Ph.D Thesis

Ph.D StudentShidlovsky Zach
SubjectDeveloping a Water Treatment Unit, Based on Advance
Oxidation Processes and Biological Techniques,
While Utilizing Reporting Bacteria as
Part of a Control Loop
DepartmentDepartment of Chemical Engineering
Supervisors PROF. Yaron Paz
PROF. Sima Yaron


Water purification is challenged by the co-presence of contaminants, some of which are too toxic for bacteria or too opaque for photocatalytic processes. This challenge may be overcome by utilizing biological and photocatalytic reactors operating in sequence. Realization of this approach should be based on a steady-state model for calculating the optimal design under nominal conditions, a non-steady-state model responding to fluctuating conditions and fast measurements of toxicity and turbidity.

The representative model system used starch and chloramphenicol to describe a system that is both turbid and toxic. The model system consisted of a chemostat, which utilized Bacillus pumilus as the bio-medium, and an annular reactor as operating units. The kinetics of these reactors was studied and modeled using a genetic-algorithm. A steady-state model tested the feasibility of integrating the biological and the photocatalytic treatments, by comparing several configurations for the combined system. A toxicity sensing unit, based on the illuminating signal emitted from genetically modified Escherichia coli, was developed. A close-loop, consisted of a photocatalytic reactor and the toxicity sensing unit, tested the system's behavior under non-steady-state conditions.

The photocatalytic reactor removed 68.3% of the antibiotics using contact time of 38 min (A0=6 ppm). The estimated reaction rate, at the absence of starch, was 0.12 ppm/min. The nominal outlet concentration value is smaller than the minimal inhibitory concentration of the working bacteria in this system (2.2 ppm). It was found that non-recycled single branch configurations (whether the biological or the photocatalytic treatment is introduced as first treatment) are usually the optimal solutions. The sequence of the combined treatment was found to be dependent on the feed composition, and can be switched gradually by adjusting the recycling of the sequential system. A scaled-up and automated toxicity sensing unit was developed. The sampling period of this sensor was 17.8 min. It was demonstrated that it was possible to develop a control system that governs the operation of a photocatalytic reactor, thus enabling to maintain a level of bacterial-toxicity that facilitates to use the output stream of a photocatalytic reactor as an input stream to a bioreactor connected in series. On the other hand, there is still a large room for improvement. This improvement includes increasing both the measuring frequency and the reliability of the sensor. In order to reduce the settling time of the close-loop of the photocatalytic reactor and the toxicity unit from 5 to 2 hr, it is recommended to develop a toxicity reporting system that is able to analyze the toxicity having sampling period shorter than 10 min. In order to increase the reliability, the signal can be measured from a toxicity array unit, enabling to increase the number of repetitions without increasing the sampling period. Alternatively, the toxicity can be analyzed by spectral changes upon intake of resazurin, known to be quickly reduced in active cells, e.g., the PrestoBlue assay. Improving the toxicity sensing unit will enable the on-line operation of the pre-designed photocatalytic-biological system that adjusts the configuration of the system according to the feed characteristic.