The thermo-mechanical behaviour of polymethyl methacrylate in roll-to-roll hot embossing of microfluidic channels
2020-05-28T11:16:27Z (GMT) by
The roll-to-roll (R2R) hot embossing technique is developed from the conventional hot embossing technique, which has been a predominant method for fabricating microfluidic channels on polymeric materials, such as polymethyl methacrylate (PMMA). The benefits of R2R hot embossing are the ability to take advantage of conventional hot embossing, as well as the potential for mass production. However, the research in R2R hot embossing remains limited, with very few studies to test or simulate the process of R2R hot embossing. This thesis presents a systematic analysis of the R2R hot embossing by investigating the thermo-mechanical behaviour of PMMA.
Both experimental and numerical methods have been used to understand the R2R hot embossing of PMMA-based microfluidic channels. For the experimental method, a series of R2R hot embossing trials have been conducted on a PMMA film with a custom-designed generic shim. The shim contains straight line features with a relief height of 40 µm and different line widths. The trial experiment runs at different embossing temperatures from 105 to 110 °C at every one degree, while other process parameters, such as nip pressure and web moving speed, are kept constant. The numerical method employs calibration data from tensile tests and DMA to simulate the formation of microfluidic channels in the cross-sectional area using the finite element simulation package, Abaqus/Standard. A Python script has been written to automatically generate input files for these simulations.
From experimental trials of R2R hot embossing it has demonstrated that at temperatures close to Tg, there are nearly no embossed features. The transfer rate, which is calculated by dividing the highest channel depth by the stamp height, increases with line widths. The highest transfer rate is 51.3% when the a 1-mm-wide line feature is embossed at 109 °C. The simulation method employs a parallel network model with viscoplastic components calibrated from test data ranging from 90 to 110 °C and strain rate ranging from 0.001 to 0.1/s. The calibrated data agrees well with the test data, and shows reasonable accuracy in predicting the cross-sectional profile of microfluidic channels. The Python script has been proved to be an efficient way for such numerical predictions under different process parameters. These findings have been generated to provide for guidance for microfluidic chip designers to modify shim layouts, and for process engineers to optimise the process parameters of R2R hot embossing.