Accepted Manuscript Process Development for Scum to Biodiesel Conversion Chonghao Bi, Min Min, Yong Nie, Qinglong Xie, Qian Lu, Xiangyuan Deng, Erik Anderson, Dong Li, Roger Ruan PII: DOI: Reference:
S0960-8524(15)00101-7 http://dx.doi.org/10.1016/j.biortech.2015.01.081 BITE 14513
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
29 November 2014 17 January 2015 19 January 2015
Please cite this article as: Bi, C., Min, M., Nie, Y., Xie, Q., Lu, Q., Deng, X., Anderson, E., Li, D., Ruan, R., Process Development for Scum to Biodiesel Conversion, Bioresource Technology (2015), doi: http://dx.doi.org/10.1016/ j.biortech.2015.01.081
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Process Development for Scum to Biodiesel Conversion Chonghao Bia,b, Min Minb*, Yong Niec, Qinglong Xieb, Qian Lub, Xiangyuan Dengb, Erik Andersonb, Dong Lia, Roger Ruanb* a
College of Engineering, China Agricultural University, Beijing 100083, China
b
Center for Biorefining, and Department of Bioproducts and Biosystems Engineering,
University of Minnesota, 1390 Eckles Avenue, St. Paul, MN, 55108, USA c
College of Chemical Engineering, Zhejiang University of Technology, Hangzhou
310014, China * Corresponding author. E-mail address: [email protected] (M. Min). [email protected] (R. Ruan), Tel.: +1 612 625 1710; fax: +1 612 624 3005 Abstract A novel process was developed for converting scum, a waste material from wastewater treatment facilities, to biodiesel. Scum is an oily waste that was skimmed from the surface of primary and secondary settling tanks in the wastewater treatment plants. Currently scum is treated either by anaerobic digestion or landfilling which raised several environmental issues. The newly developed process used a six-step method to convert scum to biodiesel, which would produce a higher value product. A combination of acid washing and acid catalyzed esterification was developed to remove soap and impurities while converting free fatty acids to methyl esters. A glycerol washing was used to help the separation of biodiesel and glycerin after base catalyzed transesterification. As a result, 70% of dried and filtered scum could be converted to biodiesel which is equivalent to about 134,000 gallon biodiesel per year for the Saint Paul Waste Water Treatment Plant at Minnesota.
Keywords: Biodiesel; Acid washing; Glycerol washing; Scum; Wastewater treatment plant 1. Introduction As a renewable fuel, biodiesel has become an attractive alternative for diesel fuel substitute. Vegetable oil has been the major source for global biodiesel production since 1990s (Krawczyk, 1996; EIA, 2014). It is still the largest proportion of feedstock of biodiesel in U.S. now. For instance, soybean oil is the largest biodiesel feedstock in U.S. during 2013 in which 5,507 million pounds was processed into biodiesel (48.8% of all kinds of feedstock); meanwhile, a total of 10,302 million pounds of soybean oil worldwide were used to produce biodiesel (U.S. Department of Energy, 2014). However, the main problem with making biodiesel a competitive fuel on the world market is its high cost. Currently, for vegetable based biodiesel, the feedstock alone takes about 70-80% of the total production cost (Cooperation and A. P. E., 2010). Many efforts have been focused using spent oil or waste oil, such as waste cooking oils, gutter oil, and wastes from animal or vegetable oil processing operations, to reduce the cost (Haas et al, 2006; Canakci & Sanli, 2008). Scum is a floatable material skimmed from the surface of primary and secondary settling tanks in wastewater treatment plants. It contains animal fat, vegetable oil, food wastes, plastic material, soaps, waxes and many other impurities discharged from restaurants, households and other facilities. Based on the report of the U.S. EPA’s Office of Solid Waste, approximate 1 to 3 billion gallons of waste grease, oil and fats are produced every year in the 30 metropolitan areas in the United States (Wiltsee, 1998). Some of those wastes (yellow grease) are collected in restaurants and
converted to biodiesel. However, approximately 60% of waste oils enter the sewer systems and end up in wastewater treatment plants. Due to a lower density than water, most of the oil floats on the surface of treatment facilities and conglomerates with other wastes to form scum, which makes scum a rich source of energy for recovery. The oil content of scum could be as high as 60%. The conventional technology for energy recovery is to co-process scum with sludge in anaerobic digestion where the produced biogas can generate electricity for plant use (Outwater & Tansel, 1994). However, this technology raises many problems in operation. For example, the scum floats on the top of the digester and forms a thick layer that impedes digester performance. As a result, many wastewater treatment plants choose to directly dispose scum in landfills. The scum disposal not only increases the cost of treatment facilities, but also causes many environmental problems. For instance, the Metropolitan Wastewater Treatment Plant at St. Paul. MN (Metro Plant) spends $100,000 a year just for land filling the scum. The landfill leachate could be a potential source for underground water pollution. In order to solve these problems, an alternative process/technology needs to be developed to recover energy from the scum and at the same time reduce the environmental impact due to landfilling scum. In this study, a new scum-to-biodiesel conversion process was developed so that the fat, free fatty acids and soap content in the scum can be successfully converted to ASTM-grade biodiesel. Since the final product is a liquid transportation fuel-biodiesel, it could gain a much higher economic value than biogas if generated through anaerobic digestion. Due to the abundance of nature gas, biogas has the selling price
close to $0.80/therm; while the biodiesel has the selling price about $3.2/therm which is three times higher than biogas (CenterPoint Energy, 2014). In addition, the byproducts such as glycerol and separated solids can be used as a heating source that not only decreases the energy input but also reduces the cost of the process. In a word, the current developed process has obvious economic and environmental advantages over other treatments and disposal processes. Currently, the most common process for biodiesel conversion is through base-catalyzed transesterification where triglyceride reacts with methanol to form fatty acid methyl ester (FAME). Potassium hydroxide (KOH) or sodium methoxide (CH3ONa) is the most common catalyst. However in the base-catalyzed transesterification process, feedstock must be free of water and free fatty acid (FFA), otherwise soap formation would reduce the biodiesel conversion yield. When using low grade feedstock such as yellow grease or brown grease, the FFA content can be as high as 30%. A pretreatment method which reduces acid value below 2 mg KOH per g of oil, must be applied before proceeding with any base-catalyzed transesterification process. Scum as a complex material contains solids, water, soap, fatty acids, oil and many other impurities. For maximizing the conversion yield, we explored several pretreatment methods in this project to clean up the scum oil. The objective of this project is to develop an effective pretreatment and conversion process to convert the scum to biodiesel.
