When the tars and BTX are removed, the other impurities can be removed by standard wet gas cleaning technologies or advanced dry gas cleaning technologies.

Wet low-temperature synthesis gas cleaning is the preferred method for the short term (van Ree et al. 1995). This method will have some energy penalty and requires additional waste water treatment, but in the short term it is more certain to be effective than hot dry gas cleaning.

A cyclone separator removes most of the solid impurities, down to sizes of approximately 5 pm (Katofsky 1993). New generation bag filters made from glass and synthetic fibers have an upper temperature limit of 260°C (Perry et al. 1987). At this temperature particulates and alkali, which condense on particulates, can successfully be removed (Alderliesten 1990; Consonni et al. 1994; Tijmensen 2000; van Ree et al. 1995). Before entering the bag filter, the synthesis gas is cooled to just above the water dew point.

After the filter unit, the synthesis gas is scrubbed down to 40°C below the water dew point, by means of water. Residual particulates, vapor phase chemical species (unreacted tars, organic gas condensates, trace elements), reduced halogen

Removes NH3, HCI, metal, part HCN, HS

/ / Removes Removes Removes Removes traces NH3 HCI, HCN H2S, and COS HCN, H2S, NH3, COS FIGURE 2.4 Three possible gas cleaning trains. Top: tar cracking and conventional wet gas cleaning; middle: tar scrubbing and conventional wet gas cleaning; and bottom: tar cracking and dry gas cleaning.

gases and reduced nitrogen compounds are removed to a large extent. The scrub­ber can consist of a caustic part where the bulk of H2S is removed using a NaOH solution (van Ree et al. 1995) and an acid part for ammonia/cyanide removal. Alkali removal in a scrubber is essentially complete (Consonni et al. 1994).

With less than 30 ppm H2S in the biomass derived synthesis gas, a ZnO bed may be sufficient to lower the sulfur concentration below 0.1 ppm. ZnO beds can be operated between 50 and 400°C, the high-end temperature favors efficient utilization. At low temperatures and pressures, less sulfur is absorbed; therefore, multiple beds will be used in series. The ZnO bed serves one year and is not regenerated (Katofsky 1993; van Dijk et al. 1995). Bulk removal of sulfur is thus not required, but if CO2 removal is demanded as well (see page 23), a solvent absorption process like Rectisol or Sulfinol could be placed downstream, which also removes sulfur. H2S and COS are reduced to less than 0.1 ppm and all or part of the CO2 is separated (Hydrocarbon Processing 1998).

Starch is washed to remove residual protein and is converted to a glucose syrup. First, the starch is jet-cooked and held for liquefaction at 90°C with alpha-amylase. In contrast to the dry-grind process, all of the alpha-amylase is added prior to jet cooking. Then glucoamylase and pullulanase, a a-1^6 debranching enzyme, are added to convert dextrin polymers to sugars. Addition of pullulanase ensures good conversion of dextrins to glucose by decreasing formation of isomaltose, a glucose dimer. Isomaltose is present at starch polymer branching points and can also be formed by “reversion” of glucose to isomaltose in a reaction catalyzed by glucoamylase. After saccharification, glucose is fermented to ethanol by an industrial strain of S. cerevisiae. In wet mills, the fermentation is often run using a series of fermenters in a semi­continuous process. Approximately 2.5 gallons of ethanol are produced from a bushel of corn in the wet-milling process.

Pervaporation is a technique that allows selective removal of volatiles from model solution/fermentation broths using a membrane. The volatile or organic compo­nent diffuses through the membrane as a vapor followed by recovery by conden­sation. In this process, a phase change occurs from liquid to vapor. Since it is a selective removal process, the desired component requires a heat of vaporization at the feed temperature. The mechanism by which a volatile/organic component is removed by pervaporation is called solution-diffusion. In pervaporation, the effectiveness of separation of a volatile is measured by two parameters called selectivity (a measure of selective removal of volatile) and flux (the rate at which an organic/volatile passes through the membrane per m2 membrane area). A schematic diagram of the pervaporation process is shown in Figure 6.4. The details of pervaporation have been described in the literature (Maddox, 1989; Groot et al., 1992; Qureshi and Blaschek, 1999b).

