Straw stems were first treated for 6 wk in small columns with P. ostreatus at 40 mg of P. ostreatus/g of stems and 1.6 g of H2O/g of dry stems, as described. These stems were then inoculated 1:10 into fresh stems, the moisture content was adjusted to 1.6 g of H2O/g of dry stems, and the mixture was added to the drums for treatment. Because of the large amount of nitrogen-limited inoculum needed for the initial small-column step, the nitrogen-limited inoculum was produced in a slightly different manner than described above. This mycelial inoculum was produced at Utah State University as previously, but the mycelia from the maintenance slants were transferred directly into the nitrogen-limited medium without first being enriched in YM broth. Thus, both of the enrichment steps in the preparation of this mycelial inoculum were carried out in the nitrogen-limited medium. The fungal pellets produced in this manner were harvested as before, shipped under refrigeration to the INEEL, and stored at 4°C until use (up to 4 wk).

After 6 wk the treated stems were removed from the glass columns and mixed by hand at 1:10 (w/w) with fresh, air-dried, uninoculated straw stems. Random samples of the 6-wk-degraded stems were dried, ground to 60 mesh, and analyzed for composition as described under Composi­tional Analyses. While it was not known how much inoculum would be necessary for the altered inoculation method, a 10 wt% inoculation of wood chips containing an active culture of the desired white-rot fungus has been successful in soil bioremediation (10). The moisture content was brought to 1.60 g of H2O/g of stems by spraying distilled water with a pressurized garden sprayer onto the fresh stems as they were mixed with the treated straw from the glass columns. The inoculated stems were then packed into 208.5-L drums at about 7.5 kg dry wt of inoculated straw per drum. Before loading the drums, a 56-cm-diameter perforated steel disk was placed in the bottom of each drum and elevated to about 5.7 cm above the bottom of the drum using screws. Humidified oil-free instrument air at 127.6 kPa was supplied at 400-500 mL/min to the bottom of each drum beneath the perforated disk; the pressure drop over each drum was about 41.4 kPa. The air exited the system separately through the centers of the lids of each drum through in-line 16-cm2 Whatman HEPA-Vent Filters with a porosity of 0.3 |im (Whatman, Newton, MA). After 6 or 12 wk of treatment, the drums were opened and several samples were removed from various locations within the straw beds. The samples were dried, ground to 60 mesh, and analyzed for composition as described under Compositional Analyses. The drums were then resealed and shipped to the Wood Mate­rials and Engineering Laboratory at WSU for analyses of various compos­ite formulations and extrusion testing. Untreated straw was also sent to WSU for these analyses. For the composite testing, the straw samples were referred to as Neat (untreated), Degrade1 (treated for 6 wk), and Degrade2 (treated for 12 wk).

To be able to study cellulose (Solka-floc) degradation without micro­bial conversion of the sugars formed, separate hydrolysis experiments were carried out in stirred flasks without cells. The conditions in these experiments were the same as during the corresponding fermentation (i. e., the same medium composition, pH 5.0, 28°C, 600 rpm). The medium composition was set according to the assumption that at the moment of Solka-floc addition, the fermentation broth would be depleted for the glucose originally present. The medium for hydrolysis experiments thus contained a single set of nutrients, 10 g/L of Solka-floc as the substrate, and added enzyme giving a desired activity. Two sources of enzyme were used: either Celluclast, a commercially available fungal cellulase prepara­tion (a kind gift from Novozymes, Denmark) or a home-produced cellu — lase enzyme prepared by collecting the supernatants of Solka-floc grown in shake-flask cultures of T. reesei Rut-C30 (Mandels medium, 28°C, pH 5.0, 4 d) by centrifugation (5600g, 10 min). The enzyme-to-substrate ratio was adjusted to mimic two chosen specific points in the batch cultivation on cellulose: the point of cellulose addition and the point at which the CO2 evolution peaked. Experiments were run in duplicate.

Results

T. reesei Rut C-30 was grown aerobically in batch cultures in two stages, using glucose in an initial phase to produce cell mass, and thereafter adding cellulose (in the form of Solka-floc) as described before. The dynamics of cellular activity and produced cellulases following the addition of Solka — floc was monitored by on-line measurements of CO2 evolution and sam­pling for determination of enzyme activity and sugar concentrations.

Truck delivery of material has a fixed cost associated with the time required to load and unload the truck, and a variable cost that is related to the time the truck is being driven and/or the distance driven. For most biomass delivery applications, truck speed is relatively constant over the route; thus, e. g., a truck picking up straw would average about 80 km/hr on rural and district roads, and a truck picking up wood chips in a forest would average about 50 km/h on logging roads. Only if the wood chips required a significant drive over highways would there be a second higher speed portion of the trip; this effect is ignored here. Figure 1 shows cost data per kilometer for truck transport of wood chips in a typical western Cana­dian setting ([3]; D. Evashiak, personal communication, 3/03); the intercept of the lines is the fixed cost of loading and unloading, and the slope is the incremental variable cost per kilometer. Table 1 provides the equations for transport costs, including straw (1). Figure 1 is adjusted to dry tonnes of biomass to make a comparison of pipeline costs easier; pipeline costs are discussed later. Typical field moisture levels for straw and wood in western Canada are 16 and 50%, respectively. The range of costs for truck transport of wood chips comes from two different types of estimate: the lower bound is from a Forest Engineering Research Institute of Canada (FERIC) study of chip transport costs from a long-term dedicated fleet, and the upper bound is based on current short-term contract hauling rates. The FERIC data are more representative of steady biomass supply to a long-term end use such as a power plant. Note that there is no change in cost with scale for any

biomass application of interest; that is, the amount of biomass moved fully utilizes multiple trucks and no savings occur with larger throughput.

