Indirect Temperature Measurement and Control Method for Cell Culture Devices

Abstract

Microfluidic devices are promising tools with which to create an environment that mimics a cell’s natural microenvironment more closely than traditional macroscopic cell culture approaches. In these devices, temperature is one of the most important environmental factors to monitor and control. However, direct temperature measurement at the cell area can disturb cell growth and potentially prevent optical monitoring, and is typically difficult to implement. On the other hand, indirect measurement could overcome these challenges. Therefore, using the system identification method, we have developed models to estimate the cell area temperature from external measurements without interfering cells. In order to validate the proposed models, we performed large sets of experiments. The results show that the models are able to catch the dynamics of temperature in a desired area with a high level of accuracy, which means that indirect temperature measurement using the model can be implemented in the future cell culture studies. The usefulness of the model is also demonstrated by simulations that use estimated temperature as a feedback signal in a closed-loop system. We also present tuning of a model-based controller and a noise study, which shows that the tuned controller is robust for typical ambient room temperature variations. Note to Practitioners —In this paper, we tackle the problem related to temperature measurement in microfluidic devices, especially but not only concerning cell culture environments. Even though it would be desirable to place a temperature sensor as close as possible to the location of interest, practical limits usually prevent this; for instance, limited space and requirements for optical monitoring. To overcome these problems in microfluidic devices, we present a novel indirect temperature measurement approach using the system identification method. The idea is to create a model that estimates temperature on the area of interest using measured outside temperature. Because it is required to measure both model input and output signals for the model development, we first fabricated a temperature sensor plate, combined it with our heating system, and measured required temperatures on several experiments. Then, we developed third-order discrete state-space models using measured temperatures and System Identification Toolbox in MATLAB. Model performances were examined and compared with measurements. Furthermore, we created a closed-loop Simulink (from MATLAB) model, and showed how desired temperature could be controlled using only measured outside temperature and the developed model. In the future research, we will implement the designed closed-loop system to our cell culture system to precisely control temperature in the cell area.