2. Material and method 2.1 Material and reagent The scum samples were collected from the Metro Plant at St. Paul, MN. Sulfuric
acid (96.4%, AR) and Hydrochloric acid (36.5-38.0%, AR) were obtained from Mallinckrodt Baker, Inc. Paris, Kentucky. Phosphoric acid (85.0%, GR) was obtained from EM Industries, Inc. Potassium methoxide (>90.0%) was purchased from Alfa Aesar. Butylated hydroxytoluene (BHT), methanol (anhydrous, 99.8%), chloroform (99.8%) and diethyl ether (99.7%) were obtained from Sigma-Aldrich, Inc. BF3-methanol reagent (14% borontrifluoride, 86% methanol), sodium hydroxide 0.5 normal in methanol, heptane (HPLC grade) and glycerol (99.9%) were obtained from Thermo Fisher Scientific, Inc. potassium hydroxide concentrate (1.0 mol/L) was obtained from Fluka Analytical Sigma Aldrich Co. Ethanol (200 proof) was from Decon Laboratories, Inc. Distilled water was obtained from Premium Waters, Inc. MN, USA. 2.2 Process design The general process for the scum to biodiesel conversion was illustrated in Figure 2 and the main processes can be summarized according to the following six steps: FilteringAcid washingAcid catalyzed esterificationBase catalyzed transesterification Glycerol washing Oil refining
Firstly large particles in scum were separated by filtration. Then acid washing followed by gravitational settling was applied to convert soap (mainly (RCOO)2Ca and (RCOO)3Al) to free fatty acid (FFA), separate oil from water and water soluble/insoluble electrolytes, and further remove fine particles. The upper layer oil was collected and subject to acid catalyzed esterification and then base catalyzed transesterification processes to form FAME. The basic reactions were described in Equations (1)-(3). A glycerol washing was then performed to aid in separating the
glycerol and other impurities from the FAME. After separating the glycerol and methanol from the FAME, the crude FAME was then refined by fractional distillation to produce high grade biodiesel that can be directly used in diesel engines. Soap + Acid → Salt + FFA Catalyst FFA + MethanolAcid →Water + FAME Catalyst Triglyceri de + Methanol Base → Glycerol + FAME
(1) (2) (3)
2.2.1 Filtering process Scum oil was obtained in solid form at room temperature. It melted at about 40oC (depending on its components) and the viscosity of scum was negatively related with the temperature. For easing the separation of the scum oil from solid particles and at the same time drying the oil, a filtration process was conducted under a relative high temperature to reduce the viscosity of the scum oil. In a drying oven of 105 °C, 157 g of scum was loaded into a polyester mesh filter bag with pore size of 100 micron and a beaker was placed beneath the filter bag to receive the oil melted from the scum sample. The filtering process lasted for 24 hours. Weights of filtered oil and remaining solids were collected every 15 min during the first 8 hrs and at the end point of 24 hrs. The moisture content was calculated by subtracting the oil and solid weights from the total scum weight. 2.2.2 Acid washing Purpose of the acid washing was to convert any soap in scum to free fatty acids, maximize the biodiesel production yield, break emulsion for better water/oil separation, and further remove impurity from the scum oil. It also made the acid catalyzed esterification easier due to less impurity interference with the reactants.
H2SO4, HCl, and H3PO4, with two H+ strength 0.2 N and 1.2 N in water were compared for the scum oil pretreatment. The ratio of the scum oil-to-acid solution was 1:1 by weight. Raw scum oil was mixed with acid solution in a set of flask with condenser on top; a magnetic stirring water bath was applied to offer stirring power and keep the temperature of the reaction system at 60°C (Leung et al., 2010). After acid washing for 1 hour, the mixture was allowed to settle for about half an hour to collect the oil in the upper phase and sediment/water in the lower phase. The upper oil was then separated for the next process. 2.2.3 Acid catalyzed esterification Acid catalyzed esterification/transesterification is a general approach for feedstock with high free fatty acid and this process is not sensitive to moisture. However, when compared to base catalyzed transesterification, acid catalyzed esterification is much slower. Therefore for feedstock high in free fatty acid, acid catalyzed esterification followed by base catalyzed transesterification was generally used and the acid catalyzed esterification became one of the pretreatment methods for reducing the FFA in feedstock. After acid washing, scum oil that contains residual water and high free fatty acid converted from soap had an average acid value (AV) of 21 mg KOH·g-1 oil. The main purpose of this step is to reduce the AV to below 2 mg KOH·g-1 oil through the acid esterification reaction described in Equation (2) and to prepare the oil for base catalyzed transesterification. In the biodiesel industry, the AV of feedstock with high fatty acid should be reduced below 2 mg KOH·g-1oil before base catalyzed transesterification (Zhang & Jiang, 2008; Cao et al., 2008). Regarding the acid catalyzed reaction, H2SO4 has been the most commonly used and widely
investigated catalyst (Lotero et al., 2005). During this process, methanol (30% of oil weight) and sulfuric acid (3% and 5% of oil weight) were added to the oil produced from last step. In a flask with a condenser at the top, a magnetic stirring water bath was applied to offer stirring power and keep the temperature of the reaction system at 60 °C (Canakci & Van Gerpen, 1999). The acid value of the oil phase was expected to drop with ongoing esterification. The acid value of oil phase was monitored every 15 minutes in the first hour, and after 150 minutes. After the reaction was completed, the mixture was settled for one hour to separate into two layers. The upper phase was the oil layer that would be collected and dried in 105°C oven over night and used for next base catalyzed reaction. The majority of methanol and all the acid were in the lower phase. 2.2.4 Base catalyzed process Due to the fast reaction rate, base catalyzed transesterification is the general process for low water feedstock (water content less than 0.05% w.t.) with low free fatty acid (AV less than 2). After acid catalyzed esterification and separation, the scum oil would meet the criteria as a feedstock for base catalyzed esterification. The reactor was as same as described in section 2.2.2. Methanol (30% of oil weight) was added into the corresponding amount of oil (obtained from acid catalyzed step) and potassium methoxide (CH3OK, 0.5% and 1% of oil weight) acted as catalyst to determine the conversion efficiency (Meher et al, 2006). Samples were collected every 10 minutes for an hour. For analyzing the FAME content, the methanol left in samples was flash evaporated first, and then glycerol washing procedure described in
section 2.2.5 was applied to induce separation of glycerol and FAME; then the supernatant containing FAME was specified by GC-MS analysis described in section 2.3.3. 2.2.5 Glycerol washing and layering In principle, methyl ester and glycerol is immiscible. When using a better quality feedstock, such as soybean oil, the FAME and glycerol would separate into two layers by gravity settling after the transesterification reaction. However, scum oil will form a stable single-phase system even after 24 hours settling at 60 oC. It was suspected that some organic compounds (for instance, surfactants or other impurities) and the soap could increase the inter-solubility of glycerol and methyl ester; therefore glycerol could not be separated. The product mixture has to be separated and the base catalyst must be removed before distillation process, otherwise the product mixture will subject to reverse reaction and re-saponification. It was found that by adding extra glycerol to the mixture could actually improve FAME/glycerol separation. Therefore, a process step that was called “glycerol washing” was applied. As a result, glycerol washing could not only bring the oil mixture to layers but also wash out most of the base catalyst and many impurities in the scum. After the base catalyzed reaction, the methanol in the mixture oil was removed by rotary evaporator. Glycerol was then added into the mixture oil with different ratio (glycerol:oil = 1:20; 1:10; 1:5; 1:1.33; 1:2; 1:1 by weight) and mixed thoroughly by magnetic stirring at 60°C for 20 minutes, and then kept settling for 12 hours at the same temperature. The upper layer contained mainly FAME and the lower layer was
glycerol and base catalyst. 2.2.6 Distillation Upper layer outcome from last step contained about 90% FAME (according to the base catalyzed condition). This FAME mixture was introduced to a customized vacuum rectification column system and refined FAME can be obtained. 2.3 Analytical methods 2.3.1 Elementals analysis Samples from acid washing process were submitted to Soil Testing & Research Analytical Laboratory at the University of Minnesota (St. Paul, MN) to perform elemental analysis. An inductively coupled plasma atomic emission spectrometer (Perkin Elmer Optima 3000, USA) was applied on all the samples to provide a total elemental analysis of acid washed scum oil samples (Yeomans & Bremne, 1991; Lee et al, 1996; Fassel & Kniseley, 1974). 2.3.2 Acid value (AV) determination Acid value (mg KOH/g oil) is an indicator of the free fatty acid content in the scum oil. Standard Test Method for Acid Value of Fatty Acids and Polymerized Fatty Acids ASTM D1980-87 (1998) were used to determine the acid value. By monitoring changes of acid value during acid catalyzed esterification process, progress of reaction was determined. 2.3.3 Fatty acid methyl ester (FAME) analysis The BF3 catalysis method was applied from AOCS Official Method Ce 2-66 (1997). This method proposes excess reactant and excess catalyst to make sure that all the fatty acid in sample was transformed into FAME.