Application of pervaporation to batch butanol fermentation has been described by Groot et al. (1984), Larrayoz and Puigjaner (1987), Qureshi and Blaschek (1999a), and Fadeev et al. (2001). Pervaporation has also been used for the removal of butanol from the fermentation broth in fed-batch reactors (Qureshi and Blaschek, 2000b; Qureshi and Blaschek, 2001a). In the fed-batch reactors concentrated sugar solutions have been used to reduce the process stream volume. It is interesting to note that acids did not diffuse through the membranes used by the above authors. Qureshi et al. (1992) used a polypropylene membrane through which diffusion of acids occurred, however, at high acid concentration in the fermentation broth.

In an attempt to improve membrane selectivity, Matsumura et al. (1988) applied a combination of liquid-liquid extraction and pervaporation to recover butanol. The extraction solvent used for this process was oleyl alcohol, which formed a liquid layer (also known as a thin membrane) on a microporous 25 pm thick polypropylene flat sheet. The oleyl alcohol also got impregnated into the sheet pores. In this combination of liquid-solid membrane, oleyl alcohol dissolved butanol relatively quickly followed by diffusion through the polypropylene mem­brane. The advantage of combining the liquid and solid membrane was that a high butanol selectivity (180) was achieved in comparison to a low selectivity (10-15) when using polypropylene film alone. It was estimated that if this solid — liquid pervaporation membrane were used for butanol separation, the energy requirement would be only 10% of that required in a conventional distillation. Unfortunately, the membrane was not stable as the oleyl alcohol that formed a thin film and was impregnated into the polypropylene film pores diffused out of the membrane.

In order to develop a stable and highly selective membrane, Qureshi et al. (1999) employed two techniques known as adsorption and pervaporation. It has been reported that adsorption of butanol onto silicalite and molecular sieves is a quick and selective process (Ennis et al., 1987). Qureshi et al. (1999) synthesized a membrane in which silicalite, an adsorbent, was included into a silicone mem­brane. By combining these, butanol selectivity was improved from 40 (silicone membrane) to 209. The membrane developed was called a silicalite-silicone membrane. The membrane was found to be stable with a working life of three years. This membrane was used with both butanol model solutions and fermen­tation broths (Qureshi et al., 1999; 2001). A comparison of various membranes suggested that this membrane may be superior to other pervaporation membranes used for butanol separation (Qureshi and Blaschek, 1999b). Some of the details
of the use of this membrane have been given in the section fed-batch fermentation (Qureshi et al., 2001) where 155 gL-1 ABE was produced in the integrated fermentation and product recovery process as compared to <20 gL-1 in a batch process. This membrane was so efficient that a butanol concentration up to 700 gL-1 was achieved in the permeate.

New materials are being developed for use as separators/ionic conductors in DMFCs. These materials generally fall into one of two categories, fluorinated and nonfluorinated. Here some current developments in both areas are presented.

In an effort to develop an inexpensive and effective separator for DMFCs and fuel cells operated on other fuels, Melman et al. developed a nanoporous proton­conducting membrane (NP-PCM) [63]. The NP-PCM is made of polyvinylidene fluoride (PVDF) and SiO2. Fuels are mixed into 3 M sulfuric acid electrolyte and circulated past the anode. Characteristics of the NP-PCM separator that bests Nafion include: membrane cost lowered by 2 orders of magnitude; pore sizes roughly 50% that of Nafion; methanol crossover cut in half; as much as 4x greater ionic conductivity; and a membrane insensitive to heavy metal corrosion products that allow for less expensive hardware and catalysts. The maximum power achieved with the cell was 85 mW cm-2 at ~243 mV on oxygen at atmospheric pressure, and cell temperature of 80°C. One disadvantage of this type of mem­brane is the peripheral systems must be corrosion resistant.