Pipeline transport of wood chips was studied in the 1960. Brebner (4), Elliott (5), and Wasp et al. (6) examined solids carrying capacity and pres­sure losses, and Wasp et al. (6) did a cost analysis for a 160-km pipeline with one-way transport, i. e., no water return. These studies were focused on the supply of wood chips to pulp mills, and hence water uptake by chips did not have a downstream processing impact. More recently Hunt (7) did an extensive analysis of friction factors in wood chip slurries in water; in the present work, we utilize his formula for the friction factor.

More recently, Liu et al. (8) completed an analysis of two-phase pipelining of coal logs (compressed coal cylinders) by pipeline. In the present article, we draw on the work of Wasp et al. (6), Liu et al. (8), and discussions with a Canadian engineering contractor (D. Williams, personal communica­tion, 3/03) to develop pipeline cost estimates for transporting water slur­ries of wood chips; these costs are also shown in Fig. 1 and Table 1.

Delivery of material by slurry pipeline has a cost structure similar to that for truck transport. The fixed cost is associated with the investment in the material receiving and slurrying equipment at the pipeline inlet, and the separation and material transport equipment at the terminus. The slope of the curve comes from the operating cost of pumping, and the recovery of the incremental capital investment in the pipeline and booster pumping stations plus associated infrastructure such as power and road access, all of which increase linearly with distance. Technically, pipeline costs would have a slight "sawtooth" shape, with a slight, discrete increase in overall cost occurring when an additional pumping station is required. Practically, most of the incremental capital cost is in the pipeline rather than pumping stations, and the sawtooth effect can be ignored. (In our analysis, the pipe­line component of the total capital cost is 85% at 50 km, and 94% at 500 km.)

One key element in the pipeline scope and estimate is whether a return line for the carrying fluid is provided. This would be required in virtually

all circumstances if the carrying fluid were a hydrocarbon (e. g., oil) and would be required for water if upstream sources were not available, as might occur in a forest cut area, or if downstream discharge of separated water were prohibited. Tables 2—4 show the scope and cost estimate included in a two-way pipeline (i. e., one with return of the carrier fluid).

Key elements at the upstream end are materials receiving from trucks, dead and live storage, slurrying, and pipeline initial pumps. Key elements along the pipeline are the slurry and return pipeline and booster pumping sta­tions. Key elements at the discharge end are slurry separation and drainage of the wood chips, and material transport to the biomass processing facil­ity. As already noted, pressure drops, pumping requirements, and the overall estimate are based on water as the carrier fluid.

Note that unlike truck transport, there is an economy of scale in slurry transport of materials, since larger throughputs benefit from an economy of scale in construction of the pipeline and associated equipment, and in lower friction losses in larger pipelines.

Figure 2 compares the total transport costs of wood chips by truck and by pipeline, for an arbitrary fixed distance of 160 km. The basis of the cost estimate is a wood chip concentration of 27% by volume at the inlet end and 30% by volume at the outlet end. The close agreement between the estimat­ing formulae of Liu et al. (8) and the results of Wasp et al. (6) for a one-way pipeline is evident. The one-way pipeline cost estimates were cross-checked against a recent estimate of two short large-diameter liquid pipelines in western Canada (D. Williams, personal communication, 3/03), and showed good agreement. Figure 2 shows the impact of scale on pipeline costs, as compared with the cost of truck transport, which is independent of scale. (The formulae of Liu et al. (8) and the data from Bantrel [D. Williams, personal communication] suggest a capital cost scale factor for pipelines of

0. 59—0.62; the data of Wasp et al. (6) as not specific enough to calculate a comparable figure.) Figure 2 also shows the significantly higher cost for a two-way pipeline that returns carrier liquid to the inlet end.

From Figs. 1 and 2 it is clear that the marginal cost of transporting biomass by pipeline at a concentration of 30% is higher than truck transport at capacities <0.5 million dry t/yr (one-way pipeline) and 1.25 million dry t/yr (two-way pipeline) at a distance of 160 km. The implications of this finding are discussed in the next section.

Truck plus pipeline transport cost of woodchips with carrier return pipeline

Truck plus pipeline transport cost of woodchips

without carrier return pipeline

Truck transport of woodchips — FERIC

Э — Truck transport of woodchips — Short term contract hauling

Distance (Kms)

Fig. 3. Comparison of integrated truck/pipeline transport vs truck-only transport of wood chips at capacity of 2 million dry t/yr.

Ellen I. Burnes,1 John Hagen,1
Dennis Wichelns,*1,2 and Kristen Callens1

1 Department nf Agricultural Economics,

California State University,

Fresno, CA 93740,

E-mail: dwicbelns@csufresnn. edu; and
2Califnrnia Water Institute, California State University,
Fresno, CA 93740

Abstract

In this article, we estimate the costs of using alternative feedstocks to produce ethanol in a 40 million-gal facility in California’s San Joaquin Val­ley. Feedstocks include corn imported from Midwestern states and locally grown agricultural products such as corn, grapes, raisins, oranges, and other tree fruits. The estimated feedstock costs per gallon of ethanol include $0.92 for Midwestern corn, $1.21 for locally grown corn, $6.79 for grapes, $3.36 for raisins, $3.92 for citrus, and $1.42 for other tree fruit. Adjusting for coproduct values lowers the estimated net feedstock costs to $0.67/gal of ethanol for Midwestern corn, $0.96 for locally grown corn, $6.53 for grapes, and $3.30 for raisins. We also examine the potential increases in net revenue to raisin pro­ducers, made possible by having an alternative outlet available for selling surplus raisins.