FAME in all samples was analyzed by GC-MS using the method from Li et al., (2011). Oil samples (about 100 mg) were weighed into 10 ml volumetric flasks and diluted with chloroform to 10 ml. The GC-MS (Agilent 7890-5975C, USA) equipped with a HP-5 column and a mass detector was used. Chromatographic data were recorded and addressed by a built-in Agilent data analysis software. Components were identified in NIST Mass Spectral Database. Quantification was carried out by comparing the peak area with that of the GLC-10 and GLC-30 standard mixtures (Sigma-Aldrich, Co.). 2.4 Biodiesel fuel testing The distilled biodiesel fuel was sent out and tested by Iowa Central Fuel Testing Laboratory (Fort Dodge, IA) using ASTM D 6751/BQ-9000 Full Spec Test Package for B100. The results were listed in Table 2. 3 Result and discussion 3.1 Filtering process At the end of the process, water, oil, and solids amounted 28.2%, 56.0%, and 15.8% of the total scum weight respectively. The specific filtering curves were illustrated in Figure 2. The weight changes (% of total scum weight) of the filtered oil, the water evaporated and the solid left in the filter bag were recorded over time. Water content started to decline at the first 15 minutes and the total water evaporated at 24 hours was 28.2% of the scum weight. The dash line in the Figure 2 indicated the water evaporation rate. It showed that the highest rate of water evaporation was 0.041 %/min observed at 2.5 hours. The rate declined to below 0.023 %/min after 5.75 hours. Filtered oil was received in the beaker, continuously increasing to reach 56.0%
after 24 hours. Rate of oil filtration curve was a derivative of oil weight percentage over time. It was observed from this curve that the rate of oil collection was very low during the first hour, then climb up to the maximum of 0.702 %/min at 2 hours, after that, the rate decreased to below 0.023 %/min after 5.75 hours. Weight of solids in the filter bag was also shown in Figure 2. This curve dropped from 100% to 15.8% indicating 15.8% solids in scum. The curvature of the solids curve was negatively correlated to the oil curve during the first 6 hours which indicated that the weight loss of scum solids was mainly due to the separated oil. During 6-24 hours the solids curve was positively related to the water curve which indicated that the solid weight loss during 6-24 hours was mainly due to the water evaporation. At the time of 6 hours, oil filtered was 52.2% of the total scum weight, which counted 93.2% of total oil. However, at time of 8 hours, 98.0% of the total oil (counted 54.9% of total scum weight) could be filtered out. When considering maximizing oil yield versus maximizing energy efficient, process time and capital cost, there was a tradeoff point in this filtering process. For maximizing oil yield, longer heating and filtering time would be better, however it also increased processing time, energy and capital cost. When using shorter time, oil yield would be slightly decreased, but the energy for evaporating water would be reduced by half and the size of the filtering equipment would be smaller due to faster processing time. The water content left in oil would be further reduced in the acid washing step. 3.2 Acid washing It was found that acid solution with 0.2 N couldn’t acidize scum oil completely; while the acid solution with 1.2 N could acidize scum oil completely, leaving 0.3-0.4
mol/L H+ surplus after this treatment. It was observed that most sediment was formed when the H2SO4 was added, some sediment with H3PO4 addition, however, no sediment appeared with HCl added system. This was because SO42- and PO43- were more likely to form insoluble salts with the metallic elements, such as CaSO4, Fe2(SO4)3, etc. When using dried and filtered scum oil as the starting material, the acid washing would account for 10% of mass loss due to remove all these metallic impurities. Table 1 showed the major impurity elements in the scum oil before and after acid washing. Many soap component elements such as Calcium, Iron, Aluminum, Magnesium, Zinc etc. were decreased or eliminated after acid washing. For example, the Ca2+ was dropped from 1483 mg/kg to less than 0.43 mg/kg after acid washing. Nitrogen content is an indicator of surfactant content as it may forms lots of active hydrophilic groups and surfactants, such as -NH3, -N2+, etc (Ananthapadmanabhan, 1993; Hayashita et al, 1994). After H2SO4 treatment, the content of the nitrogen content dropped from 1990 mg/kg to 1200 mg/kg, which was significantly superior to HCl and H3PO4. In conclusion, H2SO4 was considered to be the most suitable acid washing reagent for scum oil pretreatment. The acid washing was more effective for removing inorganic impurities than for those organic compounds containing N, P, and S. After acid washing, N content still remained 1200 mg/Kg in scum, 80% of S wasn’t affected, only P was mostly removed. These organic impurities might interfere with subsequent processes. Currently 6% sulfuric acid solution mixed with scum oil with 1:1 ratio by weight was used for the acid washing step. 3.3 Acid catalyzed esterification
Figure 3a showed the decrease of acid value of the different oil samples with various concentrations of H2SO4. The dotted curve was made using unwashed scum oil sample (just after scum filtration) with 3% of H2SO4 addition, and the circle and triangle curves were made using acid washed oil samples with 3% and 5% of H2SO4 addition respectively. It showed that the AVs of all three experiments decreased during acid catalyzed esterification reaction. AVs of unwashed scum oil with 3% catalyst had slowest decrease and ended at 8.07 mg KOH·g-1oil, meanwhile, AVs of acid washed oil with 5% catalyst had fastest decrease and ended at 2.10 mg KOH·g-1 oil. Sample of acid washed oil with 3% catalyst gained a final AV of 4.27 mg KOH·g-1 oil. The esterification reaction of the unwashed scum oil wasn’t as complete as the acid washed oil for same catalyst admixture. This was mainly because part of the sulfuric acid reacted with the soap in the scum and produced water. The consumed sulfuric acid reduced the catalyst strength and water also hindered the esterification reaction since the esterification reaction was a reversible reaction and water was one of the reaction products. That was why acid washed scum oil had better performance during the esterification reaction. It was found that acid washing and 5% catalyst was necessary for acid catalyzed esterification reaction. Figure 3b indicated the rate of AV decreasing by using derivative of AV to time. It was observed that reaction rates of all three experiments decreased as time elapsed which indicated that reactions achieved equilibrium gradually. During the first 15 min, reaction rate of acid washed oil (0.94 mg KOH·g-1 oil·min-1, 0.81 mg KOH·g-1 oil·min-1) was much higher than unacidized scum oil (0.29 mg KOH·g-1oil·min-1). However, they dropped faster in the next quarter of hour, and after 1 h, the reaction
