Automation of hollow fiber membrane fabrication system and optimization of spinning parameters for polymeric membrane in gas separation

Abstract

The applications of the membrane in various types and shapes are crucial and vital especially in gas separation applications. The membrane fabrication process is complex and very delicate, but most of the existing fabrications are not fully automated which focusing only on one parameter at a time without considering the correlation between parameters involved. Therefore, a fully automated system needs to be developed with a fast and reliable fabrication procedure for determining the performance of hollow fiber membrane (HFM). The main objective of this work was to develop an automated fabrication that can be used to produce a highly producible membrane and to identify the optimum properties of HFMs for specific applications. This research was started with development of automatic fabrication machine using programmable logic controller and related sensors and actuators. Next, a mathematical model was developed and used in optimization procedure to predict the properties of HFM for gas separation application. A full factorial model was developed using a systematic experimental strategy based on the design of experiments that emphasized on tuning and simultaneous optimization of fabrication parameters of polysulfone membrane under various controllable and uncontrollable conditions. The tuning parameters for HFM fabrication were dope solution flowrate, bore fluid flowrate, air gap distance and collection speed, while CO2 flux with CO2/CH4 selectivity and diameter of membrane are the optimized membrane properties. The cause and effect relationship between all parameters were observed to obtain an optimum gas separation performance based on scanning electron microscope and gas permeation test. Scanning electron microscope and gas permeation test were also utilized in determining the performance. Finally, experimental verification was conducted to seal the reliability of the model by comparing values predicted by the model and real experimental data for new random spinning condition. Moreover, an acceptable physical morphology and structure revealed by observing and discussing the scanning electron microscope. The mathematical model was able to predict the performance of manufactured membranes based on fabrication parameters during the fabrication of hollow fiber membranes accurately with less than 5 percent error and it shows that this model is well defined and described the phenomena within range reliability for production. As a result, optimum fabrication properties of 1.27 ml/min dope flowrate, 0.33 ml/min bore flowrate, 9.1 m/min collection speed and 4 cm air gap were introduced to obtain 33 and 38.28 gallon per unit as selectivity and CO2 permeance by this study, respectively. A quality prediction model for HFM has successfully developed with 95% reliability for prediction of responses and 82.7% desirability of the performance for the indicated optimum point mentioned. This model allows membrane quality to be predicted before the fabrication stage thus prevents waste during fabrication. Additionally, variation of HFM (size and shape) can be designed and studied using this model.