The membrane bioreactor (MBR) process is a promising technology that provides a cost-effective alternative to conventional processes in water/wastewater treatment. The present research focuses on the application the MBR technology for the purification of groundwater contaminated with petroleum hydrocarbons exemplified by gasoline. These organic contaminants include several alkanes, iso-alkanes, and aromatics exemplified by benzene, toluene, ethylbenzene and xylenes, and oxygenated additives such as methyl-tert-butyl ether (MTBE). The MBR system employed a powder activated carbon (PAC) slurry with microorganisms acclimated to petroleum hydrocarbons. The process was operated as a closed system to prevent the volatilization of gasoline contaminants from the aqueous phase into the atmosphere.

A predictive model was employed for performance forecasting and design of the MBR process in the above application under a variety of operating and process conditions. The total organic carbon (TOC) was used as a lumped parameter for representing the total concentrations of petroleum hydrocarbons. The model involved the following essential features: (a) biological reaction in bulk liquid solution, (b) film transfer from bulk liquid phase to the biofilm-liquid interface, (c) diffusion with biological reaction inside biofilm, (d) adsorption equilibrium occurring at the biofilm-carbon interface, and (e) diffusion transport of contaminant within the activated carbon particles. The biological reaction in the bulk liquid phase as in the biofilm was represented by Monod kinetics, and the model assumed that biodegradation occurred in the bulk fluid phase as well as in the biofilm immobilized on the PAC particle. This modeling approach could be employed in upscaling the process from laboratory-scale to pilot-scale and eventually to full-scale using the techniques of dimensional analysis and similitude. The upscaling technique would involve dimensionless groups developed specifically for the adsorption and biodegradation sub-processes associated with the MBR process. The modeling and upscaling technique described above would serve as an efficient, time-saving, and cost-effective approach to design the MBR process.

The model parameters were determined from independent laboratory-scale experiments and correlation technique. The adsorption equilibrium and rate parameters for the model were determined by independent batch reactor studies. The biokinetic parameters for the degradation of gasoline components were determined from chemostat studies using continuous stirred tank reactors (CSTR). The laboratory-scale MBR experiments involved experiments in the presence and absence of activated carbon slurry for comparison purposes, and the organic contaminant removals as well as the permeate fluxes were measured in each case. These experiments established the effectiveness of the MBR technology in achieving nearly complete removals of organic contaminants under optimal conditions of PAC dosages and biomass concentrations. The results of these experiments also provided the means for model verification under different operating conditions. Model simulation studies were performed with respect to certain parameters including those representing liquid film mass transfer, biofilm transport, biological degradation kinetics, influent concentration, and reactor hydraulic retention time. These studies provided a qualitative evaluation of the parameters influencing the process dynamics under different conditions.

Key words: Membrane bioreactor, petroleum hydrocarbons, membrane fouling, modeling, biofilm transport, biodegradation kinetics, upscaling