Nafion is made by copolymerizing tetrafluoroethylene (TFE) and perfluorovi — nyl ether (PSEPVE) containing a sulfonyl fluoride. Whereas Nafion provides good performance in PEFCs at temperatures below 100°C, the material is expen­sive due to the relatively difficult polymerization and expense of the monomer PSEPVE. Also, chemical modifications to tailor Nafion properties has proven difficult. Yang and Rajendran describe an effective copolymerization strategy of TFE and ethylene to produce less expensive melt-processable terpolymers that are easily hydrolyzed and acidified to give polymers of high ionic conductivity [42]. The conductivity is similar to or slightly higher than that of Nafion. The membrane also uptakes more water (by weight) than Nafion under the same conditions. An unoptimized 5-mil thick membrane tested in a DMFC yielded performance comparable to a DMFC made with Nafion 115. The power output of the DMFCs were quite similar, though the methanol crossover through the experimental membrane is 9% higher than for Nafion 115 (i. e., 10.9 x 10-4 vs. 9.9 x 10-4 g min-1 cm-2). The authors speculate that the PEM may be optimized by making a more homogeneous thickness and cross-linking.

A review of nonfluorinated PEMs for use in DMFCs was prepared by Roziere and Jones [38]. The number of nonfluorinated polymer materials for application in higher-temperature fuel cells (i. e., > 80°C) is limited by thermal instability. Thermally stable polymers tend to have either polyaromatic or polyheterocyclic repeat units. Examples of these include polybenzimidazole (PBI), poly(ether ketone)s (PEK), poly(phenyl quinoxaline (PPQ), polysulfone (PSU), and poly(ether sulfone) (PES). The chemical structures of some common nonfluori — nated polymers are shown in Figure 9.5. These polymers are thermally stable but are poor ionic conductors until modified. Modification methods include acid and base doping of the polymer, sulfonation of the polymer backbone, grafting phos- phonated or sulfonated functional groups onto the polymer (where sulfonated polymers generally contain “s” in the acronym), graft polymerization onto the polymer followed by functionalization of the graft material, and total synthesis. Many of these membranes exhibit good proton conductivity (0.01 to 0.1 S cm-1) when well-hydrated but a balance between the sulfonation level imparting ionic conductivity, adequate membrane strength, and membrane swelling must be met before long-term use in fuel cells is possible. Roziere and Jones suggest that sPEKs and related polymer blends hold the most promise for use in DMFCs. They cite unpublished performance data of a DMFC made using polyaromatic membrane with comparable power density to Nafion cells and little degradation after start/stop regimes at temperatures >100°C over a period of weeks.

There is a limited amount of DMFC performance data for fuel cells made with these PEMs, and the conditions under which the PEMs are tested in DMFCs are quite varied. Relevant DMFC performance data included here are listed in the “Performance Targets and Efficiencies” section of this chapter. A more rig­orous set of evaluations for candidate PEMs in operational DMFCs should include durability tests including startup and shutdowns over extended periods.

An engaging two-paper study by Silva et al. evaluates inorganic-organic hybrid membranes made from sulfonated poly(ether ether ketone) (sPEEK)

PEK

[64,65]. One goal of the study is to demonstrate a systematic and complete approach to DMFC membrane development, characterization, and, ultimately, the real-world testing in DMFCs that is thus far lacking for other promising materials. The sPEEK membranes have a sulfonation degree of 87% and zirconium oxide content that varies between 2.5 and 12.5 wt%. The group evaluates the membranes using standard analytical techniques (e. g., impedance spectroscopy (proton con­ductivity and proton transport resistance), pervaporation (permeability to meth­anol), and water swelling) and as PEMs in DMFCs. The organic/inorganic hybrid sPEEK/ZrO2 membranes exhibit good proton conductivity and the addition of ZrO2 particles can tailor the electrochemical performance of the membranes. This makes the organic/inorganic hybrid sPEEK/ZrO2 membranes a possible alterna­tive to perfluorinated membranes. It was found the proton transport resistance increased as the wt% ZrO2 in the membrane increased and that proton conduc­tivity followed the opposite trend. Water uptake decreases as the inorganic com­ponent increases, following the same trend as the proton conductivity. This supports the observation of the importance of sorbed water in proton conduction in sulfonated matrices. Methanol permeability decreases as the inorganic content of the membrane increases. The permeability of O2, CO2, and N2, reduction products and oxidant stream constituents were evaluated as a function inorganic content of the membrane. It was found that as ZrO2 content increases, the per­meability of O2 and CO2 decreases and N2 permeability is unaffected. In general, the desirable characteristic of decreasing permeability of reactants, products, and oxidant stream constituents as the ZrO2 content increases were observed but with a corresponding decrease in proton conductivity.

