Design and application of calorimeters for monitoring biological processes in stirred tank bioreactors

Developing and improving online monitoring techniques for processes is always a matter of interest for industrial and research applications. Independent of kind and complexity of the reaction, measuring the heat generation (calorimetry) is a universal tool for process monitoring. Therefore, two new calorimetric measurement techniques (chip calorimeter and reactor calorimeter) for process monitoring in stirred tank reactors were developed, validated and applied to several biological systems. The validation was done by exhaust gas analysis and the application of the oxycaloric equivalent. By assuming the rule of Thornton, the enthalpy of combustion of organic compounds is linearly correlated to the relative degree of reduction (Thornton's rule: -115 kJ per mole accepted electrons). Since four moles of electrons are absorbed by one mole oxygen, a constant value for the oxygen related enthalpy of about -460 kJ mol-1 results. This is denoted as the oxycaloric equivalent. Consequently, if the enthalpy of the anabolic processes is neglected, the metabolic heat generation can be calculated based on the oxygen transfer rate. As a model organism for validating the calorimetric measurements, the genetic modified Escherichia coli strain VH33 was chosen. The genetic sequence of the phosphotransferase system (PTS) in the bacterial chromosome of this strain was deleted. The galactose permease (GalP), as an alternative uptake mechanism, was amplified on a genetic level to allow the passage of glucose across the plasma membrane. Therefore, this strain had a strictly reduced overflow metabolism and was the optimal organism for validating the calorimetric measurements of fermentations by exhaust gas analysis and the oxycaloric equivalent. One possibility for calorimetric monitoring of fermentations in stirred tank reactors is the chip calorimeter, developed by the TU Bergakademie Freiberg, which can be connected to every reactor system as long as a septum is available. The measurements of the device were validated by exhaust gas analysis (oxycaloric equivalent) of E. coli VH33 cultures. The measured results of the chip calorimeter were very precise (0.001 W L-1). However, the upper limit of the device was a heat generation of 1 W L-1 due to oxygen depletions in the measurement chamber at higher biomass concentrations. Therefore, the chip calorimeter was enhanced by an automatic sample dilution system, which mixed the sample with oxygen containing NaCl solution before measuring in the chamber. With this modification, the upper limit of the measurable heat generation was increased up to 4 W L-1. An alternative for measuring the heat generation of fermentations is to modify or design a whole stirred tank reactor as a calorimeter – the so called reactor calorimeter. This measurement principle based on the thermodynamic balancing of the cooling / heating water cycle of the reactor system. Consequently, all heat sources and sinks of the system should be known. In general, there are two ways for designing reactor calorimeters: 1) measuring the mass flow of the cooling liquid and the temperature difference between cooling cycle inlet and outlet; 2) determination of the heat transfer coefficient and the temperature difference between reactor inside and cooling cycle. Both methods were realized by modifying a 50 L stirred tank reactor and were validated and compared with each other by cultivating E. coli VH33. Ultimately, it could be shown that the second way, based on the determination of the heat transfer coefficient has a higher signal quality and should be chosen for all following applications. Finally, the modified reactor was used to investigate the potential of the calorimetric signal for online detection of product formation during fermentations. It could be demonstrated by determining the ratio between heat generation and oxygen consumption, that lysine formation of Corynebacterium glutamicum DM1730 and itaconic acid formation of Ustilago maydis can be online detected during fermentations in stirred tank reactors