Evaluation of processes for conversion of biomass-to-fuels using biochemical versus thermo-chemical processes

Transcription

1 KET050 Dept of Chemical Engineering, Lund Institute of Technology Evaluation of processes for conversion of biomass-to-fuels using biochemical versus thermo-chemical processes June 4, 2009 Principal investigators: Josefine Cragnell, Kristina Green, André Holmberg, Magnus Karlsson, Nathalie Mönegård-Jakobsson Tutors: Børre T. Børresen, StatoilHydro Guido Zacchi, Christian Hulteberg, Hans T. Karlsson Department of Chemical Engineering

2 Abstract This report contains an evaluation between two processes producing biofuels from wood. The first process, was based on an earlier study performed at the department of chemical engineering at LTH, produces ethanol via SSF, Simultaneous Saccharification and Fermentation, and the second process goes via an APR, Aqueous Phase Reforming, and produces biodiesel. The second process was designed and simulated with Aspen Plus. The evaluation was performed on a technical/economical basis. The economical calculations of the simulated process were performed using Ulrich s method. The main task was to compare the production of ethanol, with the production of fuel that is produced in the APR-process. The comparison was performed based on the same amount of feedstock, tons of wood per year, which represents m 3 ethanol per year. Two different cases of the APR-processes were evaluated, case one producing diesel and gasoline and case two producing naphtha and energy gases. For case one the quantity produced was m 3 / year and the production cost was 1121 /MWh corresponding to 10.4 /litre. For case two the production cost was estimated at 266 /MWh. For the ethanol process the production cost is 89.3 / MWh which is equal to /litre. A sensitivity analysis was made on the ethanol process and the case one process. This analysis indicated that the APR-process is more sensitive to the investment cost and operating cost than the ethanol process. Due to low selectivity for the different reactions in case one and that the fact that all the hemicellulose is combusted to produce process heat the energy efficiency is 11.7%. For case two the energy efficiency is 41.7% and for the ethanol process it is 60.3%. Case one has more advanced apparatus than the ethanol process, due to its resistance to high pressure, 80 bar. 2

3 Table of Contents Abstract Introduction Background Processes Ethanol process Pre-treatment Simultaneous saccharification and fermentation (SSF) Recovery of ethanol By-products Processes using APR Pre-treatment Enzymatic hydrolysis Alkane production of aqueous-phase processing Isomerisation Dehydration Hydrogenation Aldol condensation Aldol condensation and Hydrogenation Formation of C 12 -alkanes from glucose Formation of C 9 and C 15 alkanes through aldol crossed condensation with acetone Four-phase dehydration/hydration (4PD/H) reactor APR Acetone production Hydrogenolysis Steam reformer for acetone production Solid residue Process proposals Case one Case two Results Ethanol Process Economical evaluation Sensitivity analysis Energy use

4 5.2. Case one Economical evaluation Sensitivity analysis Energy use Fossil part of the biodiesel from acetone Case two Economical evaluation Energy use Comparison between the processes Discussion Virent Comparison Case one and two Conclusions Acknowledgements Works Cited Appendix A Appendix B Calculation methodology for economy Ulrich s method Annuity method Energy calculations Appendix C Thermodynamic data Physical data

5 1. Introduction This project was performed by five master s degree students at the chemical engineering department at the faculty of engineering at Lund University and is performed in cooperation with StatoilHydro ASA, Oslo. The aim of the project was to make a techno-economical comparison and evaluation of two different processes for production of biofuels using wood as raw material. The first process is production of ethanol from wood with the following steps; Pre-treatment, enzymatic hydrolysis, fermentation, separation and drying. The second process is aqueous phase reforming (APR-process) with subsequent synthesis of fuels, including pre-treatment, hydrolysis, APR, fuel processing and separation and drying. StatoilHydro proposed to use a process developed by Virents Energy Systems for this, but due to lacking information regarding the process it was rejected. Therefore two suggestions for the APR-process were made. In the first suggestion acetone is added to increase the carbon number of the product using only one APR-unit, which produces hydrogen. This process proposal produces diesel range liquid fuels. In the second suggested process another APR-unit was introduced. The second APR was fed with hydrogen and sugar alcohols to produce alkanes C 1 -C 6 where the heavier ones can be used as a liquid fuel additive and the lighter ones can be used as e.g. town gas. The two APR-cases have been simulated in Aspen Plus, which is a process modelling tool for conceptual design, optimization and performance monitoring for chemical industries. The processes was then evaluated economically and technically based on the results from those simulations. The purpose of this study was to perform a technical and economical evaluation on the two processes regarding the energy content of the final products. The processes were based on a feedstock of tons of wood/ year. In the first process, m 3 ethanol /year is produced and the energy content of that ethanol (1230 TJ) will be compared with the biofuel produced using the APR-process. The ethanol process is evaluated at the department of chemical engineering at LTH while the APR-process had to be designed and evaluated by the project group. Issues like heat and mass balance, investment and operating cost and sensitivity analysis have been considered. 2. Background Conversion of various biomass sources to fuel may represent attractive business opportunities. Furthermore, an increased use of fuels produced from biomass may be environmentally beneficial with respect to emissions of greenhouse gases. Several processes for production of fuels from biomass have been developed, but for the time being the commercial competitiveness of these processes are questionable. This is in particular the case for processes using lignocellulosic materials, like wood or straw. The following section will supply an insight in the importance of investigating these processes. Fluctuating oil prices and limited supply of petroleum are both reasons for the interest in biofuel production. It is also of strategic importance for many countries to decrease their dependence on imported oil. Another reason for the increasing attention to biofuel production is the amount of work it creates locally. It is expected that for every percent of the European Union's fossil fuel consumption that is exchanged by biofuels, between to new jobs can be created. As a way to aid their economic development some non-oecd (Organization for Economic Co-operation and Development) countries, for example Brazil, have developed a biofuel industry of their own. Their aim is to produce 5