Index Entries: Biofuels; renewable energy; raisins; ethanol; feedstocks.

Introduction

Prior to the winter of 2003, the primary oxygenate added to gasoline sold in California was methyl tert-butyl ether (MTBE). Since that time, refiners in California have been discontinuing the use of MTBE, while increasing their use of ethanol as an oxygenate. As MTBE use is discontin­ued, most of the ethanol that will be used in its place likely will be imported from other states. An economic analysis of the potential for producing

*Author to whom all correspondence and reprint requests should be addressed. Applied Biocbemistry and Bintecbnnlngy 95 Vol. 113-116, 2004

ethanol in California is timely and appropriate, given that MTBE is being discontinued, and that California has a large agricultural industry that may benefit from increased demand for some of its products.

In the early 1980s, motivated by fuel shortages, geopolitical uncer­tainty, and high fuel costs, California developed the capability of produc­ing fuel ethanol. Producers demonstrated the ability to make ethanol from local feedstocks including agricultural waste, industrial waste, and other biomass sources. Five ethanol production facilities were constructed in California during that time. Three of those facilities were closed within 10 yr, when fuel prices declined, feedstock costs rose, and subsidies for ethanol production were ended. Nonetheless, California gained valuable experience while the plants were operating. In particular, producers dem­onstrated that ethanol could be produced in California, provided that sub­sidies were available. Producers also learned that plant location and the choice of feedstocks are important firm-level decisions and that regional economics and political considerations influence the financial viability of ethanol production.

As the most productive agricultural region in the United States, it seems appropriate to reconsider the role that ethanol production might play in California. In this article, we estimate the costs of using alternative feedstocks to produce ethanol in a 40 million-gal facility in California’s San Joaquin Valley. We consider seven materials that might be used as an ethanol feedstock: citrus, grapes, raisins, deciduous tree fruit (peaches, plums, and nectarines), corn, almond hulls, and whey. Any of these, except corn, can be described using one or two of the following statements: (1) it is currently produced in surplus amounts in the San Joaquin Valley (grapes and raisins), (2) it is a culled product (grapes, citrus, tree fruit), or (3) it is a byproduct (almond hulls and whey). Ethanol production from any of these sources would enhance the farm-level economics of the primary crop activity by generating new demand for culls, surplus, and byproducts that would otherwise be wasted or sold for a minimal price.

We calculate the costs of using alternative feedstocks on a per-gallon- of-ethanol basis, to enable ready comparison of the relative costs and ben­efits of each feedstock. We assume that ethanol producers must purchase feedstocks at the prevailing market prices of the culls, byproducts, and surplus products. Our results suggest that the cost of producing ethanol using California agricultural products or Midwestern corn is higher than the current price of ethanol. Hence, a public subsidy would be required to encourage ethanol production using any of the feedstocks we examine.

An alternative view of the ethanol question is: would crop producers be willing to sell a portion of their surplus production at a price that ethanol producers would be willing to pay in the absence of a subsidy program? That question is particularly pertinent when producers must store their surplus production for some period of time before it is sold in a primary market. The net price received by producers in those markets declines with the length of time that the surplus production is held in storage. Hence, producers might gain by selling a portion of their surplus production to ethanol producers at a price below the primary market price. We examine this possibility for the case of raisin production and storage in California’s San Joaquin Valley.

Surplus raisin production occurs often in California, and the sale of surplus raisins is restricted somewhat by a federal marketing order. The high cost of storing raisins reduces the net revenue earned by farmers, and the carryover of production from one year to the next can have a depressing impact on future raisin prices. We find that between 1992 and 2001, an ethanol industry would have generated greater revenues for raisin produc­ers in 6 of those 10 yr. We consider only the farm-level benefits of having an alternative outlet available for selling a portion of the surplus raisin production. In particular, we examine the decision to allocate surplus rai­sins from storage, either to the food market or to the ethanol market. We do not consider raisin production costs, because those have already been paid. We assume that the ethanol plant exists, and that its owners would be willing to purchase raisins for use as a feedstock.

Our goals in this article are (1) to estimate the costs of using alternative feedstocks to produce ethanol in a 40 million-gal facility in California’s San Joaquin Valley, and (2) to demonstrate the impact of storage costs on the decision to sell surplus crops for use in ethanol production at a price below the price available in primary markets. We describe the availability and cost of the agricultural products that might be used as feedstocks for etha­nol production in California. We use that information to estimate the costs of producing ethanol and to examine the impact of storage costs on crop marketing decisions.

James Kamm

University of Toledo,

Toledo, OH 43606,

E-mail: jkamm@utnet. utoledo. edu

Abstract

Directly burnable biomass to be used primarily in steam boilers for power production has been researched and demonstrated in a variety of projects in the United states. The biomass typically comes from wood wastes, such as tree trimmings or the byproducts of lumber production, or from a cash crop, grown by farmers. Of this latter group, the main emphasis has been utilizing corn stover, or a prairie grass called switchgrass, or using tree seedlings such as willow. In this article, I propose an alternative to these energy crops that consists of several different herbaceous plants with the one consistent prop­erty that they annually generate an appreciable bulk of dried-down burnable mass. The fact that they are a set of plants (nine are offered as candidates) gives this energy crop a great deal of flexibility as far as growing conditions and annual harvest time line. Their predicted yield is impressive and leads to speculation that they can be economically feasible.

