1 Department of Chemical and Biochemical Engineering, Technical University of Denmark2 Center for Microbial Biotechnology, Department of Systems Biology, Technical University of Denmark3 Department of Systems Biology, Technical University of Denmark4 Novozymes A/S
Ecient fermentations at industrial scale are usually preceded by an enormous amount of research work aimed at optimizing the productivity of the strain in question. Before that, the question of the selection of the correct strain already accounts for a substantial amount of work. Today, most screening is done in microtiterplates which allow for cultivations similar to those in shake asks, however, due to the much smaller volume, microtiterplates are much more streamlined for parallel, machine-controlled operation. Out of these cultivations, a number of strains are selected for further investigation, which typically means performing cultivations in larger and larger scales. As the size of the reactor increases from shake ask to bench-scale reactors to pilot-plant installations the number of strains decreases until only one strain is left at production scale which hopefully is the ideal strain. However, precisely this scaling-up can give problems, as strains may behave dierently in a shake ask than in a production scale reactor. There is therefore a need for a small-scale production platform which can oer more reliable upscaling; Or in other words a platform which is better at mimicking full-scale operation is needed. Apart from the fermentation industry, research also depends on well-controlled cultivations with tight measurement and control in order to obtain meaningful data about the strain metabolism. Microbioreactors have the potential to be the platform needed to fulll the above requirements: The working volumes are relatively small, typically <1 mL, and they can be operated in dierent operating conditions such as batch or continuous cultivations. Additionally, their small size oers a number of possibilities: Under the presence of good mixing, one can assume the contents of a microbioreactor to be free of gradients (e.g. nutrients, oxygen) which allows for a precise determination of the state of the cultivation. Additionally, the large surface to volume ratio opens up the possibility for quick changes in temperature, so that e.g. the in uence of step changes on the metabolism can be investigated. The advance of miniature online measuring techniques makes it possible to measure at least the basic culture variables such as dissolved oxygen (DO), cell density (OD) and pH continuously and without disturbing the cultivation. Online measurements are at this scale very susceptible to the presence of bubbles|as is a microbioreactor itself as already small bubbles can disturb the ow in microchannels. Bubble-less aeration through a membrane elegantly solves both problems and also separates the broth from external in uences. The microbioreactor developed here was designed to fulll the above requirements; Additionally, much focus was put on the single-use aspect. This includes both being cheap in fabrication and in operation, and also requires the reactor to be sterilizable by industrial methods. It consists entirely of polydimethylsiloxane (PDMS) and contains two optical sensor spots for the measurement of DO and pH as well as a micro-stirrer for agitation of the broth. It has previsions for the measurement of cell density (by means of optical density measurent) as well as membrane aeration. Both temperature and pH can be controlled online and automatically. The device has outer dimensions of 14 mm diameter and 4.2 mm height. The reactor chamber is a cylinder with 8 mm diameter and 2 mm height resulting in a culture volume of 100 L. The uidic connections are done by piercing the reactor side walls with needles|the PDMS will tightly enclose the needle to prevent leakage. The reactor chamber is sealed with a semi-permeable membrane (thickness approximately 80 m), through which the gases can diuse. Both oxygen and o-gases are exchanged this way. Additionally, pH can be controlled by the addition of CO2 or NH3 to the aeration gas ow to lower or increase pH respectively. The density of the culture broth is measured by a transmittance measurement|light is shone through 0.5 mm of culture broth, and the intensity of the transmitted light is measured. This gives an indication of the amount of cells in the broth. Both DO and pH are measured with uorescent sensor spots: Oscillating light is shone onto the sensor spots which in turn emit oscillating uorescent light with a certain phase shift respective to the exciting light. This phase shift relates to the DO or pH of the broth, respectively. Mixing is solved by means of a small magnetic stirrer bar which, contrarily to what is seen in other microbioreactor solutions, rotates freely within the reactor. Experiments had shown that a stirrer bar rotating in the middle of the reactor will only force the broth into a swirling motion where the outer edges of the reactor do not have enough updraft anymore. The freely spinning stirrer bar however will hit the wall and ricochet chaotically into the reactor chamber again. Thus, over time, all of the reactor oor will be covered which prevents the formation of dead zones. Temperature is controlled by means of an external (and thus re-usable) heating plate which contains both a temperature sensor and a resistance heating wire. As the oor of the microbioreactor only consists of a membrane which oers virtually no heat resistance, this allows for a precise control of the broth temperature. In order to provide benchmarking data to be able to evaluate the reactor performance, batch cultivations were done in both shake asks and bench-scale reactors. Finally, corresponding cultivations were performed in the microbioreactor. Additionally, as an entirely theoretical case study of something completely new, the application of the topology optimization methodology on microbioreactors and the resulting gains in productivity was studied.
Main Research Area:
Gernaey, Krist, Eliasson Lantz, Anna, Stocks, Stuart