6 fuels for local use as well as for export, which is an approach that many more are considering. Because of the management control measures such as tax incentives by national governments, biofuels accounted for over 1.5% of global transport fuels in (1) Data has been showing significant increase in the atmospheric concentration of greenhouse gases, such as carbon dioxide, nitrous oxide and methane over the last century. This is primarily due to the combustion of fossil fuels. These greenhouse gases have a warming effect on the planet. (2) Temperature data for the last thousand years has been collected from around the globe using tree rings, corals, ice cores and historical records. For the last 140 years, it is estimated that the global average surface temperature has increased by C. It is also likely that the 1990s have been the warmest decade and 1998 the warmest year since This together with a wide array of other data suggests that the planet is getting warmer (2). It is not known to which degree this warming can be attributed to anthropogenic emissions of greenhouse gases, though it is very likely that a major part of it can be attributed to just that. (2) (3) If this theory is correct it is important to reduce the amount of greenhouse gases emitted by human activities. The increasing concern for the greenhouse effect, the depletion of fossil fuels and rising oil prices, builds a need for new substitutes that provide energy. A fuel that can replace the fossil fuels is bioethanol. A comparison has been made for the energy content between gasoline, diesel and ethanol (Table 1). Table 1 Comparison of different fuel types (4) Density (g/l) at 51 C Energy(kJ/l) LHV at 51 C Gasoline No. 2 Diesel Ethanol With a given volume of ethanol it is possible to drive about 75-80% of the distance compared to the same amount of gasoline even though it only contains two thirds of the energy content. (4) This means that the consumer has to fill their fuel tank more frequently but it will be slightly reduced since an engine that runs on ethanol has higher efficiency than an engine that runs on gasoline. (5) The ethanol used for refuelling cars (E85) is expected to give 88% less emission of greenhouse gases compared to gasoline, if the ethanol is derived from cellulose. This assumes that the cellulosic ethanol is over 100% renewable. However E85 contains 15-25% of gasoline which also has to be regarded. Table 2, made by Ford Sweden, shows that driving a car with ethanol give 75% less emission of carbon dioxide versus gasoline. (6) Table 2 Comparison of different motor types (6) Car model Motor carbon dioxide Decrease of emission Ford Focus 1.8 Gasoline 169 g/km hp Ford Focus 1.6 Diesel 127 g/km 21% 109 hp Toyota Prius 1.5 Hybrid - gasoline/ 104 g/km 35% 77 hp Electricity Ford Focus hp Flexi Fuel Bioethanol/ gasoline 42 g/km 75% 6

7 Second generation bioethanol is ethanol produced from biomass such as forest residues and agricultural waste. During the combustion carbon dioxide is emitted, but since new biomass will grow, the same amount as emitted will be assimilated. Bioethanol has many other advantages such as it is liquid at room temperature and can therefore be blended with gasoline. Without any modification of the gasoline engine it is possible in most cases to add a concentration up to 20% ethanol. The world leaders today for producing ethanol are Brazil that uses sugar canes as raw material and United States that uses corn. In Europe, Spain and France are the leading countries that utilize grain-based raw materials. In order to replace gasoline and diesel with ethanol in Europe a better raw material which do not compete with food production has to be used. Materials that fulfil this requirement are woody biomass, herbaceous crops and agricultural residues. Only in Sweden which is covered by 55% forest it would be possible to replace 26% of gasoline or 40% of diesel with biofuels, e.g. ethanol. (5) Since the use of fossil fuel started, carbon dioxide that has not been present in the natural cycle between atmosphere and vegetation in a very long time is released. During the last century, man has pumped large amount of oil and that implicate that we have contributed with carbon to the atmosphere that ancient growth and animals have adsorbed during millions of years. (7) When burning wood all carbon dioxide that the tree has received from the air and the ground is released compared with combustion of fossil fuel. When the carbon dioxide is not part of a natural cycle it has nowhere to go except for being left in the atmosphere. During the 1990s one fourth of the carbon dioxide that was released in the air by combustion of fossil fuels was absorbed by the sea, and the absorption in earth and vegetation was of the same order of magnitude. About 40% was however for the time being, left in the atmosphere. When producing fuel from wood, the carbon dioxide that is released is a part of the natural cycle and will be received by the land and sea. The most important is to plant new trees in the same proportion as it is cut down. (7) 3. Processes The processes that were investigated in this study were one for producing ethanol and two different process alternatives, producing diesel and gasoline and producing biogas and light naphtha, using APR Ethanol process The most common way to produce ethanol is from sugar-, starch- or cellulose-based raw materials. The sugar-process is the simplest, followed by the starch-process and finally the cellulose-process. (8) In the cellulose-based ethanol production the feedstock is much cheaper, but the process is more complex. An advantage compared to sugar- and starch-based raw materials is that the cellulose-based raw materials do not compete with food production. In this report, the cellulose-process with spruce as raw material is studied. The process is based on SO 3 -catalyzed steam pre-treatment followed by simultaneous saccharification and fermentation (SSF). (8) Figure 1 shows the schematic process flowsheet. 7