Index Entries: Biomass; biofuel; energy crop; sclerified stalked plants; stiff stalked plants.

Introduction

The prospect of biomass as a fuel source is an alluring one. In the first place, it is geopolitically simple; most countries desiring to utilize it can provide their own biomass. In addition, many forms of biomass-to-energy conversions are CO2 friendly, adding no net CO2 to the atmosphere or at least no additional CO2 other than would have naturally taken place. Prob­ably the most exciting aspect of biomass fuels is that they are replaceable; there is no doomsday worry about using up all of EartMs resources. Indeed, some types of biomass are not only replaceable but also renewable; they are or can be created at the same rate that they are used. [4]

For all of the advantages of biomass fuel sources, there are two dis­tinct drawbacks. First, biofuels do not economically compete with conven­tional (oil, gas, coal) fuels. It costs more to generate electricity from biomass compared to coal, and it costs more to power automobiles using biomass fuels compared to gasoline. Second, and probably more important, the sources of biomass are quite diffuse and may not be available in sufficient quantities to make a national impact as an energy use. Biomass is defined as organic material from animals, such as manure, or plants such as trees, grasses, and agricultural crops. Common examples are sawdust as a byproduct of milling wood, rice husks as a byproduct of food production, and pallet and wood crate discards. Ohio claims to produce <1% of its electricity with biomass, and it does it with "forest wastes, such as tops and limbs, and wood wastes, such as sawdust, chips, barks, and edgings" (1). Most of this electrical generation is done within the wood-manufacturing industry (2) and is primarily used internally by the company that gener­ates it.

Plant biomass can also come from "energy crops." Energy crops are "crops developed and grown specifically for fuel. These crops are selected to be fast growing, drought and pest resistant and readily harvested to allow competitive prices when used as fuel" (3). The prospect that farmers will use portions of their vast acreage to produce a material that can be used for fuel seems to have a ring of creditability to it. Whether the power indus­try can pay the farmers enough to turn their heads away from the familiar markets of soybean, wheat, and corn is another issue. Farmers can provide the quantity of material to make biofuel a significant factor in the energy source equation.

Energy crops, or biofuel feedstocks, "under development in the US include hybrid poplar, willow, switchgrass, and eucalyptus" (3). Indeed, prospective energy crops appear to come from both ends of the spectrum of the plant kingdom—monocot herbaceous plants, on the one hand, to more sophisticated hardwood trees, on the other. Switchgrass is a prairie grass (monocot) that grows favorably in the plains states and southwestern states. It dries down to give a burnable product, and the crop has benefited from significant amounts of research performed in states such as Iowa (4), Wisconsin, and Texas and from the Department of Energy through Oak Ridge National Laboratory. On the other hand, willow is a tree, a woody plant whose trunk has annular rings such that it adds woody matter annu­ally to its stem. It appears to be the favored energy crop under investigation in East Coast states, particularly New York where coburning with coal has produced some electricity (5). Both of these energy crops have their propo­nents, both have been the subject of a considerable amount of research and demonstration over the past 10-15 yr, and both are looking more and more promising.

The intent of this article is to introduce another type of plant into the energy crop mix. Of this new type, I suggest nine specific species. These species come from the center of the plant kingdom, between monocots (grasses) and woody plants. All are herbaceous, either annual or perennial. The one common characteristic to all is that their stems die off annually and dry down to give a brittle stiff and relatively hard skeleton. These "remains" of the plant consist primarily of organic carbon-rich compounds and become a source of energy (fuel). Herein, they are referred to as stiff stalked plants (SSPs).

Before introducing and discussing these plants on an individual basis, and in order to understand the significance of the measurements made on and the comments made about each, some consideration should be given to plant botany as it relates to the characteristics of a good energy crop. Plant botany or physiology takes on a completely different thrust when viewed from the perspective of using the dead and dried stem for the ulti­mate purpose. A plant skeleton is really much different, and much simpler, than its living counterpart.

The negative values of the estimated net returns suggest that surplus California agricultural products may not be viable ethanol feedstocks. However, this conclusion is not satisfying when one considers that current surplus conditions generate high storage and product transformation costs. For example, the cost of storing raisins is $11.00/t per month, and there were more than 200,000 t of raisins in storage at the end of the 2001 produc­tion season. Oranges and grapes that are either surplus or culled are stored as juice or concentrate for later use in the food and beverage industry.

Storage costs reduce the net revenue received by farmers when they sell their produce in primary markets. Hence, producers might gain net revenue by selling a portion of their surplus production in a secondary market, such as that for ethanol, rather than paying substantial storage costs while waiting for the sale of their produce in a primary market. This problem can be viewed as one of determining the net revenue maximizing strategy for allocating surplus production between storage and a second­ary market. We examine this problem for the case of raisin production and marketing in California’s San Joaquin Valley. We assume that surplus pro­duction can be stored for later sale in the primary market, or sold for use in ethanol production.