rate of these three samples all dropped below 0.05 mg KOH·g-1oil·min-1, which was too slow for industrial application. Therefore it was reasonable to set the duration of acid catalyze esterification reaction at 1 hour. Based on the results, the acid catalyzed esterification had the best result when uses 5% oil weight of H2SO4 as catalyst and 30% oil weight of methanol to react with scum oil at 60°C for 1 hour. 3.4 Base catalyzed trans-esterification Figure 3c showed the conversion rate of the scum oil sample during base catalyzed transesterification reaction. The conversion rates of the scum oil sample with 0.5% and 1% potassium methoxide (CH3OK) increased during the first half hour, and after 30 minutes, reached a plateau. Therefore, the duration of base catalyzed reaction (60°C, 30% methanol) was set at 30 minutes to ensure reaction reaches equilibrium. The final conversion rate of scum oil sample with 1% CH3OK was 92.6% (this data indicates the ratio of produced FAME weight to raw oil used in base catalyze reaction). For 0.5 % CH3OK, conversion rate was 76.7%. CH3OK showed strong basicity in methanol, it certainly would be consumed by residual free fatty acid and water in the system. For instance, an oil sample with an acid value of 3 mg KOH·g-1oil will consume 0.3 % KOH (of oil weight). Hence, much less catalyst will be utilized. This phenomenon was more obvious when the base catalyst was less than 0.5%. For the base catalyzed trans-esterification, the best condition was using 1% oil weight of CH3OK as catalyst and 30% oil weight of methanol to react with scum oil at 60°C for 30 minutes. 3.5 Glycerol washing and layering process Figure 4 showed how different amounts of glycerol addition affected the oil
mixture layering. The oil mixture samples (11.5 g) used to carry on this experiment were collected from base catalyzed reaction product (part of methanol was removed by evaporation). X-axis was the ratio of additional glycerol to mixture oil and Y axis showed each layer’s weight and the position of boundary layers between phases. It showed that below 1:5 ratio, the system maintained in one phase status. From the ratios of 1:5-1:3.3, mixed oil separated into three layers. The upper layer was mainly methyl ester, while the middle and lower layers were mixtures of glycerol and methyl ester. Both the ratios of 1:2 and 1:1 leaded the mixed oil into two layers separation which meant glycerol and methyl ester phases were completely separated. Square dot showed the weight of glycerol admixture. It was obvious that the volume of the glycerol layer was more than the volume of glycerol added. Adding glycerol separated glycerol from the reaction broth of base catalyzed reaction. The final glycerol layer was composed of KOH (from CH3OK), pigments and other impurities that dissolved in the glycerol. During the glycerol washing, many impurities leached into the glycerol phase. The upper phase still had impurities in the FAME which must be separated by distillation. It was worth noting that KOH had a high solubility in glycerol phase (>16%, w/w, 60°C), however, in methyl ester the value was as low as less than 0.2%. This indicated that KOH had a strong tendency to remain in glycerol phase rather than stay in methyl ester phase after separation. For the glycerol washing step, it had the best effect when added one part of glycerol to two parts of oil mixture and washed the remaining glycerol and other impurities out of the oil. 3.6 Refining process During the distilling process, the bottom temperature was increased from 142-210 ˚C,
and top temperature was increased from 120-160˚C, the vacuum varied between 0.3-0.4 mmHg. The distillation yield was 88.2%. 3.7 GC analysis of biodiesel produced from scum The majority of FAMEs formed with different fatty acids were C14, C16 and C18. The weight percentages of each FAME were as follows: C14:0 (5.63%), C16:0 (32.45%), C16:1 (2.27%), C18:0 (15.59%), C18.1 and C18:2 (41.35%), and other (2.7%). Among all, octadecanoic acid (C18) and hexadecenoic acid (C16), which were main components of biodiesel, accounted for 91.66%. Saturated fatty acid (SFA) accounted for 53.67% which was much higher than refined vegetable oil biodiesel (6.5-15.7%, except palm oil) (Ramos et al, 2009). A higher SFA percentage leads to a higher Cetane Number (CN) and higher cloud point. The higher CN might improve the combustion performance of biodiesel, but the higher cloud point would affect the cold weather performance of the fuel. 3.7 Biodiesel yield and fuel quality When using dried and filtered scum oil as the starting material, the acid washing resulted about 90% yield, acid esterification-base transesterification-glycerol washing resulted about 89% yield, and the distillation resulted about 88% yield. So the total biodiesel yield from the dried and filtered scum oil was about 70% which was equivalent to about 1.24 ton per day biodiesel production or about 134,000 gallon biodiesel per year for the Metro Plant. Based on the biodiesel B100 selling price at the Twin city that ranged from $3.8-$4.9 in 2014 (Hennessy, 2014), the potential biodiesel production could bring about $518,000-$668,000 revenue to the Metro Plant at Saint Paul, MN.
Table 2 provided the fuel quality of the distilled B100 fuel; expect the sulfur content, total acid number and oxidation stability, the fuel passed all ASTM specification. The sulfur content was 33.6 ppm that was two times above ASTM limit. Seemingly sulfuric acid in acid washing process did not reduce the sulfur content effectively (see Table 1). Probably ion exchange resins or magnesium silicate powder as absorbent may reduce the sulfur content after the refining process. Total acid number was 1.43, three times higher than the ASTM limit. However, this was not a serious problem. In the batch distillation process, the acid number of samples from different distillation stage increased when distillation was progressing. It was due to the temperature rise. In this project, the last 5% of the distillate contained highest total acid value (>10 mg KOH/g oil) and the acid value of the front 95% of the distillate had acid value lower than the standard (2
Oil Methanol
Methanol recycle
Acid value
Neutralize
AV≤2
Base catalyzed transesterification 1% CH3OK, 30% methanol, react at 60 °C for 30 mins
Methanol recycle
Glycerol wash and settling separation
Glycerol
Glycerol recycle
Glycerol:oil=1:2, wash at 60 °C for 20 mins Raw glycerol Raw FAME
Waste glycerol Distillation
FAME
Figure 1. Process flow for converting scum to biodiesel (FAME).
Figure 2. Weight distribution and rate curve of filtering scum at 105°C.
a
b
c
Figure 3. Acid value to time curve in different acid catalyze condition (a), and the rate of AV reduce to time curve in different acid catalyze condition (b). Effect of the amount of base catalyst on the conversion rate with time elapse (c).
Figure 4. Effect of glycerol-oil ratio on layering in glycerol washing process.
Table 1. Element analysis of raw scum oil before and after acid washing.
