The cleaning efficiency of a microfiltration membrane was characterized in terms of both chemical and hydraulic cleanliness and the existing relationships between flux recovery and the amount of proteins deposited on the membrane surface left after cleaning were searched. A 0.1 mu m tubular ceramic microfiltration membrane fouled by a 3.5 wt% whey protein concentrate suspension was cleaned using sodium hydroxide. Hydraulic cleanliness was evaluated using three parameters: (i) percent flux recovery, (ii) percent irreversible removed fouling and (iii) a hydraulic cleanliness criterion representing the ratio of residual fouling resistance to initial membrane resistance. Chemical cleanliness was evaluated by means of the measurement of the amount of residual proteins deposited on the membrane surface following cleaning (expressed as mu g cm(-2) of membrane surface area). Three filtration operating parameters defining the cleaning operation were investigated: crossflow velocity, cleaning time and transmembrane pressure. Crossflow velocity had a slight beneficial influence on flux recovery but had no significant effect on the chemical cleanliness level (p > 0.05). In contrast, flux recovery was greatly affected by transmembrane pressure which had a detrimental effect on hydraulic cleanliness. With an applied transmembrane pressure during cleaning, hydraulic cleanliness was not achieved in accordance with the presence of residual proteinaceous matter representing from 0.4% to 6% of the initial amount after fouling. An optimum cleaning time of 20 min has been highlighted for the chemical cleanliness whereas hydraulic cleanliness was not significantly affected by cleaning time (p = 0.5). No linear correlation was found between the hydraulic and chemical cleanlinesses as a function of the cleaning operating conditions investigated. Percent flux recovery was insufficient in indicating accurately the chemical cleanliness. (c) 2009 Elsevier B.V. All rights reserved.