V-/EPA United States Environmental Protection Agency Municipal Environmental Research Laboratory Cincinnati OH 45268 Research and Development EPA-600/S2-81-242 Dec. 1981 Project Summary Physical and Chemical Characteristics of Synthetic Asphalt Produced from Liquefaction of Sewage Sludge J. M. Donovan, R. K. Miller, T. R. Batter, and R. P. Lottman Direct thermochemical liquefaction of primary undigested municipal sewage sludge was carried out to produce a low molecular weight steam-volatile oil, a high molecular weight synthetic asphalt, and a residual char cake. The latter product is capable of supplying the thermal energy requirements of the conversion process. The steam- volatile oil has immediate value as a synthetic fuel oil. The synthetic asphalt may prove to be a useful cement for paving with further research, or it can be used as a fuel or coking stock. It is outwardly similar to petroleum asphalt, but chemically different. The thermochemical liquefaction process should be capable of opera- ting in a technical and environmentally acceptable manner in conjunction with many existing wastewater treat- ment facilities. The overall feasibility of the process depends on the value of the oil and synthetic asphalt products as petroleum replacements and on the costs associated with disposal of sludge. Projected economics indicate that the process has considerable promise for many potential sites in the United States at the present time. This Project Summary was develop- ed by EPA's Municipal Environmental Research Laboratory, Cincinnati. OH, to announce key findings of the research project that is fully docu- mented in a separate report of the same title fsee Project Report ordering information at back). Introduction Disposal of sewage sludge is an increasing problem for many munici- palities in the United States. Currently, there is a need to implement alternative disposal technologies. The alternative technologies need to be energy efficient and,.if possible, some product of value should be recovered from sludge. Direct thermochemical liquefaction has the capability of meeting these requirements. Thermochemical liquefaction investi- gations by various authors have shown that organic biomass can be converted to bitumen, heavy oil, and distillate oils having combustion heats closely approximating petroleum products. The utility of the high molecular weight fraction of the liquefaction product (because it is a substantial part of the liquefaction products) is often questioned. To this end, the present study was conducted to determine the value of this fraction as synthetic asphalt for use as a paving cement. A synthetic asphalt product would be valuable since petroleum asphalt has increased in value with escalating ------- petroleum costs and it is being increas- ingly used (e.g., production of olefins) as a petrochemical feed stock. The low molecular weight product of biomass liquefaction has always been viewed as a valuable synthetic fuel. In the case of sludge liquefaction, the low molecular weight oil product was conveniently separated during the reaction period by steam auto-distilla- tion. This steam-volatile oil was obtained in significant quantities using a relatively simple process. This process also led to simplified separation of excess water present because of the use of wet sludge. Thermochemical Liquefaction of Primary Sewage Sludge Previous work in the area of direct thermochemical liquefaction of sewage sludge and other forms of biomass has resulted in the production of high molecular weight fractions that are similar in appearance to petroleum asphalt. A high molecular weight fraction derived from pure cellulose in our Battelle-Northwest laboratory showed promise as a replacement or extender for petroleum asphalt. Because of the high cost of pure cellu- lose and other biomass forms, other alternative feed materials were suggested. Primary undigested sewage sludge, because of its low or negative cost, was a logical choice. Experimental Procedure Although the intended feed stock for the experimental work was to have been dried primary sludge from Honolulu, the actual feed stock was fresh, primary sludge dewatered with polymer. The fresh sludge was preserved by adding chloroform and, later, by freezing. It was then sent by air to our laboratory at Richland, Washington. Fresh sludge received at our laboratory was prepared for reaction by adding a base (either Na2C03 or CaO). This material was then loaded into an Inconel linerand blanketed with an inert gas (Argon). The liner was sealed and placed into an autoclave. Water was added between the liner and autoclave to assist in heat transfer and to balance the differential pressure across the liner that would result from heating to reaction temperature. Pressure inside the liner was controlled to equal the vapor pressure of water in the annular space; this prevented either implosion or explosion of the liner. After being prepared in this manner, the autoclave was heated to reaction temperature (2 to 3 hours), held at reaction temperature for 1 hour, and cooled 4 to 6 hours. During the heating, reaction, and cooling periods, gas and steam were emitted from the liner because of the pressure in the liner caused by the gas generation accom- panying thermochemical liquefaction. The predominant gas formed and dis- charged was carbon dioxide, but steam and steam-volatile oil were also dis- charged with the gas. These gases were condensed in a water trap and saved for later analysis. Although hydrogen sulfide was monitored during the reactions, none was detected. After cooling, the autoclave was opened, the liner removed and opened, and the product taken out. This product could be separated into an aqueous supernatant liquid and a char cake either by settling or centrifugation. In our laboratory work, the latter was more convenient and was used most often. The supernatant liquid was analyzed for its volatile constituents and-was sub- jected to treatability analyses. The char cake contained the high molecular weight fraction of the liquefaction product, which was intended to become synthetic asphalt. As a result of being intermixed with the char and ash, the high molecular weight fraction was not acceptable for use as a replacement or extender for petroleum asphalt since it would not melt when heated nor mix with heated petroleum asphalt. To overcome the problems caused by the char and ash, the high molecular weight fraction was separated by sol- vent extraction. Previous liquefaction work, and specifically work with lique- faction products from cellulose, indi- cated that acetone was an acceptable solvent. Soxhlet extractors were used, and extraction times were 8 to 24 hours. The extracted char cake was crumbly after the solvent was removed in a rotary evaporator under vacuum. The resulting material, whose appearance resembled heavy crude oil, was not acceptable for direct use as a petroleum asphalt replacement or extender in preliminary work because its viscosity was not high enough. The use of vacuum distillation to remove residual low molecular weight liquefaction products from the extracted product solved the low viscosity problem. Although only a very small amount of low molecular weight material was removed, the viscosity of the product was substantially increased. Synthetic asphalt samples prepared in this manner were sent to Dr. R. Lottman at the University of Idaho for testing and analysis as paving cements. Liquefaction Test Results Reaction conditions and resulting product yields are given in Table 1. Pressure may be estimated from the vapor pressure - temperature relation- ship for water. The mean total yield (oil + asphalt) for conversions carried out at 320°C with Na2CO3 was 20.4% with a standard deviation of 7.1%. Total yield seemed to increase with temperature since the yields at 295°C and 345°C differ from the mean by 1.5 and 2.3 standard deviations, respectively. Several runs at 320°C were done to provide the University of Idaho with a large, consistent sample for a complete series of asphalt paving material testing. Unfortunately, this limited the. amount of data taken at other points| which in turn limited our ability to des- cribe yield as a function of temperature. Lime (CaO) is apparently an excellent liquefaction adjunct* having yields approximately equal to those of carbonate under the same experimental conditions. Unfortunately, asphalt testing results indicated that the one sample we produced using lime was an inferior product. Because of the inherent variations in using sludge as a raw material, perhaps additional trials with lime would give better results. Elemental analyses of the products produced by liquefaction (Table 2) allows a comparison with conventional feeds and petroleum asphalt. The sulfur concentration in the synthetic asphalt and oil samples places them in approxi- mately a grade four heating oil category. Both the sulfur and nitrogen contained in the synthetic asphalt samples and oils result from the use of sludge. Our analyses of