The Study On Monte Carlo Simulation In Microfluidic Chip

Microfluidics is a new technology developing very fast since the 90 years last century. As a new crossed science, it combines analytical chemistry, MEMS, computer, electronics, material, biology, medicine and so on. Microfluidics is an important part of the micro-nano technology, and also a technology platform for researching system biology. Microfluidics has many important application, such as environment detect, clinic detect drug filter, finding a new gene, DNA sequencing, disease diagnosis. Although, research on the microfluidic is started very late in China, it developed very rapidly with the help of the sustentation fund of the nation. National Nature Fund and the 863 Plan have launched a lot of money into BioMEMS research.In this background, our laboratory has started national nature fund task"DNA multi-field dynamics research and application in biology separate chip(No.10572053)", 863 Plan task"biology micro- and nano-fluidic system's design software exploitation(No.2006AA04Z305)", University doctor fund task"DNA flow theoretical research and application in the microfluidic chip(No.200401830 57)".This paper is part of research content of these three tasks. In our work, we facus on the application of Monte Carlo method used in the microfluidic devices. First, we study the Monte Carlo simulation of chromatin fiber, and second, The DSMC investigation of the micro-scale flow in the microfluidic devices.First,we used Monte Carlo method simulating the chromatin fiber. The simulation spends a large part of time in calculating the Debye-Huckel potential, in order to save computation time, we used a tabulation of the double integral Debye-Huckel equation. Then during the simulation, a liner interpolation was used to abtain the values of Debye-Huckel equation corresponding to random four parameter values. We obtained the chromatin fibber structure from simulation. Equilibrium ensembles of 60-nucleosome chains at physiological ionic strength were generated by a Metropolis-Monte Carlo algorithm. For a DNA linked at the nucleosome stem and a nucleosome repeat of 200 bp, the simulated fiber diameter of 30.17 nm and the mass density of 6.34 nucleosomes per 11 nm fiber length are in excellent agreement with experimental values from the literature. We studied how the solution concentration affects the chromatin fiber structure. We find, when the solution concentration is between 0.001mol/L and 0.01mol/L, it is hard to form 30-nanometer chromatin fiber structure. When the solution concentration is between 0.01mol/L and 0.15mol/L, the nucleosomes form compact chromatin fiber structure. We also studied how the twist angle between adjacent nucleosomes (β) affect the chromatin fiber structure. We find, when the twist angle is increased, the mass density increased a little under a limit value of 6.2 nucleosomes per 11 nm, the mean tilt angle between the nucleosomes axis and the fiber axis is also increased when the twist angle is increased.Second, we used DSMC method studied the micro-scale flow in the microfluidic devices. We developed DSMC code for simulation micro-scale flow. The DSMC code we used was verified by comparing the results to analytical solutions. We studied the three factors which can affect the accuracy of the simulation results, the number of cells in the computational domain, the number of particles per cell, which is referred as the mesh size, and the number of real molecules represented by each simulated particle, which is also called"scaling factor". It can be concluded that 20 simulated particles per cell as well as the corresponding scaling factor is appropriate to achieve the accuracy of the numerical results, which is also consistent to the suggest of Bird. We find that the cell dimension can much greater than the local mean free paths in directions where the gradients of the flow properties are small, and while the time step could also been enlarged accordingly. The contracted channel is commonly used in microfluidic chip, so we analyzed the flows in contracted channel. We studied the ratio of the extent of the contraction witch can affect the flows, when the ratio is decreased, the velocity of flows is increased wholely, and the contour curve of the pressure distribution of the expanse part of the channel is more remarkable. We also studied how the differences of the wall temperature affect the flows. When the wall temperature is increased, the gas was heated more efficiently, so the temperature distribution is more uniform.