Injection and flow control system for microchannels C. Fütterer,* a N. Minc, a V. Bormuth, a J.-H. Codarbox, a P. Laval, a J. Rossier b and J.-L. Viovy a a Institut Curie, 11 rue P. et M. Curie, 75005 Paris, France. E-mail: claus.fuetterer@curie.fr b ESPCI, rue Vauquelin, 75005 Paris, France Received 23rd December 2003, Accepted 16th March 2004 First published as an Advance Article on the web 11th May 2005 In spite of considerable efforts, flow control in micro-channels remains a challenge owing to the very small ratio of channel/supply-system volumes, as well as the induction of spurious flows by extremely small pressure or geometry changes. We present here an inexpensive and robust system for flow control in a microchannel system, based on a dynamic control of reservoir pressures at the end of each channel. This system allows flow equilibration with a time constant smaller than one second, and is also able to maintain stable flux from stopped flow to many ml min 21 range over several hours. It is robust to changes in ambient pressure and temperature. This system further includes a feature for sub-microliter sample injection during the experiment. We quantify flow control in elastomer and thermoplastic channels, and demonstrate the impact on one application of the system, namely the reproducible, automated separation of large DNA by electrophoresis in a selforganized magnetic bead matrix in a microchannel. 1 Introduction Microfluidics and “lab on chips” are presently the subject of an intense research activity. Numerous protocols developed for “macroscopic” handling of fluids and biological objects like cells and molecules are currently miniaturized and integrated on chips for increased throughput, better automation, and reduced reagent consumption. 1,2 In addition, microfluidics also provide access to features difficult to achieve in macroscopic systems, such as increased surface/volume ratios and stable laminar flows, even at high shear rates. However, the control of flows in microchannels as they are used in microfluidic and lab-on-chip systems raises very specific problems, as compared to the “macro” world. The total volume in the microchannel or microchannel array is very small, generally much below 1 ml, and the hydrodynamic resistance can be high, so that conventional flow control tools such as syringes or peristaltic pumps are in general poorly adapted; even the most accurate are at the limit of their resolution and reproducibility. Hence, their use generally leads to strong hysteresis and long time constants for reaching steady state, associated with the strong flow resistance of microchannels, the mechanical elasticity of the tubing and of the material in which the microchannels are embedded. Reverse flow can, in principle, be realized using such pumps by aspiration, however this leads to nucleation of bubbles. Push–pull arrange- ments of syringes or peristaltic pumps can be assembled, but even then synchronization would be problematic and unstable regimes difficult to avoid. Also, since the ratio of connection and microchannel volumes is extremely large (typically ì 1000), micro bubbles, slight changes in the connection geometry or in the environment (particularly important for elastic tubing and elasto- mer chips) can provoke dramatic spurious flows in the channels. Further, syringe pumps work at constant flow, so that rapid flow exchange is difficult and, in the case of changes in the hydro- dynamic channel resistance (accidental or wanted occlusion, change of fluid viscosity), they can lead to an uncontrolled rise in pressure causing destruction of embedded structures in the channel or even channel breakdown. Not only is the application of syringe pumps in microfluidics seriously limited, but in addition high precision syringe pumps are expensive. To overcome these problems, it has been proposed to integrate microfabricated valves and pumps into PDMS microchannel arrays. The valves disconnect the array from the outer world in order to damp uncontrolled flows, and the pumps transport the liquids on “chip”. 3,4 This elegant approach, however, complicates the microfabrication, and it may raise problems for high precision flow control, in particular when the surface properties of the PDMS (and thus the adhesion between the PDMS surfaces making the valve) are modified by the transported solutions or analytes. In addition, problems may arise in some applications, for instance in conjunction with many electromigration methods, where electrical fields have to be applied while avoiding the introduction of bubbles caused by electrolysis. Another alternative for fluid transport in microchannels is electroosmosis. 5,6 The implementation is relatively simple and cheap (however, additional electrodes have to be integrated into the system and a high voltage power source provided), but it requires a good stability, homogeneity and control of surface charge density, which is difficult to achieve with real biological samples. In addition, important applications of lab-on-a-chip technologies, such as protein separation or DNA amplification (PCR), achieve optimal performance in channels with neutral walls inhibiting electroosmosis. Finally, the rate of electroosmotic transport requires strong fields (up to several kV cm 21 ), which can in turn lead to excess heat generation. In the present article, we propose a new approach for flow control in microchannels. Our system is based on a pneumatic pressuriza- tion of reservoirs whereby the rise and decay time constants of the system are tunable by a leakage valve. In particular the decay time can now be dramatically reduced without complicating the setup. The principle and the theoretical basis are presented in Section 2. In Section 3 we propose a convenient injection system. Our realization of the system is demonstrated in Section 4. In Section 5 we characterize the system, and we present in Section 6 the improvements of an important biological application, large DNA separation, which we obtained with our approach. Finally, conclusions are drawn in Section 7. 2 Principle and theory of pressure control Our system aims at alleviating the previously discussed problems by connecting a pressurized reservoir with a controlled leak. The principle scheme of the regulation system is presented in Fig. 1, in the case of a single channel and unidirectional flow for the sake of simplicity. Using a common analogy between electricity and This journal is © The Royal Society of Chemistry 2004 MINIATURISATION FOR CHEMISTRY, BIOLOGY & BIOENGINEERING DOI: 10.1039/b316729a 351 Lab Chip , 2004, 4 , 351–356