8 Figure 1 Scheme of ethanol process (8) Pre-treatment The purpose of the pre-treatment step is to increase the availability of the cellulose and hemicellulose fibres, which are surrounded by lignin, to enzymatic attack. There is a use of SO 2 to open up the lignocellullosic matrix and release the hemicelluloses sugars. It is not only important to recover the cellulose and hemicellulose, but also the lignin for further utilization as either chemical feedstock or solid fuel. (8) (9). The wood stock is mechanically chipped to 2-10 mm chips and then impregnated with the gas. Then the feed is preheated to 95 C with low pressure steam before high pressure steam is injected into the reactor with a temperature of 210 C. The biomass is exposed to the 19 bar steam for five minutes (8) (5). Then the pressure is rapidly reduced with flashes in two steps, first to 4 bar and then down to 1 bar (atmospheric pressure).the flash streams contain a large amount of volatile compounds which are formed during the pre-treatment. The streams are condensed in heat exchangers to utilize the secondary steam before being sent to the wastewater treatment. (8) The heat losses in the pre-treatment step are assumed to be 10% higher than for an adiabatic system. The steam consumption for spruce as feedstock is 0.82 kg/kg dry feedstock. The capacity of production in the process investigated is m 3 ethanol/ year. For this 181 kton dry raw material (DM)/ year is needed. Table 3 shows the composition of the raw material and the amount of sugars released in percent of ingoing component in the raw material. (9) Table 3 Composition of raw material (DM) and sugar released in percent of ingoing component in the raw material (9) Component Raw material composition % Dilute acid hydrolysis reaction Sugar released % Glucan water Glucose 18 Xylan water Xylose 42 Galactane water Galactose 57.3 Arabinane water Arabinose 22.4 Mannan water Mannose 58.8 Lignin 27.7 Soluble lignin 11.8 Ashes 2.0 Soluble ashes 6.9 Acetyl groups 1.5 Remaining 4.3 8

9 Simultaneous saccharification and fermentation (SSF) In the SSF-step the enzymatic hydrolysis and fermentation are performed simultaneously in the same vessel. This process step is chosen because of its high productivity and its high ethanol yield, which both leads to a lower ethanol production cost compared to separate hydrolysis and fermentation (SHF). The risk of contamination is also lower in the SSF. A disadvantage is that the optimal temperature for the enzymes actually is somewhere between 40 and 50 C, but because of the presence of yeast the temperature is limited to 37 C. The SSF step operates at 37 C in agitated fermenters. The pre-treated slurry is mixed with the yeast stream and enzymes, nutrients and more water is added (8). Another drawback is that the yeast is very difficult to recycle in the SSF so new yeast has to be produced continuously. (8) The yeast grows in separate fermenters on a part of the hydrolysate from the pretreatment mixed with molasses to get a higher sugar concentration. The sugar from the hydrolysate that is used to produce yeast will lead to a small decrease in the ethanol production. (9) The total cycle time for a batch, including draining and cleaning for the SSF-step is set to 84 hours. (8) Recovery of ethanol The recovery of ethanol is done with distillation and dehydration. Distillation and molecular sieve adsorption are used to produce ethanol with a concentration of 99.8 wt%. The water-insoluble solids (WIS) are separated in a filter press from the stillage stream after distillation. (8) The entire solid residue is dried to 88% DM and used as fuel to heat up all endothermic reactions and other energy demanding parts of the processes. The solid is dried with a superheated steam at 4 bar as the drying medium and the secondary steam can be used in other parts of the process. It is important to preheat the feed to nearly saturation temperature to avoid steam condensation. (8) The dried solid have very low ash content and the heating value is higher than normal wood pellets. In the SSF-based enzymatic process traces of yeast and enzymes may be left in the solids that can affect the fuel characteristics as formation of nitrous oxide (8) By-products In the pre-treatment the sugar degradation products 5-Hydroxymetylfurfural (HMF) and furfural are produced. The acetic acid is realised from the hemicellulose.these substances constitute the largest amount of the by-products and affect the process. Studies have been made that proves that furfural and HMF, up to a concentration of 4.6 respectively 6.0 g/l, only affects the productivity and not the yield of ethanol. Concentrations of acetic acid below 6 g/l favour the ethanol production and increase the yield. But higher concentrations decrease the ethanol yield. (8) 3.2. Processes using APR By using APR and other catalytic processes it is possible to produce kerosene, diesel, and gasoline-range alkanes from biomass derivatives. The APR unit produces C 1 -C 6 alkanes and H 2 for the down-stream catalytic reaction steps that complete the product fractions. By producing gasoline and diesel from biomass, it is possible to use the existing infrastructure and engines without modifying them. This is an advantage compared to other alternative fuels such as biogas and almost pure ethanol and methanol. (10) Pre-treatment The pre-treatment used is uses alkali and steam to break down the hemicelluloses. Nearly 100% of the hemicellulose is removed in the pre-treatment and the celluloses and lignin proceed to the enzymatic hydrolysis. (11) 9