Non-metal
Metallic
mg/kg
a-d
Scum oil
H2SO4 a
HCl
b
H3PO4 b
12.18±1.06c
Ca
1483.2±18.1
S0960-8524(15)00101-7 http://dx.doi.org/10.1016/j.biortech.2015.01.081 BITE 14513
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
29 November 2014 17 January 2015 19 January 2015
Please cite this article as: Bi, C., Min, M., Nie, Y., Xie, Q., Lu, Q., Deng, X., Anderson, E., Li, D., Ruan, R., Process Development for Scum to Biodiesel Conversion, Bioresource Technology (2015), doi: http://dx.doi.org/10.1016/ j.biortech.2015.01.081
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Process Development for Scum to Biodiesel Conversion Chonghao Bia,b, Min Minb*, Yong Niec, Qinglong Xieb, Qian Lub, Xiangyuan Dengb, Erik Andersonb, Dong Lia, Roger Ruanb* a
College of Engineering, China Agricultural University, Beijing 100083, China
b
Center for Biorefining, and Department of Bioproducts and Biosystems Engineering,
University of Minnesota, 1390 Eckles Avenue, St. Paul, MN, 55108, USA c
College of Chemical Engineering, Zhejiang University of Technology, Hangzhou
310014, China * Corresponding author. E-mail address: [email protected] (M. Min). [email protected] (R. Ruan), Tel.: +1 612 625 1710; fax: +1 612 624 3005 Abstract A novel process was developed for converting scum, a waste material from wastewater treatment facilities, to biodiesel. Scum is an oily waste that was skimmed from the surface of primary and secondary settling tanks in the wastewater treatment plants. Currently scum is treated either by anaerobic digestion or landfilling which raised several environmental issues. The newly developed process used a six-step method to convert scum to biodiesel, which would produce a higher value product. A combination of acid washing and acid catalyzed esterification was developed to remove soap and impurities while converting free fatty acids to methyl esters. A glycerol washing was used to help the separation of biodiesel and glycerin after base catalyzed transesterification. As a result, 70% of dried and filtered scum could be converted to biodiesel which is equivalent to about 134,000 gallon biodiesel per year for the Saint Paul Waste Water Treatment Plant at Minnesota.
Keywords: Biodiesel; Acid washing; Glycerol washing; Scum; Wastewater treatment plant 1. Introduction As a renewable fuel, biodiesel has become an attractive alternative for diesel fuel substitute. Vegetable oil has been the major source for global biodiesel production since 1990s (Krawczyk, 1996; EIA, 2014). It is still the largest proportion of feedstock of biodiesel in U.S. now. For instance, soybean oil is the largest biodiesel feedstock in U.S. during 2013 in which 5,507 million pounds was processed into biodiesel (48.8% of all kinds of feedstock); meanwhile, a total of 10,302 million pounds of soybean oil worldwide were used to produce biodiesel (U.S. Department of Energy, 2014). However, the main problem with making biodiesel a competitive fuel on the world market is its high cost. Currently, for vegetable based biodiesel, the feedstock alone takes about 70-80% of the total production cost (Cooperation and A. P. E., 2010). Many efforts have been focused using spent oil or waste oil, such as waste cooking oils, gutter oil, and wastes from animal or vegetable oil processing operations, to reduce the cost (Haas et al, 2006; Canakci & Sanli, 2008). Scum is a floatable material skimmed from the surface of primary and secondary settling tanks in wastewater treatment plants. It contains animal fat, vegetable oil, food wastes, plastic material, soaps, waxes and many other impurities discharged from restaurants, households and other facilities. Based on the report of the U.S. EPA’s Office of Solid Waste, approximate 1 to 3 billion gallons of waste grease, oil and fats are produced every year in the 30 metropolitan areas in the United States (Wiltsee, 1998). Some of those wastes (yellow grease) are collected in restaurants and
converted to biodiesel. However, approximately 60% of waste oils enter the sewer systems and end up in wastewater treatment plants. Due to a lower density than water, most of the oil floats on the surface of treatment facilities and conglomerates with other wastes to form scum, which makes scum a rich source of energy for recovery. The oil content of scum could be as high as 60%. The conventional technology for energy recovery is to co-process scum with sludge in anaerobic digestion where the produced biogas can generate electricity for plant use (Outwater & Tansel, 1994). However, this technology raises many problems in operation. For example, the scum floats on the top of the digester and forms a thick layer that impedes digester performance. As a result, many wastewater treatment plants choose to directly dispose scum in landfills. The scum disposal not only increases the cost of treatment facilities, but also causes many environmental problems. For instance, the Metropolitan Wastewater Treatment Plant at St. Paul. MN (Metro Plant) spends $100,000 a year just for land filling the scum. The landfill leachate could be a potential source for underground water pollution. In order to solve these problems, an alternative process/technology needs to be developed to recover energy from the scum and at the same time reduce the environmental impact due to landfilling scum. In this study, a new scum-to-biodiesel conversion process was developed so that the fat, free fatty acids and soap content in the scum can be successfully converted to ASTM-grade biodiesel. Since the final product is a liquid transportation fuel-biodiesel, it could gain a much higher economic value than biogas if generated through anaerobic digestion. Due to the abundance of nature gas, biogas has the selling price
close to $0.80/therm; while the biodiesel has the selling price about $3.2/therm which is three times higher than biogas (CenterPoint Energy, 2014). In addition, the byproducts such as glycerol and separated solids can be used as a heating source that not only decreases the energy input but also reduces the cost of the process. In a word, the current developed process has obvious economic and environmental advantages over other treatments and disposal processes. Currently, the most common process for biodiesel conversion is through base-catalyzed transesterification where triglyceride reacts with methanol to form fatty acid methyl ester (FAME). Potassium hydroxide (KOH) or sodium methoxide (CH3ONa) is the most common catalyst. However in the base-catalyzed transesterification process, feedstock must be free of water and free fatty acid (FFA), otherwise soap formation would reduce the biodiesel conversion yield. When using low grade feedstock such as yellow grease or brown grease, the FFA content can be as high as 30%. A pretreatment method which reduces acid value below 2 mg KOH per g of oil, must be applied before proceeding with any base-catalyzed transesterification process. Scum as a complex material contains solids, water, soap, fatty acids, oil and many other impurities. For maximizing the conversion yield, we explored several pretreatment methods in this project to clean up the scum oil. The objective of this project is to develop an effective pretreatment and conversion process to convert the scum to biodiesel.
