 D' Cruz, A., M. Kulkarni, L. Meher, and A. Dalai. "Synthesis of Biodiesel from Canola Oil Using Heterogeneous Base Catalyst." Journal of American Oil Chemists Soc. 84 (2007): 937-943. Appendix A % Conversion to FAME for various feedstocks in stainless steel vials in the presence of nickel II oxide with a residence time of 2 hours.

100 SBO FFA SBO + FFA 80 Olive Oil % Conversion 60 40 Corn Oil Algae Oil 20 0 Appendix B NMR spectroscopy for pure methanol NMR spectroscopy for pure tall oil