10 Enzymatic hydrolysis To cleave the cellulose chains into sugars, cellulolytic enzymes are used. The process is operating under very mild conditions which result in higher yield, cheaper construction materials and reduction of the formation of toxic by-products. This is to be compared to the acid hydrolysis, which is processed under much harsher conditions. The disadvantage is however that the enzymatic process is slow and pretreatment of the material, in order to open up the structure, is necessary. (8) Compared with the SSF-step that is used in the ethanol process, a more optimal temperature for the enzymes in the hydrolysis step (40 50 C) can be selected. (8). 90% of the cellulose is assumed to be converted into glucose which proceeds with the lignin and water to a filter press, where the solids (mainly lignin) is separated from the liquid stream containing the main part of the glucose. A loss of 5% of glucose is assumed to be left in the solid fraction. After the filter the lignin goes to a boiler where it is combusted to support the process with energy. (11) Alkane production of aqueous-phase processing The following section describes how heavier alkanes are formed from C 6 -sugar. The mechanism is showed in Figure 2. The different steps in the figure will be described closer in the following sections below. Figure 2 Formation of heavier alkanes from C 6 -sugar. (12) Isomerisation Before the dehydration can take place, the glucose must undergo an isomerisation process since glucose has a low yield to HMF in the dehydration step, see dehydration. In the isomerisation step glucose is converted into fructose. 10

11 Asghar Molaei Dehkordi, Iman Safari, and Muhammad M. Karima (13) did a study were the converted glucose to fructose. This study shows that with an inlet concentration of 0.1 mol/ l it is possible to convert glucose to fructose with a fractional conversion of 49.9% in a batch reactor. The temperature should be 60 C and the catalyst loading 20 g/l. The catalyst used was Novo Nordisk Sweetzyme type IT, a commercial IGI enzyme (immobilized glucose isomerase). However it is possible to reach 37% fractional conversion with the same conditions except that the inlet concentration is set to 1.25 mol/l, see Figure 3 for further information (13). The lifespan of the enzyme is expected to be kg fructose/kg enzyme (14). Figure 3 Fractional conversion of glucose to fructose for 0.1 mol/l and 1.25 mol/l. (13) Dehydration Dehydration is a reaction where water is removed from the reactant. It is known that dehydration of fructose to HMF has a higher reaction rate and a higher yield than glucose to HMF. There is no fructose in the used feedstock, and therefore glucose has to undergo an isomerisation, to be converted into fructose before the dehydration, see above (15). The dehydration of fructose to HMF can be done catalytically. A homogeneous acid catalyst, such as mineral acid and organic acid, is effective and has showed high HMF yield with high fructose conversion. But there are major problems concerning separation, recycling and corrosion on the equipment. Solid acid catalysts, such as H-form zeolites and metal phosphates, can be recycled and have high HMF selectivity, but the fructose conversion is very low. Studies have been done on dehydration of fructose with different solvents. Solvents can be divided into four main groups: Water, organic solvents or organic-water mixtures, ionic liquids and biphasic water/organic systems. The best conceivable solvent from an environmental perspective is water. Unfortunately water is ineffective as a solvent for fructose dehydration to HMF since it is non-selective. The main by-products which are formed are levulinic acid and formic acid (16). A study has been done on biphasic water/organic solvents for dehydration of fructose by Roma n- Leshkov and Chheda (17). In the study a common mineral acid, HCl, was used as a catalyst. The solvents used were water, dimethylsulfoxide, DMSO, and poly(1-vinyl-2-pyrrolidinone), PVP, in the aqueous phase and methylisobutylketone, MIBK, and 2-butanol in the organic phase. The selectivity and fructose conversion were high, 85% and 89% respectively. The problem with the system was that the product was 11

12 hard to separate from the high boiling temperature DMSO and from the organic phase which HMF also dissolved in (17). Xinhua Qi with co-workers did a study where fructose dehydrates to HMF with an ion-exchange resin catalyst Dowex 50wx8-100 which is a strong acid cation exchange resin. This catalyst showed to be stable up to 150 C and no catalyst deactivation or decreasing selectivity was observed after five runs. The best result, in terms of HMF yield, was achieved when the solvent composition was 30 wt% water and 70 wt% acetone, organic-aqueous solvent, and when the fructose concentration was 2 wt%. The yield and conversion that were achieved were 73.4 and 95.1% respectively. This run was performed at 150 C and during 15 min. The fructose-to-catalyst weight ratio was 1. The by-products produced, in this run, levulinic acid and formic acid had only a yield of 5.7% and 2% respectively. (16) A microwave-assisted organic synthesis, MAOS, has been used for many reactions as a heating source since it have showed to give higher yield and selectivity for the wanted product under a given time than a conventional oil bath heater in the laboratory scale. MAOS was also believed to give a higher yield and selectivity for the specific reaction, which is considered, than conventional oil bath heater. This was confirmed in this study where the two heaters were compared. (16) The resin used is known to be potentially dangerous at high temperatures with decomposition products such as aromatics, hydrocarbons, organic sulfonates and sulphur oxides. The resin was tested between C under microwave irradiation and was shown to be stable up to 150 C, as stated above. (16) This study which is done by Xinhua Qi with co-workers showed that the catalyst Dowex is stable up to 150 C and give a high yield and selectivity for HMF. MAOS was also stated to give a higher yield and selectivity than conventional oil bath heater in laboratory scale. This result is based on a laboratory scale study and the result is assumed to be useful in a full scale industry. (16) Hydrogenation In the hydrogenation step most of the glucose is being converted into sorbitol with mannitol as byproduct, either with H 2 produced in the APR or from an external source. (18) With a feed that contains 40 wt% glucose over g ruthenium catalyst (1.8% Ru/ C) in a Trickle-Bed Reactor (Figure 4) it is possible to reach 98.5% selectivity for sorbitol at 100% conversion of glucose after 39 h. The remaining 1.5% that does not become sorbitol is actually sorbitol isomerised to mannitol. The residence time for the reactant is g ru h/ ml, g ruthenium and hour/ml feed, if this is increased the selectivity to sorbitol will decrease and if the residence time is decreased to below g ru h/ ml the glucose conversion will decrease drastically. The glucose feed of 40 wt%, equals 2.6 mol glucose/ l, and the glucose solution flow rate is 36 ml/h. This solution is fed into the reactor with a hydrogen stream of 20 l/h at 100 C and 80 bar. If the hydrogen stream is 20 l/ h the catalyst will be deactivated after 693 h on stream, therefore the hydrogen flow should be increased to 100 l/h. (18) It is possible to use both nickel and ruthenium catalyst but nickel demands higher temperature and pressure which results in higher plant cost and the risk of caramelization of the feed. Nickel also tends to leak into the sugar solution which demands an implementation of an expensive purification step. (19) 12