volatiles (reported next) show that sulfur and nitrogen are "Previous liquefaction work has referred to alkaline adjuncts as catalysts However, our previous work has indicated that the alkali is a reactant, and that the overall reaction is highly dependent on pH. Because of this, alkalis used for liquefaction an referred to as adjuncts in this report ------- Table 1. Reaction Conditions and Yields Experimental Designation HS-3 HS-4 HS-5 HS-6 HS-7 HS-8 HS-9 HS-10 Dry Ash-Free Sludge, Kg 2.61 2.52 2.41 3.99 2.01 2.28 2.49 1.33 Steam ' Volatile Oil gm 330 480 -0- 450 -0- 250 450 250 Synthetic Asphalt, gm 268 440 250 330 310 50 430 164 Total Yield %* 23 37 10 20 15 13 35 31 Reaction Temp. °C 320 345 295 320 320 320 320 320 5% by Weight NatCOs /Va2C03 /Va2CO3 /Va2CO3 Na2C03 Na2C03 CaO NatC03 * Weight of light oil and synthetic asphalt as percent of dry, ash-free sludge. Table 2. Elemental Analyses (Percent by WeightJ Element C H N 0 S Petroleum Asphalt AC-10 87. 11. 0.4 1. - Synthetic* Asphalt 74. 10. 4. 9. 0.8 Steam-Volatile^ Oil 77. 12. 3. 7. 0.9 Char Cake HS-10 26. 3. 0.9 10. - * Average values from Experiments HS-4. HS-5, HS-6, HS-9, and HS-10. t Average values from Experiments HS-9 and HS-10. substituted into the aromatic and ali- phatic constituents of the volatile products. As a result of this, we suspect that the synthetic asphalt and oil products also contain a wide range of substitutecLsulfur and nitrogen com- pounds. For use as fuel, this is probably of minor concern. For use as synthetic asphalt, however, the presence of ni- trogen may be limiting because of poten- tial interactions between petroleum asphalt and the more polar synthetic asphalt (if synthetic asphalt is to be blended with petroleum asphalt). Also, since nitrogen ischemically substituted and since the average molecular weight of synthetic asphalt is lower than petro- leum, there may be a more pronounced tendency for synthetic asphalt to be soluble in water and for water solubility _m synthetic asphalt to be higher than hat in petroleum asphalt. Table 3 illustrates heats of combustion for steam-volatile oils, synthetic asphalt and char cake. The low combustion heat of char cake (Table 3) is due to the large concentration of ash in the char cake. The HS-10 char cake contained- 60% ash before combustion. Heats of combustion for the synthetic asphalt and oil reported in Table 3 are approximately 90% of the values for petroleum equivalents. Although not measured, the viscosity of steam-volatile oil was approximately that of No. 2 heating oil, judging from its pouring properties at room temperature. Synthetic Asphalt Test Results Data from testing the various synthetic asphalts produced by sludge liquefaction varied widely. The data presented here are for sample HS-7, one of the better samples. Before proceeding with the testing, the HS-7 synthetic asphalt sample was melted and washed with hot water. The sample lost 30% of its original weight during washing; olive-brown solubles were removed in the wash water. The synthetic asphalt sample was then dried at 60°C. The washed sample began to melt at 50°C and became completely liquid at 80°C. It was sticky and adhered well to cardboard when subjected to freezing temperatures. When frozen, the sample was brittle but no more so than petro- leum asphalt. When exposed to room temperature, it regained its putty-like consistency much more rapidly than petroleum asphalt. Part of the sample was used directly with aggregate, and another sample was prepared from a mixture of 50% high-grade (AC-10) petroleum asphalt and 50% synthetic asphalt. During blending, an adverse reaction between the petroleum and synthetic asphalts was noticed; there seemed to be a rapid increase in apparent viscosity when the two were mixed. Both the 100% synthetic and the 50- 50 blend were mixed with hot aggregate and oven cured at 60°C for 16 hours. When the loose mixes were removed from the oven, the blend appeared duller in appearance. After the two mixes were subjected to heating arid compaction, they were cooled to room temperature and bent and pulled apart for a preliminary assessment of adhesion. The 100% synthetic asphalt showed good adhe- sion whereas the 50-50 blend mix was ------- Heats of Combustion Synthetic Asphalt HS-9 HS-10* Calculation based on total mass including ash. poor and crumbly. Further testing Table 3. excluded the blend because of its poor performance. Compacted mix specimens were made with petroleum asphalt and with Type 100% synthetic asphalt. Both were sub- jected