General Process. Figure 1 shows a schematic flowsheet of the enzymatic biomass — to-ethanol process. Lignocellulosic biomass consists mainly of cellulose, hemicellulose and lignin. In the production of ethanol the sugars in the hemicellulose and cellulose are converted to ethanol. As a by-product, lignin is formed which can be used as a solid fuel.

In the pretreatment stage the hemicellulose is solubilised into sugars. The cellulose structure is opened up and becomes more accessible to enzymes. Prior to pretreatment, the raw material must be chipped. There are many types of pretreatment methods and steaming has been recognised as one of the most effective for lignocellulosics (12, 23-27). Although companies such as Stake Technology (Stake Technology Ltd, Canada) currently sell equipment for the steam pretreatment of wood, the technology is still considered to be in the development stage (28). Furthermore, the use of catalysts to enhance the cellulose accessibility and the yield of hemicellulose sugars of a number of feedstocks has been recognised (14, 15, 29, 30). If catalysts such as sulphuric acid or sulphur dioxide are used, the corrosive effects of long-term usage on the equipment must be considered. The effect on the quality of the lignin by-product, due to possible sulphonation, must also be taken into account. If a catalyst is employed, the raw material must usually be pre-steamed to expel trapped air from the pores in the material. This can be performed using low-pressure steam.

In the hydrolysis step, the cellulose is hydrolysed by enzymes into glucose. To reduce the cost of concentrating the end-product, a high sugar concentration is desirable, which means a high dry matter content of the cellulosic material entering the hydrolysis stage. However, a high dry matter content in the reactor makes stirring difficult which reduces mass transfer. Fed-batch hydrolysis, i. e. gradual feeding of the material into the hydrolysis vessel, is an option that may be attractive.

Large quantities of enzymes are required for the hydrolysis to be fast due to the low specific activity of the cellulase system. The cellulose-degrading enzymes should preferably be produced on the same raw material as that used for ethanol production. A minor part, about 5%, of the pretreated material is used as a carbon source for some cellulase-producing organism, such as Trichoderma reesei (31-33). Unfortunately, this fungus does not excrete sufficient amounts of one of the enzyme components, (i — glucosidase. As the consumption of P-glucosidase is much lower than that of the other cellulases, this enzyme may either be produced by another organism such as Aspergillus phoenicis (34, 35) or be purchased to avoid increased complexity in the process. Extensive research is necessary to find more effective enzymes which will reduce the hydrolysis time. It is also essential to find inexpensive substrates for enzyme production.

After hydrolysis, the solid residue is separated from the hydrolysate and washed. The residue consists mainly of lignin which can be dried and used as a high-quality solid fuel. Separation can be performed either through filtration in a filter press (36) or by using a decanting centrifuge (18). The sugars in the hydrolysate are then fermented to ethanol.

Hexoses, mainly glucose, mannose and galactose, are generally fermented by Saccharomyces cerevisiae in continuous fermenters with yeast recycling. The technology and equipment required for glucose fermentation in a biomass-to-ethanol facility are believed to be similar to those used in the sugar — and starch-to-ethanol processes or in the production of ethanol from spent sulphite liquors.

The pentose fermentation step is essentially still at the laboratory stage, although significant advances have been made in understanding the mechanism and in the modification of pentose-fermenting micro-organisms (37-39). None of the wild-type, adapted or engineered strains, have to our knowledge, been able to routinely ferment the pentose-rich, water-soluble stream obtained from steam-pretreated hardwoods or softwoods.

An alternative to separate hydrolysis and fermentation is simultaneous saccharification and fermentation (SSF), i. e. operation of the hydrolysis step in combination with the fermentation step. This operation reduces the end-product inhibition resulting from glucose and cellobiose build-up by continuously converting the glucose to ethanol. Although ethanol and other fermentation products may decrease the activity of individual cellulase enzymes, this inhibition is much weaker than the inhibition caused by equivalent glucose or cellobiose concentrations (40-42). The conditions for running SSF are a compromise between the optimal conditions for cellulose hydrolysis and those for fermentation. It is probably necessary to produce new yeast and enzymes for each SSF batch because of the difficulty in separating the cells and enzymes from the unhydrolysed solid residue (43).

The ethanol in the fermenter is rather diluted and must be recovered by distillation, which is considered to be a mature technology. The experiences from production of alcoholic beverages can be used to design and construct distillation units for recovery of fuel ethanol.

Four major by-product and liquid waste streams are expected from the enzymatic hydrolysis process: a waste stream from the pretreatment unit; spent fungal mycelia from the enzymatic production unit; unhydrolysed cellulose residue containing spent enzyme and lignin; and stillage from the distillation column which contains solubilised non-volatiles from the raw material, carbohydrate- and lignin — degradation products as well as by-products from the fermentation stage. The stillage waste stream is the largest in volume. It has high Biological and Chemical Oxygen Demands, (BOD7 and COD, respectively), and a low pH. It often contains high levels of colouring compounds and has been noted for its high corrosiveness (44). Biological treatment of the stillage stream generally involves anaerobic treatment prior to aerobic treatment and a tertiary treatment to remove colouration. Recycling of liquid streams would minimise the fresh water requirement and lower the amount of waste water produced. However, this recycling leads to an increase in the concentration of various substances in the hydrolysis and fermentation steps (45). To avoid the accumulation of non-volatile inhibitors in the process, the stillage stream could be evaporated prior to recirculation (36). The volatile fraction from the evaporation step is then recycled and the non-volatile residue incinerated. This is described in more detail in the sections ” Recycling of process streams” and ’’Process integration”.