2. Material and method 2.1 Material and reagent The scum samples were collected from the Metro Plant at St. Paul, MN. Sulfuric
acid (96.4%, AR) and Hydrochloric acid (36.5-38.0%, AR) were obtained from Mallinckrodt Baker, Inc. Paris, Kentucky. Phosphoric acid (85.0%, GR) was obtained from EM Industries, Inc. Potassium methoxide (>90.0%) was purchased from Alfa Aesar. Butylated hydroxytoluene (BHT), methanol (anhydrous, 99.8%), chloroform (99.8%) and diethyl ether (99.7%) were obtained from Sigma-Aldrich, Inc. BF3-methanol reagent (14% borontrifluoride, 86% methanol), sodium hydroxide 0.5 normal in methanol, heptane (HPLC grade) and glycerol (99.9%) were obtained from Thermo Fisher Scientific, Inc. potassium hydroxide concentrate (1.0 mol/L) was obtained from Fluka Analytical Sigma Aldrich Co. Ethanol (200 proof) was from Decon Laboratories, Inc. Distilled water was obtained from Premium Waters, Inc. MN, USA. 2.2 Process design The general process for the scum to biodiesel conversion was illustrated in Figure 2 and the main processes can be summarized according to the following six steps: FilteringAcid washingAcid catalyzed esterificationBase catalyzed transesterification Glycerol washing Oil refining
Firstly large particles in scum were separated by filtration. Then acid washing followed by gravitational settling was applied to convert soap (mainly (RCOO)2Ca and (RCOO)3Al) to free fatty acid (FFA), separate oil from water and water soluble/insoluble electrolytes, and further remove fine particles. The upper layer oil was collected and subject to acid catalyzed esterification and then base catalyzed transesterification processes to form FAME. The basic reactions were described in Equations (1)-(3). A glycerol washing was then performed to aid in separating the
glycerol and other impurities from the FAME. After separating the glycerol and methanol from the FAME, the crude FAME was then refined by fractional distillation to produce high grade biodiesel that can be directly used in diesel engines. Soap + Acid → Salt + FFA Catalyst FFA + MethanolAcid →Water + FAME Catalyst Triglyceri de + Methanol Base → Glycerol + FAME
(1) (2) (3)
2.2.1 Filtering process Scum oil was obtained in solid form at room temperature. It melted at about 40oC (depending on its components) and the viscosity of scum was negatively related with the temperature. For easing the separation of the scum oil from solid particles and at the same time drying the oil, a filtration process was conducted under a relative high temperature to reduce the viscosity of the scum oil. In a drying oven of 105 °C, 157 g of scum was loaded into a polyester mesh filter bag with pore size of 100 micron and a beaker was placed beneath the filter bag to receive the oil melted from the scum sample. The filtering process lasted for 24 hours. Weights of filtered oil and remaining solids were collected every 15 min during the first 8 hrs and at the end point of 24 hrs. The moisture content was calculated by subtracting the oil and solid weights from the total scum weight. 2.2.2 Acid washing Purpose of the acid washing was to convert any soap in scum to free fatty acids, maximize the biodiesel production yield, break emulsion for better water/oil separation, and further remove impurity from the scum oil. It also made the acid catalyzed esterification easier due to less impurity interference with the reactants.
H2SO4, HCl, and H3PO4, with two H+ strength 0.2 N and 1.2 N in water were compared for the scum oil pretreatment. The ratio of the scum oil-to-acid solution was 1:1 by weight. Raw scum oil was mixed with acid solution in a set of flask with condenser on top; a magnetic stirring water bath was applied to offer stirring power and keep the temperature of the reaction system at 60°C (Leung et al., 2010). After acid washing for 1 hour, the mixture was allowed to settle for about half an hour to collect the oil in the upper phase and sediment/water in the lower phase. The upper oil was then separated for the next process. 2.2.3 Acid catalyzed esterification Acid catalyzed esterification/transesterification is a general approach for feedstock with high free fatty acid and this process is not sensitive to moisture. However, when compared to base catalyzed transesterification, acid catalyzed esterification is much slower. Therefore for feedstock high in free fatty acid, acid catalyzed esterification followed by base catalyzed transesterification was generally used and the acid catalyzed esterification became one of the pretreatment methods for reducing the FFA in feedstock. After acid washing, scum oil that contains residual water and high free fatty acid converted from soap had an average acid value (AV) of 21 mg KOH·g-1 oil. The main purpose of this step is to reduce the AV to below 2 mg KOH·g-1 oil through the acid esterification reaction described in Equation (2) and to prepare the oil for base catalyzed transesterification. In the biodiesel industry, the AV of feedstock with high fatty acid should be reduced below 2 mg KOH·g-1oil before base catalyzed transesterification (Zhang & Jiang, 2008; Cao et al., 2008). Regarding the acid catalyzed reaction, H2SO4 has been the most commonly used and widely
investigated catalyst (Lotero et al., 2005). During this process, methanol (30% of oil weight) and sulfuric acid (3% and 5% of oil weight) were added to the oil produced from last step. In a flask with a condenser at the top, a magnetic stirring water bath was applied to offer stirring power and keep the temperature of the reaction system at 60 °C (Canakci & Van Gerpen, 1999). The acid value of the oil phase was expected to drop with ongoing esterification. The acid value of oil phase was monitored every 15 minutes in the first hour, and after 150 minutes. After the reaction was completed, the mixture was settled for one hour to separate into two layers. The upper phase was the oil layer that would be collected and dried in 105°C oven over night and used for next base catalyzed reaction. The majority of methanol and all the acid were in the lower phase. 2.2.4 Base catalyzed process Due to the fast reaction rate, base catalyzed transesterification is the general process for low water feedstock (water content less than 0.05% w.t.) with low free fatty acid (AV less than 2). After acid catalyzed esterification and separation, the scum oil would meet the criteria as a feedstock for base catalyzed esterification. The reactor was as same as described in section 2.2.2. Methanol (30% of oil weight) was added into the corresponding amount of oil (obtained from acid catalyzed step) and potassium methoxide (CH3OK, 0.5% and 1% of oil weight) acted as catalyst to determine the conversion efficiency (Meher et al, 2006). Samples were collected every 10 minutes for an hour. For analyzing the FAME content, the methanol left in samples was flash evaporated first, and then glycerol washing procedure described in
section 2.2.5 was applied to induce separation of glycerol and FAME; then the supernatant containing FAME was specified by GC-MS analysis described in section 2.3.3. 2.2.5 Glycerol washing and layering In principle, methyl ester and glycerol is immiscible. When using a better quality feedstock, such as soybean oil, the FAME and glycerol would separate into two layers by gravity settling after the transesterification reaction. However, scum oil will form a stable single-phase system even after 24 hours settling at 60 oC. It was suspected that some organic compounds (for instance, surfactants or other impurities) and the soap could increase the inter-solubility of glycerol and methyl ester; therefore glycerol could not be separated. The product mixture has to be separated and the base catalyst must be removed before distillation process, otherwise the product mixture will subject to reverse reaction and re-saponification. It was found that by adding extra glycerol to the mixture could actually improve FAME/glycerol separation. Therefore, a process step that was called “glycerol washing” was applied. As a result, glycerol washing could not only bring the oil mixture to layers but also wash out most of the base catalyst and many impurities in the scum. After the base catalyzed reaction, the methanol in the mixture oil was removed by rotary evaporator. Glycerol was then added into the mixture oil with different ratio (glycerol:oil = 1:20; 1:10; 1:5; 1:1.33; 1:2; 1:1 by weight) and mixed thoroughly by magnetic stirring at 60°C for 20 minutes, and then kept settling for 12 hours at the same temperature. The upper layer contained mainly FAME and the lower layer was
glycerol and base catalyst. 2.2.6 Distillation Upper layer outcome from last step contained about 90% FAME (according to the base catalyzed condition). This FAME mixture was introduced to a customized vacuum rectification column system and refined FAME can be obtained. 2.3 Analytical methods 2.3.1 Elementals analysis Samples from acid washing process were submitted to Soil Testing & Research Analytical Laboratory at the University of Minnesota (St. Paul, MN) to perform elemental analysis. An inductively coupled plasma atomic emission spectrometer (Perkin Elmer Optima 3000, USA) was applied on all the samples to provide a total elemental analysis of acid washed scum oil samples (Yeomans & Bremne, 1991; Lee et al, 1996; Fassel & Kniseley, 1974). 2.3.2 Acid value (AV) determination Acid value (mg KOH/g oil) is an indicator of the free fatty acid content in the scum oil. Standard Test Method for Acid Value of Fatty Acids and Polymerized Fatty Acids ASTM D1980-87 (1998) were used to determine the acid value. By monitoring changes of acid value during acid catalyzed esterification process, progress of reaction was determined. 2.3.3 Fatty acid methyl ester (FAME) analysis The BF3 catalysis method was applied from AOCS Official Method Ce 2-66 (1997). This method proposes excess reactant and excess catalyst to make sure that all the fatty acid in sample was transformed into FAME.
