The effects of ultrasound and heat pretreatments on ethanol yields from cassava chips were investigated. Cassava slurries were sonicated for 10 and 30 s at the amplitudes of 80, 160, and 320 mu m(pp) (peak to peak amplitude in mu m) corresponding to low, medium. and high power levels, respectively. The sonicated and non-sonicated (control) samples were then subjected to simultaneous liquefaction-saccharification and ethanol fermentation. Cassava starch-to-ethanol conversion efficiencies showed that higher ethanol yields were directly related to sonication times, but not to power levels. Significantly higher ethanol yields were observed only for sonicated samples at the high power level. The ethanol yield from the sonicated sample was 2.7-fold higher than yield from the control sample. Starch-to-ethanol conversion rates from sonicated cassava chips were also significantly higher; the fermentation time could be reduced by nearly 24 h for sonicated samples to achieve the same ethanol yield as control samples. Thus, ultrasound pretreatment enhanced both the overall ethanol yield and fermentation rate. When compared to heat-treated samples, the sonicated samples produced nearly 29% more ethanol yield. Combined heat and ultrasound treatment had no significant effect on overall ethanol yields from cassava chips. Ultrasound is also preferable to heat pretreatment because of lower energy requirements, as indicated by energy balances. Integration of ultrasound application in cassava-based ethanol plants can significantly improve ethanol yields and reduce the overall production costs. (C) 2009 Elsevier Ltd. All rights reserved.