Jorissen et al. evaluated DMFCs made with nonfluorinated PEMs (sPEEK and sPEK using PBI and basically substituted PSU (bPSU)) prepared by a thin- film method [24], in which the catalytic layer is sprayed directly onto the PEM. The authors used unsupported electrocatalysts at relatively high loadings (1.5 to 6.5 mg cm-2) to overcome the effects of flooding at both electrodes. The fuel cells were operated at 110°C with little or no cathode humidification and perfor­mance was compared to DMFCs made with Nafion 117 and 105. None of the DMFCs prepared from the test PEMs had optimized catalyst layer/PEM inter­faces, yet a number of the test cells performed similarly to the control cell made with Nafion 105. The best performing of these cells was a membrane composed of PEK, PBI, and bPSU that generated 85% of the power density (230 mW cm-2 @ 500 mV) of the Nafion 105 cell (270 mW cm-2 @ 500 mV). The control cell made from Nafion 117 exhibited poor performance generating less than half the power density of Nafion 105 cell under the same conditions.

Another nonfluorinated material of promise is poly(vinyl alcohol) (PVA). The material is chemically and thermally robust and quite inexpensive relative to Nafion. In a recent article by Khan et al., the synthesis and characterization of PVA-based membrane is described [66]. Membranes were based on PVA and its ionic blends with sodium alginate (SA) and chitosan (CS). The membrane ion exchange capacity (IEC, in milliequivalents of ion per gram dry polymer) were determined to be ~0.5 for PVA, 0.6 for PVA-CS and 0.8 for PVA-SA. All of the PVA-based membranes were found to have lower methanol permeability than Nafion. The PVA-CS membrane had the lowest permeability at 6.9 x 10-8 cm2 s-1 as compared to Nafion 117 with a permeability of 2.76 x 10-7 cm2 s-1. The other two membranes had permeability intermediate to PVA-SA and Nafion. Contrasting the desirable characteristic of lower methanol permeability than Nafion 117, the proton conductivities of the PVA-based membranes are signifi­cantly lower (~0.01 S cm-1) than that for Nafion 117 (0.1 S cm-1). The investi­gators hold out the possibility of improving proton conduction in the PVA mem­branes through doping.

Lee et al. also investigated the use of PVA-based materials. The group pre­pared membranes made from PVA, SiO2, and sulfosuccinic acid (SSA) [67]. The SSA acted both as the cross-linking agent and as the source of hydrophilic sulfonate groups. The SSA content of the membranes was varied from 5 to 25 wt %. Both the proton conductivity and methanol permeability decreased as the wt % of SSA increased to 20 wt %. At over 20 wt % SSA both trends reversed. Similar to what Khan et al. found, the membranes had proton conductivities on the order of 1 x 10-3 to 1 x 10-2 S cm-1 and methanol permeability of 1 x 10-8 to 1 x 10-7 cm2 s-1. In a related study, Lee and his group expanded the previous work by adding poly (acrylic acid) (PAA) to the hybrid membranes [68]. The thought was the addition of unreacted carboxylic acid groups may improve proton conductivity, but the PAA did not markedly improve upon the original hybrid membranes.

The roof of the aquaponics greenhouse is designed to capture rainwater and to melt and capture snow. Water is then transported through pipes to a central water — storage tank. Since this geographic area is subject to periodic droughts, the need to keep a supply of water is essential during the summer months.

Floor of the Structure

The floor of the aquaponics greenhouse is 61 1 thick concrete fitted with special plumbing that allows for the passage of hot water during the winter and cold water during the summer. It is connected to both the outdoor furnace and the recouperators on the microturbines. It has 31 1 of special insulation below it. Sump tanks, which gather flood water from the growbeds, are beneath the floor and accessible through a manhole in the top. It is then pumped back into the fish tank.

FUEL FLOW CHART

FIGURE 14.1 Flow chart of system.

Balanced Diet

Many types of plants can be grown in this system, if all of their cultural require­ments are met (light, heat, nutrients, and water). There are no weeds, so each plant can reach its full potential. The beds are comfortable to work with since they are hip height and require no stooping. As such, the project can potentially be used to provide a complete diet. Initially, the project will produce lettuce, broccoli, tomatoes, parsley, oregano, basil, cucumbers, peppers, onions, swiss chard, and spinach.