13 Figure 4 A general picture of a trickle-bed reactor (20) Aldol condensation Aldol condensation is one of the main reactions for forming C-C bonds from hydrocarbons containing carbonyl groups (21) (22). In aldol condensation, larger compounds are formed under mild temperatures, generally between 300 K and 370 K (15). Aldol crossed-condensation takes part between different aldehydes and ketones while aldol self-condensation occurs between two equal aldehydes or two equal ketones. The aldol condensation can either be catalyzed by an acid or a base. When a base catalyst is used in aldol self-condensation the following mechanism occurs. First the base attaches the α-hydrogen to form water and then an enolat ion is formed. The enolat ion attacks the unsaturated carbon site related to the carbonyl group in the other aldehyde or ketone. The product then retracts the hydrogen from the base (23), see Figure 5. An aldol crossed-condensation occurs in the same way. Dehydration can occur after aldol condensation. The result will be an unsaturated aldehyde or ketone. Factors like reaction temperature, solvent, reactant molar ratio, structure of reactant and choice of catalyst determines how large the produced compound will become. HO O H - OH O - O OH O O OH O O O HO O O OH HO O O OH + - OH O- OH HOH Figure 5 HMTHF, aldehyde, undergoes aldol self-condensation. Generally industrial aldol condensation is performed with a homogenous base catalyst like sodium or calcium hydroxide. This performance will produce a large wastewater stream that has to be neutralized, which increase the process cost. To decrease the process cost and to achieve a cleaner process a stable solid base catalyst is desired (15) Aldol condensation and Hydrogenation MgO-ZrO 2 catalyst has proved to be stable in aqueous-phase aldol condensation condition. The product precipitates from the aqueous solution since the product becomes less water-soluble. This problem can be solved by having hydrogenation present in the same reactor. The catalyst then has to be modified, which is done by depositing Pd on it to form a bifunctional metal base Pd/MgO-ZrO 2. 13

14 Aldol crossed-condensation and hydrogenation have been performed in the same batch in the presence of Pd/MgO-ZrO 2. Thirteen experiments were performed with furfural and acetone and another four were performed with HMF and acetone. The molar ratio, the aldol crossed-condensation temperature and the organic/catalyst ratio were varied in the furfural and acetone experiments while only the aldol crossedcondensation temperature was varied in the HMF and acetone experiments. In the HMF and acetone experiments the molar ratio was constant, 1:1.The hydrogenation temperature was held constant at 393 K in all the experiments. The best result, considering selectivity for higher products and overall carbon yield in the aqueous phase, was achieved for the furfural and acetone experiments when the aldol crossedcondensation was performed at 353 K a molar ratio of 1:1 and the organic/catalyst ratio was 6. This experiment differs from the other in respect to solvent during the hydrogenation. The solvent used was hexadecane while it was water in the other experiments. (24) The selectivity for higher products and overall carbon yield in the aqueous phase were calculated based on C 5 units in the furfural experiments and C 6 units in the HMF experiments. (24) For the HMF and acetone experiments the best result, considering selectivity for higher products and overall carbon yield in the aqueous phase, was achieved when the temperature was 326 K. The highest overall carbon yield and the highest selectivity for the given condition above for HMF and acetone was 94% and 61% and for furfural and acetone the highest overall carbon yield and the highest selectivity was 71% and 85%, see 14

15 Appendix A, table 6. (24) The aldol condensation was stopped after hours. The hydrogenation reaction went on until the overall carbon yield in the aqueous phase had reached a constant value. In order to ensure this the hydrogenation reaction was stopped after 26 hours, for all the runs. (24) The catalyst Pd/MgO-ZrO 2 was tested for recycling. Three runs were performed without regenerating the catalyst in between and the result showed that the catalyst retained most of its activity and selectivity. The catalyst can be totally regenerated through calcination (24) Formation of C 12-alkanes from glucose Hydrocarbons up to six carbons can be produced from glucose and xylose in APR. Hydrocarbons with six carbons have a low value as fuel additive because of its high volatility. For production of higher hydrocarbon from glucose and xylose another pathway has to be used. Glucose and xylose contain a carbonyl group each but they do not undergo aldol self-condensation since the carbonyl group undergoes intermolecular reactions to form ring structures. It is possible to dehydrate glucose and xylose over a mineral or acid solid catalyst. Glucose then forms HMF, and xylose forms furfural (21). If HMF and furfural can be selectively hydrogenated in the furan ring double bonds (15), then HMF and furfural can be converted to 5-hydroxymethyl-tetrahydrofurfural, HMTHF, and tetrahydrofuran-2 carboxyaldehyde, THF2A, respectively. HMTHF and THF2A can undergo aldol selfcondensation over a base catalyst. The base catalyst can either be a mineral or a solid. If a solid is used no formation of water will take place. The large organic compounds produced need to be hydrogenated over a metal catalyst, e.g. Pd, to increase its solubility in water. The organic compounds can then be converted to alkanes by dehydration/hydrogenation in a four-phase dehydration/hydration (4PD/H) reactor (21). It has been showed that the selectivity for hydrogenation of merely the furan ring in furfural is poor, less than 6%. This is believed to depend on the steric hiderance that can arise when a large molecule, such as furfural, should react. (25) Formation of C 9 and C 15 alkanes through aldol crossed condensation with acetone There is another pathway for HMF and furfural than through hydrogenation to form higher alkanes. Since neither HMF nor furfural has an α-hydrogen atom they cannot undergo aldol self-condensation but they can react with another aldehyde or ketone that does (21). Acetone can react with HMF and furfural through aldol crossed-condensation with a base catalyst. The product will be a large organic compound, C 8 or C 9, depending on if HMF or furfural is used. Before the final dehydration/hydrogenation, where the straight chain alkanes are produced, the large organic compounds have to become more water-soluble. This is done by a hydrogenation step. It is possible to produce higher alkanes ( > C 9 ) with acetone by doing a subsequent aldol crossedcondensation with HMF or furfural before the hydrogenation. After that a dehydration/hydrogenation step in a (4PD/H) reactor that produces up to C 15 alkanes is performed (15). 15

