The size of the stereoselectivity pocket of Candida antarctica lipase B limits the range of alcohols that can be resolved with this enzyme. These steric constrains have been changed by increasing the size of the pocket by the mutation W104A. The mutated enzyme has good activity and enantioselectivity toward bulky secondary alcohols, such as 1-phenylalkanols, with alkyl chains up to eight carbon atoms. The S enantiomer was preferred in contrast to the wild-type enzyme, which has R selectivity. The magnitude of the enantioselectivity changes in an interesting way with the chain length of the alkyl moiety. It is governed by interplay between entropic and enthalpic contributions and substrates with long alkyl chains are resolved best with E values higher than 100. The enantioselectivity increases with temperature for the small substrates, but decreases for the long ones.