Lurgi claims to develop a cheaper way to make ultra-clean diesel fuel from synthesis gas via methanol. The process first converts methanol into propyleneth — esis; this is followed by olefin oligomerization (conversion to distillates), then product separation-plus-hydrogenation. The intermediate methanol-to-propylene step so far is only proven at demonstration scale.

The process would yield mostly kerosene and diesel, along with a small yield of gasoline and light ends. The near-zero sulphur/polyaromatics diesel fuel result­ing from this process would differ from more conventional Fischer-Tropsch diesel only in cetane numbe (>52 via “Methanol-to-Synfuel” versus >70 cetane for FT diesel). The incidental gasoline stream not only would be near-zero sulfur, but also have commercial octane ratings (92 RON, 80 MON) and maximally 11% aromatics (Peckham 2003).

Methanol to Gasoline

In the 1970s, Mobil developed and commercialized a methanol to gasoline (MTG) process. A plant was built in Montunui, New Zealand in 1985 and sold to Methanex. It produced gasoline until 1997 when the plant was permanently idled. If the gasoline is to be sold without additional blending, then further treating is necessary to reduce the amount of benzenes.

Deborah A. Samac, Hans-Joachim G. Jung, and JoAnn F. S. Lamb

USDA-ARS-Plant Science Research, University of Minnesota, St. Paul

CONTENTS

Current Alfalfa Cultivation and Utilization……………………………………………………. 80

Development and Cultivation of Alfalfa for Biomass…………………………………… 82

Chemical Composition of Alfalfa………………………………………………………………….. 83

Genetic Impacts on Composition…………………………………………………………………… 86

Alfalfa Leaf Meal…………………………………………………………………………………………… 87

Protein and Fiber Separation…………………………………………………………………………… 87

Pretreatment of Alfalfa Fiber………………………………………………………………………….. 88

Conversion Response after Dilute Acid Pretreatment…………………………………….. 89

Alfalfa Biotechnology and Genomics…………………………………………………………….. 90

Conclusions…………………………………………………………………………………………………….. 93

Acknowledgments…………………………………………………………………………………………… 94

References………………………………………………………………………………………………………. 94

Alfalfa (Medicago sativa L.) has considerable potential as a feedstock for pro­duction of fuels, feed, and industrial materials. However, unlike other major field crops such as corn and soybeans, which are commonly refined for production of fuel and industrial materials, refining of alfalfa remains undeveloped. Instead, alfalfa is primarily processed and used on-farm in the form of dried hay, silage, and fresh forage known as “greenchop,” or is grazed by animals in pastures. In many countries, including the United States, alfalfa is used as a basic component in feeding programs for dairy cattle and is an important feed for beef cattle, horses, sheep, and other livestock. Known as the “Queen of the Forages,” alfalfa provides highly nutritious forage in terms of protein, fiber, vitamins, and minerals for ruminant animals. If alfalfa is developed to its full potential as a feedstock for biorefining, a major shift may occur in the manner in which alfalfa is produced and used for feeding farm animals.

High-blend alcohol fuels have been used in vehicles for many years in different regions of the world. Brazil is probably the most well-known region, having established government policies in the 1970s to develop this fuel.2 During the 1990s, 4.5 million automobiles operating on 93% ethanol (balance gasoline), were in use in Brazil.3 In the United States, E85 has received increasing attention and use due to stricter emissions standards and worries about energy security. For example, in 1992 the federal government passed the Energy Policy Act (EPAct).4 This legislation was established with the goals of enhancing the nation’s energy security and improving environmental quality. The EPAct encourages the development and use of alternative fuels that are not substantially derived from petroleum.

Alternative fuels are defined to include alcohols at blends of 85% or more of alcohol (such as ethanol) with gasoline. The U. S. Department of Energy (DoE) is charged with the responsibility of implementing this act. The EPAct contains both voluntary and mandatory provisions designed to develop an alternative fuel economy. The EPAct’s voluntary activities are administered through the DoE Clean Cities Program, which helps create markets for alternative fuels and alter­native fuel vehicles (AFVs) through public/private partnerships in more than 80 U. S. cities.5

The mandatory EPAct provisions consist of four programs: The State & Alternative Fuel Provider Program; The Federal Fleet Program; Alternative Fuel Petitions Program; and the Private & Local Government Fleet Program. These programs give the DoE the power to require Federal and State governmental agencies to purchase AFVs as a percentage of their vehicle acquisitions. The Private & Local Government Fleet Program even gives the DoE the authority to impose AFV acquisition requirements on private and local government fleets, although this program has not been implemented.