16 Four-phase dehydration/hydration (4PD/H) reactor The large water-soluble organic compounds can be converted into liquid alkanes (C 7 -C 15 ) over a bifunctional catalyst (Pt/SiO 2 -Al 2 O 3 ) in a (4PD/ H) hydrogenation flow reactor (Figure 6). The reactor has one water inlet stream containing the large organic compound, one hexadecane inlet stream, one hydrogen inlet stream and a solid catalyst. The hexadecane stream serves as a remover of the produced alkanes from the catalyst surface in order to prevent forming of coke (21). Figure 6 Four phase reactor APR Hydrogen production from reforming of alkanes (C 1 -C 6 ) is dictated by the thermodynamics for the steam reforming of the alkanes to form H 2 and CO and the shift reaction that forms CO 2 and H 2 from CO and water. The steam reforming is only favourable at temperatures above 675 K, and for methane only above 900 K (26). APR of biomass derived chemicals, such as sorbitol, glycerol and glucose, can be used to produce hydrogen and light alkanes (27). Reforming of oxygenates such as ethylene glycol, glycerol and sorbitol is favourable at much lower temperatures than those required for similar carbon number alkanes (26). For example the equilibrium constant at 500 K is very favourable for the production of H 2 and CO 2 from sorbitol in the presence of water. However, at this low temperature the production of hydrogen is complicated by side reactions like methanation and Fischer-Tropsch reactions, which consume CO 2 and H 2 to produce alkanes and H 2 O (28). The standard Gibbs free energy for the vapour phase reforming of a few alkanes and oxygenated hydrocarbons are shown together with the water-gas shift reaction in Figure 7. Hydrogen production from APR of oxygenated hydrocarbons, with a C: O ratio of 1:1 has the net reaction according to reaction 1. C n H 2y O n n CO + y H 2 (26) (1) The methanation reaction (2) comprises CO H 2 CH H 2 O (28) (2) Fischer-Tropsch reaction 3 2n + 1 H 2 + n CO C n H 2n+2 + n H 2 O (29) (3) 16

17 Figure 7 ΔG /RT vs. temperature for production of CO and H 2 from vapour phase reforming of alkanes; oxygenated hydrocarbons with a C:O ratio of 1:1; and water-gas shift. Dotted lines show values of ln(p) for the vapour pressures vs. temperature of CH 3 (OH), C 2 H 4 (OH) 2, C 3 H 5 (OH) 3, and C 6 H 8 (OH) 6 (pressure in units of atm) (26). Aqueous-phase reforming is a bifunctional reaction pathway that makes use of a metal catalyst, such as Pt. This catalyzes the formation of hydrogen and carbon dioxide by C-C bond cleavage and a following water-shift reaction, and an acid catalyst, either solid acid or mineral acid, which catalyzes the dehydration of the reactant. The metal catalyst also catalyses a hydrogenation of the dehydrated reaction intermediates. This leads to the overall conversion of for example sorbitol to alkanes, hydrogen, CO 2 and water. If the reaction conditions are chosen appropriately, a product consisting mainly of alkanes such as butane, pentane and hexane will be derived from the APR of sorbitol. Hydrogen does not need to be produced in-situ if it is co-fed into the reactor instead (27). If hydrogen is co-fed into the reactor a higher selectivity for the heavier of the C 1 -C 6 alkanes is achieved since the rate of hydrogenation increases with higher H 2 partial pressure and the rate of C-C bond cleavage decreases (27). The conversion of sorbitol to alkanes, CO 2 and water is exothermic, although only 5% of the heating value is lost in the reaction. After this conversion the dehydrated product contains only 30% of the mass of the sorbitol reactant. (27) A good catalyst for the production of hydrogen through APR must cleave C-C bonds and promote removal of adsorbed CO by water gas shift but should not promote hydrogenation of CO or CO 2 and should not facilitate C-O bond cleavage. Glycerol and ethylene glycol decomposes at room temperature over Pt catalyst to form adsorbed CO, however Pt is expensive. A high selectivity for hydrogen through APR of ethylene glycol, glycerol and sorbitol can be achieved with a tin promoted Raney-Nickel catalyst. Its activity, selectivity, and stability for H 2 production are comparable to that of Pt/Al 2 O 3. Along with the hydrogen and light alkanes, small amounts of CO, corresponding to the equilibrium, as well as small amounts of alcohols, organic acids and aldehydes are produced. (30) APR is also the key step in Virents bio forming process. The oxygenated hydrocarbons in the aqueous feed reacts with water over a proprietary heterogeneous catalyst yielding hydrogen, carbon dioxide, alcohols, ketones, aldehydes, and by-product alkanes, organic acids, and furans. The APR in the 17

