Pilot scale trials of an emerging membrane technology, the extractive membrane bioreactor (EMB) are described. An EMB unit was installed at a chemical production facility to treat a spent caustic scrubber liquor containing monochlorobenzene (MCB). In its first configuration with 60 m(2) of membrane area the pilot unit worked well, destroying 98-99% of the organics in the MCB waste stream at a flow rate of 50 1 h(-1) with consequent evolution of chloride ion. However, mass transfer coefficients were relatively low, around 1x10(-6) m s(-1). Mass transfer experiments were undertaken at lab scale to determine the parameters limiting performance. These were found to be fluid film resistances and the build-up of biofilms on the surfaces of the membrane tubes. It was also found that steady-state biofilm thickness could be achieved through the use of fluid shear to remove biofilms. This information was used to re-configure the pilot plant using a new generation of 20 m longx50 mm diameter shell and tube membrane modules. With 60 m2 of these new modules installed, operating at Re (tube)=1070 and Re (shell)=2140, the mass transfer coefficient was improved by an order of magnitude to between 7x10(-6) and 9x10(-6) m s(-1) MCB removal of 98-99% was obtained from a flow of 500 1 h(-1), i.e. approximately 10 times the flow of the first configuration, using the same membrane area. Tube-side mass transfer coefficients were well described by existing correlations. The mass transfer coefficients on the shell side of these new modules were tested and found to be well described by Sh(s) = (0.53 - 0.58 phi) * Re-s(0.53) * Sc-.(0.33)