Raisin production and marketing in California are conducted within the framework of a federal marketing order. Each year, the Raisin Advisory Committee determines the quantity and value of raisins that are allocated for sale in the "free market," and the quantity and value of raisins held in "reserve." Approximately 240,000 t of raisins are allocated annually to the free market, while 60,000 t are targeted for the reserve ([6]; M. Pello, per­sonal communication, 3/6/03). Between 1991 and 2000, total raisin pro­duction in California ranged between 240,000 and 437,000 t. The amount of raisins allocated to the free market remained consistent during those years, whereas the amount allocated to the reserve pool ranged from 0 t in 1998 to 205,000 t in 2001.

Before 1980, the results of federal agricultural research were freely available and widely shared. Virtually all of the research was directed toward improving crop production and yields as well as harvesting and storage costs of crops intended for food and feed markets.

There are considerable uncertainties regarding cost-benefit analyses. Nevertheless, it is instructive that the many studies done both inside and outside the United States Department of Agriculture (USDA) found its pre — 1980 R&D efforts very effective and influential.

USDA economists found that publicly funded agricultural research earned an annual rate of return of at least 35 %.1 A 1966 study by the Agricultural Research Service (ARS) on the impact of its research from 1941 to 1966 concluded that 109 products and processes developed by ARS had been commercialized and 26 represented major contributions in basic research. Their value was estimated by the ARS at more than $6 billion, 20 times the $309 million spent by the ARS laboratories during this period.2

A 1980 study by the Congressional Office of Technology Assessment on the benefits stemming from agricultural research concluded that, "the range of estimated rates of return is from a low of 23 percent to a high of 100 percent."3 A 1992 study by Chapman and Associates (1) examined 178 cases of ARS research projects completed from 1980 to 1990 (including coopera­tive programs or joint programs with State Agricultural Experiment Sta­tions). Of the 178 cases, benefits data were identified for 87, resulting in $14.8 billion in sales or savings. These savings were greater than the total amount spent on the ARS during that time period (1).

Another study (2) found that although the ARS had a relatively small number of patents compared to the private sector in agricultural-related areas, the ARS patents were cited more often than private patents. Thus, the ARS patents were considered more often "key" patents marking significant advances in knowledge (2).

Despite these successes, in the late 1970s there was a growing and increasingly influential school of thought that a focus on more basic research and the nonexclusive sharing of the fruits of such research was not encouraging the levels of private investment sufficient to commercial­ize new technologies. Many potentially valuable scientific advances were therefore remaining in the laboratories. Commercialization would occur only if private investors could be guaranteed exclusive access to the knowl­edge generated from what would increasingly become investments in fed­eral research efforts made by both private and public sectors.

In rapid fashion, beginning in 1980, the federal government dramati­cally changed its R&D strategies to encourage one in which private inter­ests would become increasingly influential in directly assisting public research:

1. The University and Small Business Patent Procedure Act, commonly known as the Bayh-Dole Act of 1980, gave nonprofit organizations such as universities as well as small businesses the right to retain patents for technology developed with government funds.

2. The Stevenson-Wydler Technology Innovation Act of 1980 provided federal departments, agencies, and affiliated laboratories with a leg­islative mandate to pursue technology transfer activities. Each agency was to make available not less than 9.5% of its R&D budget for technology transfer activities.

3. In 1983, an Executive Order extended the coverage of the Bayh-Dole Act to all government contractors. The Act also granted federal agen­cies the right to offer exclusive or coexclusive licenses to patents on inventions made by laboratory employees considered necessary for the commercialization of the invention.

4. The Federal Technology Transfer Act of 1986 allowed federal labo­ratories to enter into Cooperative Research and Development Agree­ments (CRADAs) with private firms. A CRADA confers two important rights to businesses: First, the right of first refusal of an exclusive license on any patentable inventions that arise from the research partnership; and, second, the right to keep research find­ings secret for 5 yr. The Act also permitted royalty income from patent licensing and assignment to be distributed directly to the inventors. The 1986 Act also made technology transfer a responsibil­ity of every laboratory scientist and engineer. It required at least one full-time equivalent technology transfer position for every labora­tory having 200 or more full-time scientific, engineering and related positions.

Today much if not most federal research, including biomass-related research, is done in partnership with private companies that have the right to exclusively own the intellectual property generated from that collaboration.

How effective has the post-1980 approach been compared to its pre­decessor?

Unfortunately, there are few if any studies that adequately address this important question. Vast changes in agricultural technologies have occurred over the last 20 yr, especially in the area of biotechnology in crops and animals. Yet, in this area federal R&D spending may have played a modest role that largely followed the massive amounts of venture capital that flowed into the biotech sector.

Efforts to compare the pre — and post-1980 R&D strategies are con­founded by the fact that the metrics used to evaluate the performance of federal research have changed. In the older period, the measures used largely reflected the impact on the country and the countryside, such as the number of acres planted in the new hybrid and the rate of adoption of a new technology by farmers or processors. The new approach largely measures the impact on the agency or its private partner, considering factors such as the number of patents issued, the number of licenses issued, and the amount of royalties received.

It is now more than 20 yr since the federal government adopted a dramatically different approach to R&D by emphasizing technology trans­fer, private partnerships, and exclusive licensing. This is sufficient time to allow evaluation of the comparative effectiveness of both approaches.

Ash content was determined as follows: At least 1 g of dry stems, ground to 60 mesh, was ashed in a muffle furnace at 650°C for 18-24 h. Ash content was calculated by weight difference. Carbohydrate and lig­nin compositions of untreated and treated straw samples were deter­mined by quantitative saccharification using the method of Saeman et al.