In addition to the EPAct, the U. S. Federal government maintains a system of tax incentives for E85 in order to encourage its use and development. For example, a Volumetric Ethanol Excise Tax Credit of 51 cents per gallon of ethanol used in fuel is currently available to transportation fuel producers.6 This law also eliminates alternative minimum tax (AMT) on the Alcohol Fuels Income Tax Credit. Small ethanol producers are also provided with a 10 cents per gallon tax credit on up to 15 million gallons of production annually. Finally, states and the U. S. Federal government offer many grants to help in the production and use of E85.

In the United States, legislation and incentives have led to the development and use of many E85 capable vehicles. The U. S. Energy Information Adminis­tration (EIA) estimates that more than four million Flexible Fueled Vehicles were on U. S. roadways in 2002.7 The annual growth in E85 capable vehicles from 1996 to 2005 was 78.8%, and the projected E85 fuel use was projected to grow by 11.5% from 2003 to 2004 according to the EIA.1

In 1839, William Grove demonstrated the first fuel cell, which employed a very simple and basic system where generation of electricity was accomplished by supplying hydrogen and oxygen to two separate electrodes that were immersed in sulfuric acid [6]. A schematic of a hydrogen/oxygen fuel cell is shown in Figure 12.1. In a fuel cell, the respective fuel is oxidized at the anode producing and discharging electrons to an external circuit, which transfers them to the cathode where they are utilized along with discharged proton to reduce oxygen to water. Fuel cells (like batteries) consist of two electrodes and at least one electrolyte; but unlike batteries, their lifetime is much longer due to the fact they are easily recharged by the addition of more fuel to the anode chamber.

Batteries store energy, so they eventually expire, whereas a fuel cell is an energy conversion device that will produce power as long as fuel is supplied. Fuel cells were first developed for use in space vessels due to the demand for higher-power density and long-term power supply, which could not be delivered by traditional batteries [7]. Electrical power in outer space was required for operation of scientific data collection and transmission instrumentation to transmit information back to earth. The average battery has a lifetime up to 30 hours, which is relatively short for operations in outer space, where demand is at least 200-300 hours of constant supplied power.

The general interest in fuel cells is due to their potentially high efficiency. Efficiency is the ratio of energy produced to the amount of energy supplied, which is always higher than the energy produced. Conversion processes of one form of energy to another are never 100%. Therefore, the energy lost is actually converted to another form of energy, since the First Law of Thermodynamics declares that

H2

energy is conserved so it is neither created nor destroyed [5]. Energy loss goes either into the form of sound, light, and/or, commonly, heat.

Applications of fuel cells are enormous and very diverse, but can be catego­rized by power output requirements. For example, high power is needed for industrial applications, medium power for domestic installations, and low power is necessary for certain kinds of vehicles and for use in space as well as portable power devices [7]. Different fuel cell types are distinguished by the electrolyte utilized and have varied applications due to differences in stability and strengths of power supply as well as lifetime and working conditions. There are six different types of fuel cells that all have different applications and operational circum­stances; these are described in Table 12.1. They are alkaline fuel cell (AFC), proton exchange membrane fuel cell (PEMFC), direct methanol fuel cell (DMFC), phosphoric acid fuel cell (PAFC), molten carbonate fuel cell (MCFC), and solid-oxide fuel cell (SOFC) [8].

To give some idea of the abundance of biomass, the energy content of all biomass fuels available today would produce an estimated 2740 Quads (one Quad equals 1,000,000,000,000,000 BTUs) (4). Swedish physicist John Holmberg claims we have no energy crisis. He believes that since human society’s energy use is only about 1/13,000 of our daily solar income, the simple solution to the “crisis” is to harvest the abundance (5). That harvest is carried out day by day through photosynthesis, and stored in the form of biomass. This project then will unite all of its other energy-dependent educational projects into a cohesive whole by demonstrating the advantages of the local production of multiple-use biomass fuels.