to dry and accelerated moisture conditioning (vacuum saturation followed by 0°C freezing and 60°C water soaking). Mechanical properties of tensile splitting strength and resilient modulus were obtained for dry speci- mens, moisture saturated specimens, and specimens after accelerated moisture conditioning. The mechanical property values for the 100% synthetic mix are close to those for the petroleum asphalt mix. The synthetic mix retained only 62% of its dry tensile strength after accelerated moisture conditioning, compared with 79% retained strength for petroleum asphalt. Even though no stripping was noted for the synthetic mix specimen, its low retained tensile strength puts it on the lower end of the scale for petroleum asphalt. The resilient modulus, however, was relatively unaffected by moisture. The synthetic mix retained 96% of its dry modulus when saturated and 100% of its dry modulus after accel- erated conditioning Petroleum asphalt under the same conditions retained 103% and 76%, respectively. Although the synthetic asphalt appeared to be duller and more putty- like than petroleum asphalt, it showed good bonding behavior with no strip- ping. The synthetic asphalt was inherently different from petroleum *100 Short ton/day Table 4. Payback Period for Several Sludge Liquefaction Process Options Steam-Volatile Oil Leachate Char Cake HS-9 HS-10 HS-6 HS-10 cal/g Btu/lb 8,730 15,700 9,060 16,300 9,380 16,900 9,410 14,000 7.760 5.400* 3.020* asphalt but showed promise because of its mechanical performance during testing. Blends of synthetic asphalt and petroleum asphalt will require more investigation because of the adverse effect they have on each other when mixed. Conceptual Design and Cost Estimates for a 91 Tonne/Day* Commercial Sludge Thermo- chemical Liquefaction Plant The conceptual plant design and cost estimate were made to help determine the overall feasibility of the process and to encourage further research and development on continuous liquefac- tion of sewage sludge. Estimates, made on the basis of the best available information at this time, are tentative since there are no pilot plant data, or even bench-scale continuous process data, to support equipment specifica- tions. In the flowsheet of the process (Figure 1), approximate product flows and temperatures are given for reference. A more detailed analysis of heat and mass balance is not currently warranted since heat of reaction, product yields, and processability of some of the streams in a continuous process are unknown. Process Design A plant capacity of 91 dry tonne/day (100 short ton/day) was chosen as being representative of the volume of primary sludge produced in large wastewater treatment plants around the United States. Therefore, the estimates are for commercial plants, not for pilot or dem- onstration plants. The process would operate on I primary sludge dewatered to at least" 30% solids. Other sludges could be used, but for these estimates, only primary sludge was considered. Dewatering to 30% solids is already practiced in many wastewater treat- ment plants. Many of these plants already use lime to condition the sludge $ Mil/ion Process With Dewatering Vith Centrifuge 1 vnha/t \&frjl IGIL Recovery Without Dewatering Centrifuge With Dewatering Without Centrifuge 1 cn/i^/f i ofJf Idlt Recover Without Dewatering Centrifuge Facility Cost 9.8 7.8 7.8 5.8 Manufacturing Costs 2.97 2.68 2.58 2.29 Sludge Disposal Credit 1.98 1.98 1.98 1.98 Oil/Asphalt Revenue 1.79 1.79 1.36 1.36 Total Revenue 3.77 3.77 3.34 3.34 Simple Payback (Years) 12 7 10 6 ------- Primary Sludge 91 Tonne/day Dry Basis Centrifuge 25°l/min t Vapor-Liquid Separator J[ 1 129 l/min Waste Water Gas 121 l/min | to Vent High Pressure Slurry Feeder Water to Secondary Treatment Light Oil Centrifuge Waste Water Char Cake Dowtherm A Condensate X Dowtherm A Vapor 350°C Cake Desolventizer 4 Dowtherm Vaporizer Solvent Recovery Ash to Disposal Heavy Oil Washer Light Oil to Storage Waste Water Synthetic Asphalt to Storage 6.4 kg/min Figure 1. Preliminary schematic for a sludge liquefaction plant Waste Water for dewatering, and, in many cases, add 10% or more lime on a dry weight basis. In experiment HS-9, lime proved to be an effective adjunct for producing steam-volatile oil, but rated poorly for production of synthetic asphalt. There- fore, lime-treated sludge can be used directly without further treatment if the primary goal is fuel production. Syn- thetic asphalt from lime-treated sludge would require more investigation based an