(11) . Two aliquots of each sample were analyzed by quantitative saccha­rification for each of the three replicate columns at each condition, for a total of 12 independent measurements of each composition. Carbohy­drate analyses were done by high-performance liquid chromatography using a Bio-Rad HPX-87P carbohydrate column as previously described

(12) . The acid-insoluble fraction from the quantitative saccharification was ashed at 650°C, and Klason lignin with extractives was calculated by weight difference. The amounts of glucan and xylan degraded per 100 g of initial weight were calculated by mass balance assuming that the sum of lignin, ash, and extractives remained constant. The percentage conver­sions of glucan and xylan (AG and AX, respectively) were then calculated by dividing by the initial basis weight of each and multiplying by 100.

A two-phase aerobic batch culture is shown in Fig. 1. The initial expo­nential growth phase on glucose lasted until about 26 h. The specific growth rate (determined from the logarithm of the CO2 evolution rate [CER]) was 0.22 h1. After 25.4 h, DNS measurements of reducing sugars indicated that the glucose was exhausted from the medium (Fig. 2). Although the CO2 evolution rapidly decreased, there was a residual CO2 evolution, which only gradually decreased to zero. This was accompanied by a reduction in the OD (Fig. 2), indicating degradation of biomass. The measured cellulase activity remained constant after the depletion of glucose at a level of about 0.3 FPU/mL.

Second-Stage Batch Cultivation on Cellulose

After 67 h, cellulose in the form of Solka-floc was added as described above. There was an immediate increase in the CER at this point. The increase continued until t = 73 h, at which point there was a sharp decrease in CER. This did not coincide with complete depletion of glucose from the medium (Fig. 3). The enzyme activity increased continually up to a value of about 2.6 FPU/mL. This maximum coincided with the depletion of glucose and occurred at about t = 100 h. From the integrated area of the CO2 evolution (Fig. 1), one can estimate that the CO2 evolved on cellulose was about 80% of the value obtained from glucose.

Fig 5. Concentration of glucose (———- ) and cellobiose (——— ) during enzymatic

hydrolysis of Solka-floc. Two different enzymes were used; enzyme was prepared from a culture of T. reesei (□) and commercially available Celluclast (A). Enzyme loading was 91.9 FPU/g substrate. Hydrolysis was carried out at 28°C and pH 5.0.

In Vitro Enzymatic Hydrolysis Rates

The initial, rather rapid increase in CER found in the second stage of two-phase batch cultivation was somewhat unexpected. We decided to compare this value to the initial rates of glucose and cellobiose formation in in vitro enzymatic hydrolysis experiments (Figs. 4 and 5). The enzyme loadings were chosen to represent the actual enzyme-to-substrate ratio relevant to the point of cellulose addition and the point of maximum CER.

For the enzyme solution prepared using T. reesei, the approximate forma­tion rate of glucose was in the former case 0.18 g/(L-h) and in the second case 0.5 g/(L-h). The formation rate of cellobiose was 0.33 g/(L-h) in the first case and <0.2 g/(L-h) in the second case. Using Celluclast in a loading revevant to the point of maximum CER, a similar value was found for glucose, but a higher value was found for cellobiose (Fig. 5).

With a typical yield of CO2 on sugar, Ysc, of 0.4 (C-mol/C-mol), one can estimate that the sugar formation rate would give a CER of 0.018 mol of CO2/h for a 2-L culture as in Fig. 1. This corresponds to a CO2 concentration in the outlet gas of 1%. Within 0.5 h after addition of cellulose (Fig. 3), the measured value was in fact 1%, in good agreement with the estimated value. Calculations for the higher enzyme activity (91.9 FPU/mL) indicate that the formed glucose and cellobiose would give a CER of 0.018 mol of CO2/h, corresponding to a CO2 concentration in the outlet gas of 1.4%, which is a bit lower than the actual observed value.

Discussion

The final cellulase activity obtained from 10 g/L of cellulose was 2.6 FPU/mL. This is in good agreement with previously reported yields for the strain Rut-C30. For example, Persson et al. (6) quote a yield of 233 FPU/g of substrate in batch cultures. There was a steady increase in cellulase activity throughout the cultivation on cellulose in the current work, despite the fact that the free glucose concentration reached a value as high as 1 g/L. How­ever, at that point of maximum glucose concentration, a sharp decrease in CO2 evolution occurred, and the glucose concentration started to decrease after that point. The reason for this may be depletion of a medium compo­nent, or it may also be related to the regulation of enzyme expression. This is supported by the fact that the rate of activity increase changes at that point.

On depletion of glucose in the initial growth phase, there was a rapid decrease in CO2 evolution, but it did not decrease to zero. Measurements of OD660 showed a decrease in biomass, suggesting that the residual CO2 evolution is the result of endogenous metabolism. A higher volumetric enzyme productivity could therefore potentially be obtained if cellulose had been added earlier, provided that there was no remaining glucose repression effect.

The main point of making a two-stage culture was to enable the study of a pulse addition of cellulose. However, separating an initial biomass formation on glucose (or on other monosaccharides) from cellulase pro­duction with cellulose as substrate has advantages also from a process point of view. By using a two-stage process, a basal cellulase activity can be obtained before the addition of cellulose. This level allows the utilization of cellulose to commence rather quickly as shown by the CO2 evolution. By contrast, a one-phase batch process starting directly from cellulose will initially be very slow owing to a very low hydrolysis rate. As has been pointed out previously, a key question is, to what extent will the cellulase expression be repressed by the glucose liberated in the hydrolysis.

