The outcome of an economic study shows that the most important factor for economic bioethanol production is maximum ethanol output (von Sivers & Zacchi, 1996). A possibility to increase the ethanol output would be the recycling of effluents with low ethanol concentration, e. g. the stillage from the distillation, which contains about 1% ethanol (Cortella & Da Porto, 2003) or low concentration effluents from membrane separation steps via the steam explosion reactor. In this case, the added water would be replaced by the effluent to be recycled. During the steam treatment, vapour-liquid equilibrium of the ethanol-water system will be reached. Due to the fact that ethanol is more volatile than water, the concentration of ethanol in the vapour phase will be much higher than in the liquid phase. The vapour-liquid equilibrium of the ethanol-water system at 1.5 MPa is shown in Fig. 6.

When the reactor is vented, the exploded biomass is separated from the vapour phase in a cyclonic separator. In the separator secondary vapour is also produced by evaporation cooling of the wet biomass. The vapour phase has to be condensed by cooling at the separator outlet to recapture the ethanol. The collected condensate can be added to the feed of the distillation column.

In a first series of experiments on the recycling of ethanol-containing effluent, the added water in the feed to the steam explosion reactor was replaced by a solution containing 10% (w/w) ethanol. Analyses of the pretreated wet straw are shown in Table 2. The samples were taken from the treated straw heap in the separator immediately after the explosion step and transferred into a gastight bottle. With the exception of ethanol no significant differences were found when 10% ethanol (w/w) solution was used. The ethanol content of 31.6 g/kg feed straw (d. b.) in the treated straw from the experiment with the addition of 10% ethanol solution (w/w) is equivalent to 31.6% of the added ethanol; the remaining 68.4% is expected to be in the condensate. It was not possible to verify this due to limitations in the drainage of such small amounts of condensate from the installed regenerative cooler.

Treated straw samples taken from the separator about five minutes after the explosion step showed a significantly lower ethanol content. The average ethanol content in these samples was 13.5 g/kg of feed straw (d. b.), whereas the concentrations of the other components were

very much the same. This can be explained by the evaporation of ethanol during the cooling of the treated straw. For example, the recycling of a 1% ethanol (w/w) solution would result in a condensate with about 5% ethanol (w/w) considering also the dilution of the liquid phase in the reactor by condensation of steam.

However, recycling of low ethanol concentration effluents could be limited by inhibitors contained in the effluent. Further tests with real effluents are therefore required.