current results. Because the thermochemical lique- faction yields varied significantly, in the design and estimation work, we used the sludge characteristics and yields obtained from HS-10 as representative. With the use of these data, the yields from a plant processing 91 tonne/day would be: Synthetic asphalt: Steam-volatile oil: Char cake: 9,000 Kg/day 13,800 Kg/day 37,000 Kg/day At a density of a pproxi mately 830 g m / L, the light oil would amount to approxi- mately 16,300 L, or just slightly greater than 100 petroleum barrels/day. Process heat requirements for this plant would be supplied by combustion of the residual char cake. The extracted cake, with heat of combustion of approximately 3,020 cal/gm, would be produced at the rate of 26 Kg/min and, therefore, would be capable of supplying 1.56 X 10'° J/hr (1.5 X 107 ------- Btu/hr) of process heat at a combustion efficiency of 80%. This amount would be more than adequate to supply process heat requirements, especially if heat recovery were used on the main stream coming out of the liquefaction reactor. The greatest heat requirement by far would be to heat the liquefaction reactor, even with heat recovery. For the flow design presented in Figure 1, the liquefaction reactor would require heat input of 1.17 X 1010 J/hr. Other minor heat requirements for solvent extraction and recycle are estimated to be less than 04 X 10'° J/hr. Total energy requirements therefore are balanced by the available energy in the residual char cake. Energy from com- bustion of the char cake would be supplied to the liquefaction reactor by m Dowtherm* vaporizer. Other process utility requirements would be limited to electrical power of about 1.8 X 106 KWH/yr and a small amount of cooling water. Estimates given for capital equipment and manufacturing costs are derived from a plant that would already include sludge dewatering equipment (a centri- fuge) and solvent extraction to recover synthetic asphalt, since many sites have dewatering equipment m place and since some plants may elect not to recover synthetic asphalt. Simple Payback Period for Different Process Options A major cost in this process would be associated with primary sludge de- watering. Estimates given in Table 4 show that if investment for this process is not required, the payback period for the plant will decrease from 10 to 12 years to 6 to 7 years. Although synthetic asphalt may be a valuable product when produced from sewage sludge, these payback esti- mates (Table 4) show that the payback period is actually shorter if asphalt were not recovered because the solvent extraction process adds more capital and operating cost than could currently be recovered by sale of synthetic asphalt. If synthetic asphalt were not recovered, it would be contained in the char cake and simply burned. The viabil- ity of synthetic asphalt recovery will change depending on the yield of syn- thetic asphalt (as yet to be determined in •Mention of trade names or commercial products does not constitute endorsement or recommen- dation for use. a continuous process for which these estimates were made) and depending on its value as an alternative to petro- leum asphalt. Either or both of these factors could significantly change the economics of synthetic asphalt recovery from sludge in future years. In Table 4, credit for sludge disposal was estimated at $66 per dry tonne ($60/short ton). Revenue from sale of oil was calculated at S0.25/L ($40/bbl) and for the synthetic asphalt, $143/tonne ($130/short ton). These prices are representative of 1980 prices for equivalent petroleum products. At current petroleum prices and with the capital and operating costs shown, it is necessary to take some credit for sludge disposal to make the process economically viable. As envisioned, the liquefaction plant would be adjacent to a wastewater treatment plant and would take all the primary sludge generated by the wastewater plant—so, in fact, some credit is due since the sludge would otherwise have to be disposed of. Because the cost of disposal will certainly rise in the future, as will petro- leum prices, the liquefaction process should become viable in future years even should it not be thought to be viable at current projected payback periods of 6 to 12 years. Recommendations The potential for direct liquefaction of sludge to produce liquid fuel at a reasonable cost is very real based on the data presented m this report. Unlike other thermal processes, such as incin- eration with heat recovery or