18 BioForming process operates at K and at bar. At these conditions, it is believed that the reactions that take place include: reforming to generate hydrogen, dehydrogenation of alcohols/ hydrogenation of carbonyls, deoxygenation reactions, hydrogenolysis, and cyclization. The hydrogen is used to defunctionalize the highly reactive carbohydrates to less reactive mono-oxygenates like alcohols, ketones, and aldehydes. (10) This means that besides C 1 through C 6 alkanes, the APR can produce a stream containing alcohols, ketones, acids, and cyclic components suitable for condensation. (10) 3.3. Acetone production In the dehydration step acetone is needed. This can either be produced, from sugars via fermentation or by heterogeneous catalysis or be purchased Hydrogenolysis To form short-chained oxygenated compounds such as glycerol, propylene glycol and ethylene glycol the water soluble carbohydrates can go through a hydrotreating step, hydrogenolysis, which uses hydrogen generated in the APR or recycled from the system. (10) Steam reformer for acetone production Steam reforming a 4:1 weight ratio of glycerol: water feed with a PtRe (1:1)/C catalyst at elevated pressures yields a product with increased amounts of oxygenated hydrocarbons in the gas phase while the production of alkanes and CO x is decreased. This indicates a shift in selectivity from C-C bond breaking to C-O bond breaking at elevated pressures. This means the production of oxygenated species becomes favoured over the formation of alkanes. (31) The reformer will produce alcohols, acetone, hydrogen and CO x from a feed of glycerol and water. The best carbon selectivity for acetone production is around 10% and is obtained at 6.5 bar and 503 K with a 5.1 wt% Pt- 4.9 wt% Re/C. The acetone is then concentrated through a distillation step, see appendix A, table 1. Because of the low carbon selectivity towards production of acetone the conclusion was made that it might be better to buy the acetone directly or use the ABE-fermentation (Acetone, Butanol and Ethanol) process instead Solid residue The treatment of the solid residues is the same for the APR-process as for the ethanol-process as for the APR-process. 4. Process proposals Because of lacking information about Virents process which was proposed in the problem definition, the project group integrated two different process proposals. These two cases where studied from a technical and economical point of view and are presented and discussed below. 18

19 4.1. Case one Lignin Combustion CO CO 2 Hydrocarbons H 2S O4 SO 2 Wood Steam P R E T R E A T M E N T E N Z Y M A T I C H Y D R O L Y S I S F I L T E R Glucose Water Hydrogenation 1 Sorbitol Glucose Water Isomerization Acetone Fructos Glucose Water APR1 Dehydration CO H 2 CO 2 Hydrocarbons water Acetone HMF Fructose Glucose Water Levulinacid Acetone PSA Destillation H 2 HMF Fructose Glucose Water Levulinacid Acetone Aldol-crossed condensation Decantion 2 Fructose Glucose Water Levulina cid H 2 H 2 Hydrogenation 2 4-PhaseD/H Alkanes C 8-C 9 Alkanes C 13-C 15 Decantation 3 water Alkanes C 8 Alkanes C 13 water Figure 8 shows the schematic flow chart for case one. Lignin Combustion CO CO 2 Hydrocarbons H 2S O4 SO 2 Wood Steam P R E T R E A T M E N T E N Z Y M A T I C H Y D R O L Y S I S F I L T E R Glucose Water Hydrogenation 1 Sorbitol Glucose Water Isomerization Acetone Fructos Glucose Water APR1 Dehydration CO H 2 CO 2 Hydrocarbons water Acetone HMF Fructose Glucose Water Levulinacid Acetone PSA Destillation H 2 HMF Fructose Glucose Water Levulinacid Acetone Aldol-crossed condensation Decantion 2 Fructose Glucose Water Levulina cid H 2 H 2 Hydrogenation 2 4-PhaseD/H Alkanes C 8-C 9 Alkanes C 13-C 15 Decantation 3 water Alkanes C 8-C 9 Alkanes C 13-C 15 water Figure 8 Schematic figure of case one Pre-treatment Pre-treatment of the wood is done with NaOH and steam to release 100% of the hemicelluloses. Enzymatic hydrolysis Enzymatic hydrolysis is done in order to convert 90% of cellulose into glucose. 19