Bench-Scale Unit. In scaling up, the rate-limiting steps must be identified. The rates of all other process steps must be carefully assessed to determine whether they may decrease on scale-up and become new bottlenecks on a larger scale. Since microbial processes are complex, it is important to evaluate how various recycling scenarios will affect the microbial environment in continuous fermentation in a fully integrated process. The final investigations should preferably be performed on a pilot scale, but it is often more cost-effective to evaluate various options and operating conditions first on a bench scale.

At Lund University, a bench-scale unit was set up in 1995 for the development of a process for ethanol production from lignocellulosics based on enzymatic hydrolysis (36). Figure 2 shows a schematic flowsheet of the unit. The different unit operations are not physically connected so the material is passed manually from one step to the next. This makes the unit very suitable for studying various process configurations and also the recycling of process streams.

The pretreatment unit consists of a 10-litre pressure vessel (corresponding to approximately 1 kg of dry wood chips) and a flash tank to collect the pretreated material. An electric steam boiler supplies steam up to a maximum pressure of 30 bar, corresponding to a saturated steam temperature of 235°C. The temperature, pressure, and hold-up time in the reactor are monitored and controlled via a computer.

Figure 1. Production of ethanol using enzymatic hydrolysis.

A laboratory filter press is used to filter and wash the pretreated material. The material is stored in a 100-L stirred tank, slurried with water, then pumped through a filter cloth at a maximum pressure of 10 bar. After filtration, the filter cake is dewatered by applying a pressure of 15-17 bar and then washed under pressure with water. The unit can be used to determine scale-up design data, such as the filter-cake resistance.

The hydrolysis unit consists of two tanks with working volumes of 20 litres and 40 litres. The tanks are stirred by propeller-type agitators which have been proven to give efficient mixing of pretreated material up to 10% dry matter (DM). Both vessels are equipped with a water jacket to maintain a constant temperature in the range 30- 90°C. The hydrolysis residue can be filtered in the filter press described above.

The enzyme production and fermentation units consist of fermenters with total volumes of 22 litre and 16 litre, respectively. The fermenters are equipped with sensors for control and data sampling of temperature, pH, dissolved oxygen and anti­foam addition. The outlet gas is analysed for oxygen and carbon dioxide.

The distillation unit consists of a 1.5 m packed column, corresponding to about 15-20 ideal stages, a jacket reboiler and a condenser. The distillation unit can be operated batchwise or continuously. The reflux ratio can be varied between 0.1 and

20.

The evaporator is designed for two different purposes: flash-sterilisation of fermentation broth, or concentration of the non-volatile components in liquids. Heat transfer is achieved via plate heat exchangers using live steam and cold water as heating and cooling media, respectively. The maximum capacity of the evaporator is 16 kg/h evaporated water. The evaporator can be used for the simulation of condensate withdrawal in a multiple-effect evaporator.

Experimental Runs. In the introductory experimental runs, willow, a fast growing energy crop, was used as substrate. The willow was pre-steamed with low-pressure steam (2 bar) prior to impregnation with gaseous sulphur dioxide in a plastic bag. The amount of sulphur dioxide added was approximately 4% of the DM. The impregnated material was then subjected to steam pretreatment at 205-210°C for 5 minutes. The pretreated material was slurried with water to approximately 5% DM, and filtered. After filtration, the filter cake was dewatered to approximately 40% DM.

Hydrolysis was performed using the fibrous material and the filtrate at a dry matter content of 5%, supplemented with Celluclast 2L and Novozym 188 (Novo AS, Denmark), corresponding to 15 FPU/g substrate. Commercial enzyme preparations were used, as the main purpose of the study was to investigate the influence of the recycling of waste streams. The temperature was maintained at 40°C and the pH at 4.8 during the entire hydrolysis time of 90 hours.

Fermentation of the willow hydrolysate was carried out in the 22-litre fermenter containing 16 litres medium at 30°C, pH 5.5, at a stirring speed of 300 rpm. The hydrolysate was supplemented with a rich medium, and inoculated with compressed baker’s yeast, S. cerevisiae, to a final cell concentration (DM) of 6 g/L.

The fermentation broth was stripped in the evaporation unit, prior to distillation, to avoid problems associated with suspended particles in the distillation column. The evaporated fraction was then distilled. The COD, the BODy and the fermentability of the stillage and the evaporation residue were determined.

The enzyme production was run in a parallel study. Enzyme production was performed in the 16-litre fermenter. A modified Mandels medium was used (46), in which the yeast extract and proteose peptone were replaced by dried yeast. A separate batch of willow, pretreated under the same conditions as described in the pretreatment section, was used as substrate for the enzyme production.

Some of the results from the experimental study in the bench-scale unit are described in the section ’’Recycling of process streams”. A more detailed description of the equipment and the experimental procedure can be found in the original publication (36). Similar experiments have also been performed using a mixture of softwoods (47).