Event Abstract Back to Event Indirect Temperature Measurement in a Cell Culture Device Antti-Juhana Mäki1*, Joose Kreutzer1, Xiaohui Zhu2, Jarmo Verho1, Tomi Ryynänen1, Yong Yue2, Jukka Lekkala1 and Pasi Kallio1 1 Tampere University of Technology, Department of Automation Science and Engineering, BioMediTech, Finland 2 Xi'an Jiaotong-Liverpool University, P.R. China Motivation For biological cell culture studies, it is important that a microenvironment of cells is suitable for a successful long-term cell culturing. This environment can be mimicked more closely with a microfluidic organ-on-chip device than using conventional cell culturing methods. Furthermore, microfluidic devices are cheaper, have shorter reaction times, require less reagents and power, and allow more precise environment control than traditional cell culture methods. [1] For example, temperature can be controlled more accurately in microfluidic devices. However, implementation of a required temperature sensor in the cell culture area in these devices can be challenging, or sensors located next to cells can for instance interrupt cells growth and prevent microscopic inspection. Therefore, we developed a model that allows us to indirectly measure temperature in the area of interest. The idea is that the developed model estimates desired cell area temperature based on measured (outside) temperature. In this study, we used system identification approach to create this model. In our tests, good temperature estimations were obtained with the model. We also examined the sensitivity of the model with different liquid volumes, and all the results were acceptable for temperature estimation. As the model is not too sensitive for liquid volume changes, it can be used not only when working with different liquid volumes, but also when there is a slow evaporation of liquid during the experiment. Material and Methods The goal of this study was to develop an indirect measurement method for monitoring temperature inside 1-well structure. For this, System Identification Toolbox in MATLAB [2] was used to create a model that estimated the desired temperature inside the structure, T_Ri, based on temperature outside the structure, T_Ro. This approach was first presented in [3], where a different cell culture structure was used. For the model development in this study, a measurement system that consist of MEA1060-Inv amplifier system including a heating element, TC02 temperature controller, a custom-made sensor plate with 14 resistors used for measuring temperature (49 mm x 49 mm x 1 mm), and 1-well structure (inner diameter of 18 mm) filled with de-ionized water (used volume ranging from 0.6 ml to 1.25 ml) was used as shown in Fig. 1. During experiments, we manually altered the set-point temperature of the heating element, and performed totally six different measurements. The first three measurements with liquid volume of 1 ml were used for the model estimation and validation. Last three measurements were used to study the sensitivity of the developed model by altering used liquid volume. Using measurement data and System Identification Toolbox, we created a third-order discrete-time state-space model that estimated T_Ri based on measured T_Ro, and compared model results to measured T_Ri. Results Developed model was first compared with two validation measurements. Mean errors between measured and estimated T_Ri were 0.12°C and 0.11°C. The result from the first validation measurement is shown in Fig. 1(d). Next, three measurements with different liquid volumes (0.6 ml, 0.75 ml, and 1.25 ml) were performed to study how sensitive the model was for liquid volume changes. Mean errors were 0.23°C (0.6 ml), 0.15°C (0.75 ml), and 0.19°C (1.25 ml). This indicated that the model was suitable for temperature estimation even with relative large volume changes, thus the model could also be used e.g. when there is liquid evaporation during the experiment. Furthermore, as the maximum short-term errors were always below 0.9°C, it can be concluded that the model estimation was always in a reasonable accuracy. Conclusion New indirect temperature measurement method using a system identification process was presented. Six different measurements were used to develop and test the proposed model. The results showed that the method can be used to estimate temperature in an acceptable accuracy. We also demonstrated that temperature estimation was usable even liquid volume was changed between 0.6 ml and 1.25 ml. Furthermore, as short-term errors were relative small, the dynamics of the developed model was very similar to the dynamics of the measured cell area temperature with different heating phases.
Read full abstract
Mar 31, 2016
Myoungsoo Kang + 2
Open Access