Acknowledgment

This work was supported in part by a fellowship from the Marie Curie Training Site QCIM. We also acknowledge the National Research Fund of Hungary (OTKA T029382) and the National Research and Development Program (NKFP-OM-00231/2001) for financial support.

References

1. Wyk, J. P. H. (1999), Biomass Bioenergy 16, 239-242.

2. Kheshgi, H. S., Prince, R. C., and Marland, G. (2000), Annu. Rev. Energ. Environ. 25, 199-244.

3. Sun, Y. and Cheng, J. (2002), Bioresour. Technol. 83, 1-11.

4. Bhat, M. K. (2000), Biotechnol. Adv. 18, 355-383.

5. Himmel, M. E., Ruth, M. F., and Wyman, C. E. (1999), Curr. Opin. Biotechnol. 10, 358-364.

6. Persson, I., Tjerneld, F., and Hahn-Hagerdal, B. (1991), Process Biochem. 26, 65-74.

7. Beguin, P. and Aubert, J. P. (1994), FEMS Microbiol. Rev. 13, 25-28.

8. Tolan, J. S. and Foody, B. (1999), in Advances in Biochemical Engineering/Biotechnology, vol. 65, Scheper, T., ed., Springer-Verlag, Berlin, Germany, pp. 40-67.

9. Yu, X. B., Hyun, S. Y., and Yoon-Mo, K. (1998), J. Microbiol. Biotechnol. 8, 208-213.

10. Ilmen, M., Saloheimo, A., Onnela, M.-L., and Pentilla, M. E. (1997), Appl. Environ. Microbiol. 63, 1298-1306.

11. Kubicek, C. P., Messner, R., Cruber, F., Mach, R. L., and Kubicek-Pranz, E. M. (1993), Enzyme Microb. Technol. 15, 90-99.

12. Suto, M. and Tomita, F. (2001), J. Biosci. Bioeng. 92, 305-311.

13. Mandels, M. and Weber, J. (1969), Adv. Chem. Ser. 95, 391-414.

14. Bigelow, M. and Wyman, C. E. (2002), Appl. Biochem. Biotechnol. 98/100, 921-934.

15. Miller, G. (1959), Anal. Chem. 31, 426-28.

16. Mandels, M., Andreotti, R., and Roche, C. (1976), Biotechnol. Bioeng. Symp. 6, 21-23.

Copyright© 2004 by Humana Press Inc.

All rights of any nature whatsoever reserved. 0273-2289/04/113/0125-0136/$25.00

Any real application of pipeline transport of biomass from a field location (as opposed to mill residue) will normally require an initial truck haul to get the biomass to the pipeline inlet. This means that the fixed costs associated with both truck and pipeline transport are incurred. Thus, e. g., truck hauling of 2 million dry t/yr of biomass to a pipeline inlet at an average haul distance of 35 km (1), as might occur in a whole-forest harvest operation, with further transport of biomass by one — or two-way pipeline would have cost curves as shown in Fig. 3. The alternative of transport by truck alone is shown by the dashed line in Fig. 3.

Since by inspection of Fig. 1 all pipelines with a capacity of <0.5 million dry t/yr (one-way) or 1.25 million dry t/yr (two-way) have a higher incre­mental cost (slope) per kilometer than the alternative of hauling by truck, it is clear that pipelines below this capacity cannot compete with the alter­native of leaving the biomass on the truck for the extra distance. In the example illustrated in Fig. 3, at 2 million dry t/yr the minimum pipeline distance to recover the fixed costs of the pipeline as compared to truck haul are 75 km for a one-way pipeline (in addition to the initial 35-km truck haul to the pipeline inlet), and 470 km for a two-way pipeline (again in addition to the initial truck haul); pipeline distances shorter than this are less eco­nomic than continued hauling by truck. Hence, pipelining of truck-deliv­ered biomass at a concentration of 30% is only feasible at both large capacity and medium to long distances.

Immersion time (hours)

Fig. 4. Carrier fluid content of biomass after different hours of immersion in carrier fluid.

Absorption of Carrier Fluid by Biomass

We performed a series of simple experiments to explore the uptake of carrier fluid by biomass. Fresh wood chips, both hardwood (aspen) and softwood (spruce), were kept sealed and cool until immersion in room temperature water or oil; they were drained and dried to determine mois­ture level. Water drainage was brief, about 1 min., although one test of a longer drainage period showed a negligible impact of longer drainage times. The oil used in our study is a heavy gas oil fraction from Syncrude Canada, with a nominal boiling range of approx 325—550°C and a viscos­ity of 1.3 Pa s at 20°C. This type of oil is typical of an industrial-grade furnace oil. Wood chips were drained of oil for 1 hr before weighing. Figure 4 shows the carrier fluid content of biomass after exposure to car­rier fluid for varying periods of time. Note that immersion time can be related to pipeline distance because at a typical slurry velocity of 1.5 m/s, the slurry would travel 5.4 km/h.

The choice of an oil carrier requires a tradeoff between the viscosity of the carrier, which drops with lower boiling range of the oil fraction, and the value of the carrier, which increases with lower boiling range. At one extreme, a diesel fraction would have low viscosity but has such a high value as a transportation fuel that its use as a thermal fuel would be cost prohibitive. At the other extreme, a residuum fraction would have low value but such a high viscosity that transport of the slurry would likely be prohibitive in operating (pumping) cost. In the present study, we have arbitrarily selected a heavy gas oil as the balance between these competing considerations.

