Numerical Simulation Of Slot Die Coating And Film Thickness Prediction For Optical Films

With the development of the semiconductor display industry,the demand for display devices is growing exponentially,and the corresponding demand for optical films is also increasing.The optical film of good quality can bring better visual experience to the display panel.In the process of optical film preparation,slot die coating is widely used at present.The flow process of the coating liquid is more complicated in the slot die coating.The study on the flow and film formation of the coating liquid during the slot die coating has been a hot topic in this field.In the process of slot die coating,film thickness is an important monitoring index after film formation.The unreasonable setting of process parameters will lead to changes in the flow field formed between the die head and the substrate.It is easy to produce product defects and uneven coating,and then affect the film quality.In the case of intermittent slot die coating,the film thickness changes in each cycle are not completely consistent under certain process parameters.Therefore,for the coating process with variable parameters and the intermittent coating process with constant parameters,the prediction of film thickness has very important research significance and practical value.To solve the above problems,the following studies are carried out in this thesis:(1)According to slot die coating theory,the flow process of slot coating fluid was analyzed,and the numerical simulation model of slot die coating was established.By simulating the flow state of coating liquid under different working conditions,the causes of normal film thickness,air entrain,film fracture and coating liquid accumulation were analyzed respectively.Then,the problem of uneven film thickness in the initial region during coating was studied.By changing the shape of the coating die head,the viscosity of the coating liquid and the contact angle between the coating liquid and the die head,the film thickness distribution in the initial region was analyzed.It lays a theoretical foundation for improving product quality.(2)The minimum film thickness and maximum substrate moving speed can only be judged by the production experience because of the nonlinear characteristics of coating fluid motion in the process of optical film preparation by slot die coating method.An Improved Beetle Antennae Search(IBAS)algorithm was proposed to optimize Extreme Learning Machines(ELM)for the slot die coating prediction.The optimized ELM model can learn the nonlinear characteristics of the system well and make accurate prediction,thus solving the traditional inaccurate empirical judgment.The feasibility and effectiveness of IBAS-ELM model are verified by the prediction of the slot die coating data.(3)In the process of intermittent slot die coating,film thickness prediction is a more concerned problem.Considering the continuous variation of film thickness during intermittent slot die coating process.The combination model of Back Propagation Neural Network and Long Short-Term Memory(BPNN-LSTM)is used to predict film thickness.Compared with the single prediction model,the combined model has higher prediction accuracy.Finally,the validity of the model is proved by intermittent slot coating prediction experiment.(4)The experimental platform of optical film preparation by slot die coating method was established.Aiming at the experimental environment of optical film coating,the control scheme of local humidity in the laboratory and the control scheme of fan filter unit are designed.The requirements of humidity and spatial cleanliness in the preparation of optical films are solved.Through the film thickness verification experiment,the consistency of the numerical simulation model established and the film thickness of the actual experiment is verified.