Production of ethanol from starch-based crops such as wheat and corn is a well-known technology. Such processes have been optimized over a long time and are reaching a level of maturity where further cost reductions, based on improvements in conversion technology, are becoming more difficult. In contrast, processes using lignocellulosic raw materials are still under develop­ment and significant reductions in ethanol production cost can be expected.

With some modifications it is possible to make a basic evaluation of a starch-based process (Fig. 1) and compare it with a process based on ligno­cellulosic material (Fig. 3). This was done by Wingren [17]. The purpose of the evaluation was not to determine the absolute ethanol production cost, but to compare the processes using the same fundamental cost basis and the same assumptions in the investment analysis. A comparison of this kind provides valuable information on the major differences between a commercial process and a process under development.

Both plants were designed for an annual ethanol production of 55 000 m3, which is a rather small plant. This value is on a pure ethanol basis, although the actual distillate was assumed to be 94% (w/w), i. e. no dehydration step was included as this would have been the same in both cases. Also, no off­sites, e. g. production of heat and electricity, were included, only the pure ethanol production facility. In the evaluation no credit was given for carbon dioxide. The cost of the enzymes in a starch-based plant is lower than in a lig- nocellulosic plant. In the study it was assumed to be 0.014 US$ L-1 ethanol, which is slightly higher than the cost reported for the enzymes in a corn — based plant located in the USA [18].

The raw material flow is higher in the lignocellulosic process, 200 000, compared to 126 500, dry metric tons y-1 for the starch-based plant, due to the lower overall ethanol yield and the somewhat lower amount of fermentable sugars in the raw material. The overall energy demand in the lignocellulosic process was estimated to be 16 MJ L-1 ethanol compared to 10 MJ L-1 for the wheat-based process. The fixed capital investment was estimated to be 99 and 53 million US$ for the lignocellulosic and the starch-based processes, respec­tively. A breakdown of costs is presented in Table 3. The estimated ethanol production cost was 0.60 and 0.58 US$ L-1 for the lignocellulosic and starch — based processes, respectively. Major differences were found in the cost of raw material, enzymes, capital, steam as well as income from the co-products. It should be noted that, although significantly higher than in the starch-based process, the enzyme cost in the lignocellulosic process was based on a pro­jected future cost. In the starch-based process the cost of the raw material constitutes as much as 65% of the total production cost. This is typically the case for well-established, mature processes. Thus, the economics of a starch — based process is very dependent on the cost of feedstock.

The lignocellulosic process is more dependent on the income from the co-products. However, the potential price of the syrup is uncertain since its fuel properties are unknown. At 12.9 US$ MWh-1 the income from this co­product was estimated to be 0.03 US$ L-1. In a scenario where the co-product instead has to be disposed of and cannot be utilized as a fuel, the ethanol production cost for the lignocellulosic process would be 0.63 US$ L-1. The

income from the pellets in the lignocellulosic plant reduces the ethanol pro­duction cost by 0.15 US$ L-1 at 20 US$ MWh-1. As in the case of the syrup, the true price of this co-product will be dependent on its fuel properties.

The results of this comparative study led to some important conclusions regarding potential cost reductions in the lignocellulosic process, compared with the starch-based process. The overall ethanol yield in the lignocellu­losic process evaluated is 68% of the theoretical based on the available glucan and mannan in the raw material, a figure that can probably be increased. In addition, a pentose — and galactose-fermenting organism could increase the ethanol production per unit raw material without increasing the capital cost. This is especially important if the raw material is rich in pentoses, e. g. as in straw or hardwood. A reduction in enzyme loading would also be rewarding provided that the ethanol yield could be maintained. Figure 9 shows a break­down of capital costs together with energy costs for the two processes. The largest difference in costs is seen in the conversion steps and in the evapo­ration step. The pretreatment step in the lignocellulosic process represents around 0.093 US$ L-1 ethanol. This cost is attributed to both a high energy demand and to the high cost of the reactor system. This shows the need to improve pretreatment and/or enzymatic hydrolysis so that less severe pre­treatment is required. The higher cost of the SSF step compared with the fermentation step in the starch-based process is due to the longer residence

time and the lower substrate concentration in the lignocellulosic process. An increase in substrate load and productivity in the lignocellulosic process would reduce this difference. The difference in cost between the starch-based process and the lignocellulosic process in the downstream processing steps (evaporation and distillation) would also be reduced if the ethanol concentra­tion in the SSF step could be increased.

3.3