gasifica- tion, liquefaction produces a fuel that is eminently storable and transportable. The quantity of net energy produced by liquefaction (i.e., steam-volatile oil and synthetic asphalt) is likely to be greater than the net energy produced by incineration with heat recovery. Batch reaction methods used for this study are not acceptable for design and scale-up of even a modest pilot plant. Therefore, a bench-scale continuous reaction system is needed to: • determine physical properties of intermediate products; • determine necessary designs for ancillary equipment; • obtain larger samples for more detailed testing; • determine liquefaction product characteristics as a function of temperature, time, alkali, and incoming sludge composition; and • determine the technical feasibility of a continuous reactor for sludge liquefaction. Reliable data for design, scale-up, and economic evaluation would result from a continuous bench-scale unit. In addi- tion, if this unit were located at the site of a wastewater treatment plant, product characteristics as a function of sludge composition could be measured and the treatability of the residual aqueous phase could be determined. Therefore, we recommended that a small lab-scale continuous liquefaction facility be built near a wastewater treat- ment plant. With this system, emphasis should be placed on developing fuel oil from sludge because this option appears to be more cost-effective in the near term. However, since the high molecular weight fraction is made along with the steam-volatile oil, investigating its utility as a petroleum asphalt re- placement or as a coking stock or fuel can continue at minimum additional R&D cost. Conclusions Seventy percent of the combustion \ energy available in sewage sludge (approximately 5000 cal/g or 9000 Btu/lb) can be converted to liquid and solid fuels analogous to petroleum by direct thermochemical liquefaction The solid fuel can be burned to provide all the necessary process heat requirements so that the process is a net energy producer. the high molecular weight product of liquefaction is, in some cases, an acceptable replacement for petroleum asphalt based on the results obtained so far. Synthetic asphalt samples desig- nated HS-1, 7, and 10 were ranked as satisfactory by our preliminary asphalt testing procedures; most others were ranked as unsatisfactory. Since the conversion procedure for the satisfac- tory and unsatisfactory samples was the same in most cases, the only plausible explanation is that inherent differences in the composition of the various samples of sludge used led to physical differences in the synthetic asphalt products. The synthetic asphalt can also be used as a potential fuel or coking stock in the event it is not immediately accep- table as a paving binder. In addition, the steam-volatile oil produced during^ ------- liquefaction has significant value as a synfuel, having 90% of the heating value of No. 2 fuel oil. Based on the current value of petroleum asphalt and heating oils, the liquefaction process would be economically viable in many existing situations, with expected pay- back periods of less than 12 years. If the process objective were changed to production of fuels rather than asphalt, a reduction in processing cost would be likely—the economics would probably be more favorable and, in addi- tion, marketing of the products would be simplified. The full report was submitted in ful- fillment of Grant No. R-806790-01 by Battelle-Northwest, Richland, WA, under the sponsorship of the U.S. Environmental Protection Agency. J. M. Donovan, R. K. Miller, and T. R. Batter are with Battelle-Northwest, Richland, WA 99352; R. P. Lottman is with the University of Idaho, Moscow, ID 83843. Howard Wall is the EPA Project Officer (see below). The complete report, entitled "Physical and Chemical Characteristics of Synthetic Asphalt Produced from Liquefaction of Sewage Sludge. "(Order No. PB 82-119 082; Cost: $9.00, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Municipal Environmental Research Laboratory U.S. Environmental Protection Agency Cincinnati, OH 45268 •frU.S. GOVERNMENT PRINTING OFFICE:1982--559-092/3364 ------- United States Center for Environmental Research Fees pajo. Environmental Protection Information Environmental Agency Cincinnati OH 45268 Protection Agency EPA 335 Official Business Penalty for Private Use $300 RETURN POSTAGE GUARANTEED _. . „, v Third-Class Bulk Rate MERL0063240 LOU W TILLEY REGION V EPA LIBRARIAN 230 S DEARBORN ST CHICAGO IL 60604 ------- |