摘要： Separation and sorting of particles (cells, beads, and droplets) is critical in a variety of biomedical applications including early disease diagnostics, therapeutics, cell tracking, and clinical research. Currently, standard microfluidic methods use 2D-chips fabricated using PDMS based soft lithography or embossing/etching materials such as Topas or glass. Often hydrodynamic flow focussing is used to spatially confine the sample flow to particular areas on the chip and so facilitate target species detection; however, 2D systems have limitations when it is necessary to make fine, micrometre scale, adjustments to flow profiles and/or the spatial position of the focused sample stream so as to match up with elements of external bulky equipment such as the laser or light paths of a spectrometer or internal fields used in electrical or acoustic actuation. These limitations can be overcome by using a 3D flow focussing method which can dynamically adjust the flow patterns within the sample chip (Figure 1a, left). For complex systems containing multiple sensor and actuation stages that are best made using hybrid construction of plug-and-play subsystems, a critical aspect of the overall functioning of the system is knowing at what points in time individual particles (e.g. cells, beads) are in the vicinity of different functional parts of the device. For high throughput and high particle densities, this becomes more difficult the larger the chip and the faster the flow. To address this, here we used a strategy of integrating several optical sensors into a hybrid device made from 3D printed and conventional soft lithography PDMS parts (Figure 1a, right), and so provide this timing data. Using these, it has then been possible to develop an automated flow- focusing and sorting microfluidic system with the flexibility and stability required for the sorting and refinement of labelled cells and beads (Figure 1b).