During water immersion, 1 kg of mixed spruce and aspen wood chips at an average 50% water content would pick up an additional 0.51 kg of water and reach a terminal moisture level of about 67%. Water uptake is quick; even after immersion for 3 h moisture levels exceed 63%. This is similar to the findings of Brebner (4) and Wasp et al. (6), who reported saturated wood values of 65%. We conducted two experiments with straw and found that moisture level rose from 14% as received to >80% after exposure of 3 h. This is similar to the findings of Jenkins et al. (9) for rice straw from California.

Absorption of water has serious implications for any process such as direct combustion that converts absorbed liquid water in the fuel into emit­ted water vapor in the flue gas, in that it reduces the lower heating value (LHV) of the biomass and requires more biomass per unit of heat released by combustion, an effect also noted by Yoshida et al. (10). Figure 5 shows the loss in LHV and the corresponding increase in biomass that must be delivered to a direct combustion-based biomass operation at 67% mois­ture level. Werther et al. (11) note some other problems with increasing moisture in the direct combustion of biomass: reduced combustion tem­perature, delayed release of volatiles, poor ignition, and higher volumes of flue gas. These secondary impacts on efficiency and operability of a direct combustion unit are not considered in Fig. 5.

One can conceptually break down biomass utilization into three com­ponent cost categories: (1) field harvest of biomass, (2) transportation from the field to the biomass processing site, (3) cost of processing/conversion. For direct combustion of truck-transported biomass from harvesting of the whole forest in western Canada at or near optimum scale, the percentage and cost per megawatt-hour are as follows: category 1: 33.4%, 15.77$/MWh; category 2: 14.3%, 6.74$/MWh; and category 3: 52.3%, 24.65$/MWh (1).

Since, as shown in Fig. 5, changing the moisture level of wood chips from 50% to 67% increases the requirement for field biomass in direct combus­tion by 78% for a given output of heat and power, it is evident that water — based pipelining of wood chips cannot be economical for direct combustion, because the increase in field harvest cost associated with the higher biom­ass requirement is larger than any possible transportation cost saving. For straw, so much water is taken up that the LHV is effectively zero; pipeline transport of straw to a direct combustion application would destroy the heating value of the fuel.

This impact is not true for a fuel process such as supercritical water gasification of biomass (12—14) that does not produce water vapor from absorbed water, since the higher heating value (HHV) value of the biomass is effectively realized by countercurrent exchange of heat between prod­ucts and feed that results in condensation of produced water. The impact of absorbed water is also not an issue for fermentation of biomass, since this is a water-based process. Pipelining of biomass to fermentation processes offers the promise of larger-scale, more economic processing of ethanol, chemicals, and byproducts such as lignin. However, the pipeline design would require more detailed assessment since saccharification in the pipe­line would be a logical processing alternative, and this would require tem­perature control during pipeline transport. This more detailed assessment is the subject of future study.

During oil immersion for 48 h, 1 kg of mixed conifer and aspen wood chips at an average 50% water content would pick up an additional 0.45 kg of oil and reach an oil level of 31%. Comparable figures for 124 hours are an uptake of 0.52 kg to reach a oil level of 34%. Direct combusting wood chips delivered in a heavy gas oil can be thought of as cofiring a mix of about two/ thirds oil and one/third wood on a thermal basis. Pipeline cost would increase because of additional pumping; the increase would depend on the viscosity of the oil fraction that was selected as the transport carrier fluid.

Discussion

Pipeline transport of oil and natural gas is clearly far more economical than truck transport, even in relatively small pipelines. Three factors com­bine to make the transport of energy in the form of biomass far less economic:

1. The density of energy in the pipeline is far lower for biomass than for oil. The present work is based on 30% biomass by volume in a carrier liquid. Wasp et al. (6) based their work on 22% biomass. Brebner (4) and Elliott (5) indicated that at about 47% concentration by volume a slurry of wood chips and water cannot flow. Given the low heat content of wood per unit volume relative to oil and the low concen­tration of wood chips in water, the energy density in a 30% wood chip slurry is about 8% compared to oil, even based on HHV, and hence far larger pipelines are required to transport the same amount of energy.

2. The pressure drop in the pipeline is high for suspended solids in a carrier fluid. For example, Wasp et al. (6) indicate that at 30% con­centration of wood and a velocity of 1.4 m/s, a wood chip slurry in a 214-mm-diameter pipeline has a pressure drop that is three times larger than for water alone.

3. Recycle of the carrier fluid will often be required in biomass trans­port by pipeline, both because large quantities of water will not be available at the inlet end and because discharge of water that has carried the biomass will, in some jurisdictions, be prohibited. This requires that a second pipeline and set of pumping stations be con­structed.

In addition to these cost elements, transport of biomass for a direct combustion application by water creates a prohibitive drop in the LHV of the fuel because of absorbed water. These issues limit the application of pipeline transport of biomass to large applications that use oil as a carrier medium, or that supply a process for which the heat content of the fuel is not degraded by the requirement to remove absorbed water as vapor, such as a supercritical water gasification process.

Transport of wood chips by oil precludes firing a high percentage of biomass owing to high oil uptake by wood chips. We consider it unlikely that a two-thirds oil and one-third wood fuel mixture would have high interest today as a power plant fuel, since even a heavy gas oil fraction has too high a value as a transportation fuel precursor to be diverted into power generation.