FAME in all samples was analyzed by GC-MS using the method from Li et al., (2011). Oil samples (about 100 mg) were weighed into 10 ml volumetric flasks and diluted with chloroform to 10 ml. The GC-MS (Agilent 7890-5975C, USA) equipped with a HP-5 column and a mass detector was used. Chromatographic data were recorded and addressed by a built-in Agilent data analysis software. Components were identified in NIST Mass Spectral Database. Quantification was carried out by comparing the peak area with that of the GLC-10 and GLC-30 standard mixtures (Sigma-Aldrich, Co.). 2.4 Biodiesel fuel testing The distilled biodiesel fuel was sent out and tested by Iowa Central Fuel Testing Laboratory (Fort Dodge, IA) using ASTM D 6751/BQ-9000 Full Spec Test Package for B100. The results were listed in Table 2. 3 Result and discussion 3.1 Filtering process At the end of the process, water, oil, and solids amounted 28.2%, 56.0%, and 15.8% of the total scum weight respectively. The specific filtering curves were illustrated in Figure 2. The weight changes (% of total scum weight) of the filtered oil, the water evaporated and the solid left in the filter bag were recorded over time. Water content started to decline at the first 15 minutes and the total water evaporated at 24 hours was 28.2% of the scum weight. The dash line in the Figure 2 indicated the water evaporation rate. It showed that the highest rate of water evaporation was 0.041 %/min observed at 2.5 hours. The rate declined to below 0.023 %/min after 5.75 hours. Filtered oil was received in the beaker, continuously increasing to reach 56.0%
after 24 hours. Rate of oil filtration curve was a derivative of oil weight percentage over time. It was observed from this curve that the rate of oil collection was very low during the first hour, then climb up to the maximum of 0.702 %/min at 2 hours, after that, the rate decreased to below 0.023 %/min after 5.75 hours. Weight of solids in the filter bag was also shown in Figure 2. This curve dropped from 100% to 15.8% indicating 15.8% solids in scum. The curvature of the solids curve was negatively correlated to the oil curve during the first 6 hours which indicated that the weight loss of scum solids was mainly due to the separated oil. During 6-24 hours the solids curve was positively related to the water curve which indicated that the solid weight loss during 6-24 hours was mainly due to the water evaporation. At the time of 6 hours, oil filtered was 52.2% of the total scum weight, which counted 93.2% of total oil. However, at time of 8 hours, 98.0% of the total oil (counted 54.9% of total scum weight) could be filtered out. When considering maximizing oil yield versus maximizing energy efficient, process time and capital cost, there was a tradeoff point in this filtering process. For maximizing oil yield, longer heating and filtering time would be better, however it also increased processing time, energy and capital cost. When using shorter time, oil yield would be slightly decreased, but the energy for evaporating water would be reduced by half and the size of the filtering equipment would be smaller due to faster processing time. The water content left in oil would be further reduced in the acid washing step. 3.2 Acid washing It was found that acid solution with 0.2 N couldn’t acidize scum oil completely; while the acid solution with 1.2 N could acidize scum oil completely, leaving 0.3-0.4
mol/L H+ surplus after this treatment. It was observed that most sediment was formed when the H2SO4 was added, some sediment with H3PO4 addition, however, no sediment appeared with HCl added system. This was because SO42- and PO43- were more likely to form insoluble salts with the metallic elements, such as CaSO4, Fe2(SO4)3, etc. When using dried and filtered scum oil as the starting material, the acid washing would account for 10% of mass loss due to remove all these metallic impurities. Table 1 showed the major impurity elements in the scum oil before and after acid washing. Many soap component elements such as Calcium, Iron, Aluminum, Magnesium, Zinc etc. were decreased or eliminated after acid washing. For example, the Ca2+ was dropped from 1483 mg/kg to less than 0.43 mg/kg after acid washing. Nitrogen content is an indicator of surfactant content as it may forms lots of active hydrophilic groups and surfactants, such as -NH3, -N2+, etc (Ananthapadmanabhan, 1993; Hayashita et al, 1994). After H2SO4 treatment, the content of the nitrogen content dropped from 1990 mg/kg to 1200 mg/kg, which was significantly superior to HCl and H3PO4. In conclusion, H2SO4 was considered to be the most suitable acid washing reagent for scum oil pretreatment. The acid washing was more effective for removing inorganic impurities than for those organic compounds containing N, P, and S. After acid washing, N content still remained 1200 mg/Kg in scum, 80% of S wasn’t affected, only P was mostly removed. These organic impurities might interfere with subsequent processes. Currently 6% sulfuric acid solution mixed with scum oil with 1:1 ratio by weight was used for the acid washing step. 3.3 Acid catalyzed esterification
Figure 3a showed the decrease of acid value of the different oil samples with various concentrations of H2SO4. The dotted curve was made using unwashed scum oil sample (just after scum filtration) with 3% of H2SO4 addition, and the circle and triangle curves were made using acid washed oil samples with 3% and 5% of H2SO4 addition respectively. It showed that the AVs of all three experiments decreased during acid catalyzed esterification reaction. AVs of unwashed scum oil with 3% catalyst had slowest decrease and ended at 8.07 mg KOH·g-1oil, meanwhile, AVs of acid washed oil with 5% catalyst had fastest decrease and ended at 2.10 mg KOH·g-1 oil. Sample of acid washed oil with 3% catalyst gained a final AV of 4.27 mg KOH·g-1 oil. The esterification reaction of the unwashed scum oil wasn’t as complete as the acid washed oil for same catalyst admixture. This was mainly because part of the sulfuric acid reacted with the soap in the scum and produced water. The consumed sulfuric acid reduced the catalyst strength and water also hindered the esterification reaction since the esterification reaction was a reversible reaction and water was one of the reaction products. That was why acid washed scum oil had better performance during the esterification reaction. It was found that acid washing and 5% catalyst was necessary for acid catalyzed esterification reaction. Figure 3b indicated the rate of AV decreasing by using derivative of AV to time. It was observed that reaction rates of all three experiments decreased as time elapsed which indicated that reactions achieved equilibrium gradually. During the first 15 min, reaction rate of acid washed oil (0.94 mg KOH·g-1 oil·min-1, 0.81 mg KOH·g-1 oil·min-1) was much higher than unacidized scum oil (0.29 mg KOH·g-1oil·min-1). However, they dropped faster in the next quarter of hour, and after 1 h, the reaction
