Research On Design Method Of Integrated Microfluidic Chip For Magnetic Resonance Microscopy System

The research scope on single-cell has gradually expanded from the surface morphology into the internal components and structural characteristics in the field of biology and medicine.The magnetic resonance imaging,as an imaging technology that can reflect the internal structure of objects and distinguish different substances without trauma,has great application potential.Magnetic resonance microscopy is a magnetic resonance imaging technology with the spatial resolution less than one hundred microns and it is one of the development trends of magnetic resonance imaging in recent years.According to the principle of magnetic resonance imaging,the spatial resolution is limited by the size and linearity of the gradient magnetic field generated by the gradient coil.By ensuring the strength of the main magnetic field and the coil drive current,adopting a gradient coil with the diameter similar to the cell size can improve the spatial resolution.As the carrier of magnetic resonance microscopy technology,the microfluidic chip widely used in the fields of biophysics,biochemistry and medicine for cell culture and detection is a good choice.Due to magnetic resonance imaging requires the chip to be in the main magnetic field environment,all devices on the microfluidic chip should avoid any ferromagnetic material.That means the passive microfluidic chip such as mixers and valves without driving force is more suitable for magnetic resonance microscopy.The topology optimization method can quickly find the topology of the feasible structure without the designer's experience,and also can optimize the size and shape of the structure to a certain extent.Compared with other traditional structural design optimization methods,it is more helpful for the designer to propose innovative structure design.Therefore,the research content of this paper will focus on the topology optimization design of integrated microfluidic chip suitable for magnetic resonance microscopy.Due to the limitation of the main magnetic field distribution of magnetic resonance imaging,the cell solution will be mixed with the nutrient solution and diluted using a passive mixer.For the design of mixers with diffusing fluids,the fluidic topology optimization method has been very mature,including traditional passive mixers and active mixers which adopt external energy.However,for the mixing problem of cell solution that does not have diffusion property at the micro-scale,the development of fluidic topology optimization methods is slow.On one hand,the cell cannot be infinitely separable in the solution at the micro-scale,making the traditional concentration definition and objective function unsuitable for describing the mixing effect of the mixer.On the other hand,for the convection-diffusion equation which used to describe the mixing process,the standard finite element method cannot be easy to solve the pure convection problem stably.For the above problems,the concept of coarse grain concentration is used to propose a new objective function,the backward particle tracking method and mapping method used to describe the mixing process so as to achieve the topology optimization design of cell solution mix.After mixing,the cell solution needs to be transported to the imaging area through a variety of microfluidic devices.In an ideal state,the fluid topology optimization results of various microfluidic devices can be manufactured using some specific micro-nano fabrication technologies.However,two problems restrict the practical application of fluidic optimization results: One problem is that small flow channels and solid islands will be generated in the optimization result due to the needs of some microfluidic device functions,which will greatly increase the manufacturing difficulty and cost.Another problem is that the topology optimization of microfluidic devices usually uses a two-dimensional model for optimization design due to the limitation of computing resources,and then stretches the optimization results into a three-dimensional model for manufacturing.The infinite height of the flow channel in the two-dimensional hypothesis of the fluid problem does not match the three-dimensional reality,which makes some functional microfluidic devices not effective enough.Based on the mimicking open and close morphological operators,the structure length scale control method is applied to limit the minimum characteristic length of the flow channel and the solid island in the optimization result,so as to ensure the manufacturability.Inspired by the electric circuit analogy method,the maximum characteristic length of the flow channel is limited by the mimicking erosion morphology operator,thereby reducing the difference between the flow field state of the two-dimensional model and the three-dimensional model.In this thesis,the feasibility and effectiveness of the above method are proved by the design of micro Tesla valve and microfluidic splitters with equivalent outlet flowrate.After the cells enter the imaging area with the fluid,the cells need to be fixed for a period of time due to the limitation of the principle of magnetic resonance imaging.The imaging area is composed of the cell capture array and the corresponding gradient coils.The traditional design methods of cell capture unit pay more attention to the flow ratio of the bypass flow channel to the cell capture flow channel,ignoring the overall viscous dissipations of the unit,and the design results strongly depend on the designer's experience.In this thesis,the fluid topology optimization method is used to propose a cell capture unit design method with minimal overall viscous dissipations on the basis of satisfying the flow ratio of channels.Three-dimensional models are built to verify the flow field state before and after cell capture.For the problem that the micro-gradient coil designed by the stream function method has many turns,large approximation errors,and difficult to manufacture,a gradient coil topology optimization model with the resistance-assisted objective are used to obtain a micro-gradient coil with a simple structure.Numerical examples have been presented to analyze the influence of parameters in the optimization model on the final results of the cell capture unit and gradient coil.