20 Filter and combustion The lignin is separated from the sugar solution. Approximately 5% of the sugar is estimated to be lost in this separation. The separated lignin is combusted in order to produce process heat. Hydrogen production Since the process uses hydrogen in several steps such as in the hydrogenation and the 4PD/ H, it is necessary to produce large amounts of it. This is described in more detail below. Hydrogenation 1 About 37% of the glucose/water stream is fed to the hydrogenation in order to obtain sorbitol. The rest of the stream is fed to the isomerization. APR The sorbitol is fed together with water into the APR which produces hydrogen that can be used in other parts of the process. Raney-NiSn catalyst is used for this process in order to obtain large amounts of hydrogen as well as to reduce cost. The reaction conditions chosen should be 538 K at 1 wt% sorbitol feed. For detailed information see Appendix A, table 2. Pressure Swing Adsorption, PSA The H 2 product from the APR is purified by PSA and fed to other process units that require hydrogen for operation. The waste CO x can be burned in a combustion chamber or fed to a water-gas shift in order to obtain more H 2 from the CO. The alkanes produced as a by product in the APR can be steam reformed, burned or sold. Alkane production Isomerisation Glucose is isomerised to fructose in order to obtain better yield and selectivity to HMF in the dehydration step. It is possible to convert glucose with a fractional conversion of 49.9%. The process step is operated at 60 C. Dehydration In this step fructose is dehydrated into HMF, a relationship of 30 wt% water and 70 wt% acetone is used as solvent. The process step is assumed to be performed at 150 C and during 15 min. The by-products produced are levulinic acid and formic acid and have a molar yield of 5.7% and 2% respectively. Acetone is purchased since it is more profitable than producing it. See appendix A, table 4 and 5. Distillation Since the aldol condensation requires a 1:1 molar ratio of HMF and acetone, a large amount of the acetone has to be distilled and recycled to the dehydration step. Aldol condensation Acetone is added to HMF and an aldol crossed-condensation takes place, which produces larger molecules with low solubility in water. This is done in a batch reactor for 24 hours. After that the water is removed and an organic phase is added in order to dissolve the monomers and dimers produced. See appendix A, table 6 20

21 Hydrogenation The monomers and dimers are hydrogenated so that their water solubility is increased. Water is added as a solvent. Repeated Dehydration/Hydrogenation (4 phase reactor) By repeated dehydration and hydrogenation of the monomers and dimers all the oxygen can be removed and replaced with hydrogen. A limited amount of C-C bond breakage inevitably occurs. The result is alkanes of 1-15 carbon atoms, of which most can be used as diesel fuel. 21

22 4.2. Case two Figure 9 shows the schematic flow chart for case two. H2SO4 SO2 Wood Steam P R E T R E A T M E N T E N Z Y M A T I C H Y D R O L Y S I S F I L T E R Lignin Glucose Mannose Xylose Combustion Hydrogenation Sorbitol Mannitol Xylose APR1 APR2 Alkanes C1-C6 Water CO2 H2 CO CO2 PSA CO CO2 PSA Alkanes C1-C6 Water Water(L) Figure 9 Schematic figure of case two Pre-treatment Pre-treatment of the wood is done in the same way as for the ethanol process with a dilute acid and steam. Enzymatic hydrolysis Enzymatic hydrolysis is done in order to obtain monosaccharides from cellulose. Filter and combustion The lignin is separated from the sugar solution. Approximately 5% of the sugar is estimated to be lost in this separation. The separated lignin is combusted in order to produce process heat. Hydrogenation of Sugar The monosaccharides are hydrogenated in order to obtain sugar alcohols such as sorbitol and mannitol. Hydrogen production Since the process uses hydrogen in the hydrogenation and the APR2, it is necessary to produce large amounts of hydrogen. APR 1 The sugar alcohols are fed together with water into the APR which produces hydrogen that can be used in other parts of the process. Raney-NiSn catalyst is used for this process in order to obtain large amounts of hydrogen as well as to reduce cost. The reaction conditions chosen are assumed to be 538 K at 1 wt% sorbitol feed. For detailed information see appendix A, table 2. 22

23 Pressure Swing Adsorption, PSA The H 2 product from the APR is purified by PSA and fed to other process units that require hydrogen for operation. The waste CO x can be sent to a combustion chamber or fed to a water gas shift in order to obtain process heat or more H 2 from the CO. The alkanes produced in the APR1 are products that can be sold. Alkane production APR 2 If an APR is fed with hydrogen as well as sugar alcohols the product tends to consist of heavier alkanes, though limited to the carbon number of the sugar alcohols, since the hydrogen does not need to be produced in-situ by reforming action. This produces, with high selectivity, alkanes C 1 -C 6 of which the heavier can be used, after isomerisation to more branched chains, as gasoline additive at great ratios. The lighter and middle fractions can perhaps be used as town gas or LPG. This process produces lighter alkanes that might not be worth much, but since this process is very simple and will produce a great deal of those alkanes it might still be an interesting pathway towards a renewable fuel from cellulose. Pressure Swing Adsorption, PSA Same as in the section above. 23

24 5. Results The following section presents the results from the economical and technical evaluations. Case two will not be as thoroughly evaluated because the main prospect is to compare the ethanol process with case one Ethanol Process The following section describes the results from the ethanol process. The economy, a sensitivity analysis and the energy demand is considered Economical evaluation The economy is calculated from an ethanol production with a feed stock of ton/year, which is converted into m 3 ethanol/year. This is evaluated in reference (9) were the economical data used are from year Investment cost The investment cost for the different steps are declared in Table 4 for producing m 3 ethanol/year. Table 4 Investment cost for the ethanol process Step Investment cost (MEUR) Share % Handling of raw material Hydrolysis/ Pre-treatment Fermentation/ SSF Distillation Evaporation Drying, pelleting etc Steam generation Water purification Store Total investment cost 127 The investment cost is very high and the largest part of the capital cost is the equipment for the hydrolysis/ pre-treatment and the fermentation/ssf. Table 5 shows the cost in MEUR for producing m 3 ethanol per year. Table 5 Total investment cost for the ethanol process Raw material demand Ethanol production Solid fuel production Investment cost [m 3 / year] [TWh/ year] [MEUR] 0.9 TWh (181 kton TS) (33.75MW) 112 Operating cost The remaining costs consist of cost for raw material, solid fuel, electricity, water, chemicals, salaries and insurances. The flow rates and costs are summarized in Table 6. 24