rate of these three samples all dropped below 0.05 mg KOH·g-1oil·min-1, which was too slow for industrial application. Therefore it was reasonable to set the duration of acid catalyze esterification reaction at 1 hour. Based on the results, the acid catalyzed esterification had the best result when uses 5% oil weight of H2SO4 as catalyst and 30% oil weight of methanol to react with scum oil at 60°C for 1 hour. 3.4 Base catalyzed trans-esterification Figure 3c showed the conversion rate of the scum oil sample during base catalyzed transesterification reaction. The conversion rates of the scum oil sample with 0.5% and 1% potassium methoxide (CH3OK) increased during the first half hour, and after 30 minutes, reached a plateau. Therefore, the duration of base catalyzed reaction (60°C, 30% methanol) was set at 30 minutes to ensure reaction reaches equilibrium. The final conversion rate of scum oil sample with 1% CH3OK was 92.6% (this data indicates the ratio of produced FAME weight to raw oil used in base catalyze reaction). For 0.5 % CH3OK, conversion rate was 76.7%. CH3OK showed strong basicity in methanol, it certainly would be consumed by residual free fatty acid and water in the system. For instance, an oil sample with an acid value of 3 mg KOH·g-1oil will consume 0.3 % KOH (of oil weight). Hence, much less catalyst will be utilized. This phenomenon was more obvious when the base catalyst was less than 0.5%. For the base catalyzed trans-esterification, the best condition was using 1% oil weight of CH3OK as catalyst and 30% oil weight of methanol to react with scum oil at 60°C for 30 minutes. 3.5 Glycerol washing and layering process Figure 4 showed how different amounts of glycerol addition affected the oil
mixture layering. The oil mixture samples (11.5 g) used to carry on this experiment were collected from base catalyzed reaction product (part of methanol was removed by evaporation). X-axis was the ratio of additional glycerol to mixture oil and Y axis showed each layer’s weight and the position of boundary layers between phases. It showed that below 1:5 ratio, the system maintained in one phase status. From the ratios of 1:5-1:3.3, mixed oil separated into three layers. The upper layer was mainly methyl ester, while the middle and lower layers were mixtures of glycerol and methyl ester. Both the ratios of 1:2 and 1:1 leaded the mixed oil into two layers separation which meant glycerol and methyl ester phases were completely separated. Square dot showed the weight of glycerol admixture. It was obvious that the volume of the glycerol layer was more than the volume of glycerol added. Adding glycerol separated glycerol from the reaction broth of base catalyzed reaction. The final glycerol layer was composed of KOH (from CH3OK), pigments and other impurities that dissolved in the glycerol. During the glycerol washing, many impurities leached into the glycerol phase. The upper phase still had impurities in the FAME which must be separated by distillation. It was worth noting that KOH had a high solubility in glycerol phase (>16%, w/w, 60°C), however, in methyl ester the value was as low as less than 0.2%. This indicated that KOH had a strong tendency to remain in glycerol phase rather than stay in methyl ester phase after separation. For the glycerol washing step, it had the best effect when added one part of glycerol to two parts of oil mixture and washed the remaining glycerol and other impurities out of the oil. 3.6 Refining process During the distilling process, the bottom temperature was increased from 142-210 ˚C,
and top temperature was increased from 120-160˚C, the vacuum varied between 0.3-0.4 mmHg. The distillation yield was 88.2%. 3.7 GC analysis of biodiesel produced from scum The majority of FAMEs formed with different fatty acids were C14, C16 and C18. The weight percentages of each FAME were as follows: C14:0 (5.63%), C16:0 (32.45%), C16:1 (2.27%), C18:0 (15.59%), C18.1 and C18:2 (41.35%), and other (2.7%). Among all, octadecanoic acid (C18) and hexadecenoic acid (C16), which were main components of biodiesel, accounted for 91.66%. Saturated fatty acid (SFA) accounted for 53.67% which was much higher than refined vegetable oil biodiesel (6.5-15.7%, except palm oil) (Ramos et al, 2009). A higher SFA percentage leads to a higher Cetane Number (CN) and higher cloud point. The higher CN might improve the combustion performance of biodiesel, but the higher cloud point would affect the cold weather performance of the fuel. 3.7 Biodiesel yield and fuel quality When using dried and filtered scum oil as the starting material, the acid washing resulted about 90% yield, acid esterification-base transesterification-glycerol washing resulted about 89% yield, and the distillation resulted about 88% yield. So the total biodiesel yield from the dried and filtered scum oil was about 70% which was equivalent to about 1.24 ton per day biodiesel production or about 134,000 gallon biodiesel per year for the Metro Plant. Based on the biodiesel B100 selling price at the Twin city that ranged from $3.8-$4.9 in 2014 (Hennessy, 2014), the potential biodiesel production could bring about $518,000-$668,000 revenue to the Metro Plant at Saint Paul, MN.
Table 2 provided the fuel quality of the distilled B100 fuel; expect the sulfur content, total acid number and oxidation stability, the fuel passed all ASTM specification. The sulfur content was 33.6 ppm that was two times above ASTM limit. Seemingly sulfuric acid in acid washing process did not reduce the sulfur content effectively (see Table 1). Probably ion exchange resins or magnesium silicate powder as absorbent may reduce the sulfur content after the refining process. Total acid number was 1.43, three times higher than the ASTM limit. However, this was not a serious problem. In the batch distillation process, the acid number of samples from different distillation stage increased when distillation was progressing. It was due to the temperature rise. In this project, the last 5% of the distillate contained highest total acid value (>10 mg KOH/g oil) and the acid value of the front 95% of the distillate had acid value lower than the standard (2
Oil Methanol
Methanol recycle
Acid value
Neutralize
AV≤2
Base catalyzed transesterification 1% CH3OK, 30% methanol, react at 60 °C for 30 mins
Methanol recycle
Glycerol wash and settling separation
Glycerol
Glycerol recycle
Glycerol:oil=1:2, wash at 60 °C for 20 mins Raw glycerol Raw FAME
Waste glycerol Distillation
FAME
Figure 1. Process flow for converting scum to biodiesel (FAME).
Figure 2. Weight distribution and rate curve of filtering scum at 105°C.
a
b
c
Figure 3. Acid value to time curve in different acid catalyze condition (a), and the rate of AV reduce to time curve in different acid catalyze condition (b). Effect of the amount of base catalyst on the conversion rate with time elapse (c).
Figure 4. Effect of glycerol-oil ratio on layering in glycerol washing process.
Table 1. Element analysis of raw scum oil before and after acid washing.
Non-metal
Metallic
mg/kg
a-d
Scum oil
H2SO4 a
HCl
b
H3PO4 b
12.18±1.06c
Ca
1483.2±18.1
