Rodrigo Martinez-Duarte1,*, # and Marc J. Madou2,3
1 Microsystems Laboratory (LMIS4), École Polytechnique Fédérale de Lausanne, Switzerland
2 Department of Mechanical & Aerospace Engineering, University of California, Irvine, United States of America
3 Ulsan National Institute for Science and Technology, World Class University Program, South Korea
* For more information contact email@example.com # Previously at University of California, Irvine
Dielectrophoresis (DEP) is a viable tool for the selective manipulation of a variety of particles for different applications including cell and microorganism sorting, tissue engineering, food safety and drug development (see Dielectrophoresis note in this same webpage). Although most of the work has been done using planar metal electrodes, other techniques, i.e. insulator-based and contactless DEP (see correspondent articles in this webpage), have emerged to overcome some of the drawbacks of using metal electrodes, namely electrode fouling , fabrication and materials cost. However, DEP still greatly suffers from the lack of high processing throughput which has hindered DEP from becoming more widely used in clinical applications. For example, high throughput is desired in cell sorting to provide a less expensive alternative to fluorescence-activated cell sorting (FACS) or magnetically-activated cell sorting (MACS); in food safety to quickly identify threats, i.e. E.coli and salmonella, with very high sensitivity; in hospitals to quickly pinpoint the cause of an infection; in drug development to enrich very large populations of particles of interest. High throughput can be achieved by increasing the cross-section of the microfluidic channel. It is simple, the more you can put in the system the sooner you will process a given sample volume. Taller, rather than wider, channels are desired to maintain a small footprint of the DEP chips and minimize costs. The continuity equation establishes that the mass flow rate Q in a micro-channel equals the product of the flow velocity υ and the cross-section area A of the channel (Q = υ·A). Therefore, the flow rate can be increased by increasing only the cross-section area but not the flow velocity. Since the hydrodynamic drag force acting upon particles flowing in a micro channel depends on the flow velocity, the maintenance of low flow velocity in the channel leads to a hydrodynamic force that can be easily overcome by a DEP force created using practical voltages and electrode gaps. Once the channel is made taller, one needs to induce an electric field throughout the whole volume of the channel to guarantee all particles flowing through get influenced by a DEP force. Indeed, the use of 3D electrodes, as tall as the channel and contained inside the channel, enables the addressing of all particles flowing in the channel. This is in contrast to planar electrodes positioned on the channel floor, ceiling or walls which only address those particles flowing close to the channel surfaces. Another constraint in the fabrication of the electrodes comes from the fact that to induce electric field gradients suitable for DEP using practical voltage levels, say tens of volts, the electrodes must be very close together. The gap between electrodes must also be uniform across the height of the electrodes and thus a 3D structure with vertical walls is desired.
Here we present 3D carbon-electrode DEP as an alternative that could lead to the desired throughput levels. Its main advantages are the properties of carbon itself and the fabrication process which allows for the low cost fabrication of closely spaced arrays of electrodes featuring heights above 100 µm. Glass-like carbon electrodes are more electrochemically stable than metal ones and thus afford the application of higher voltages, hence a stronger DEP force, across the sample without electrolyzing it . In fact glass-like carbon, also known as glassy carbon, is a preferred material among electrochemists due to its remarkable stability [3,4]. Glass-like carbon also has excellent biocompatibility and has been demonstrated both as an implantable material  and as substratum for cell culture . Furthermore, glass-like carbon is chemically very inert in almost all solvents/electrolytes. Remarkably, it withstands attack from strong acids such as nitric, sulfuric, hydrofluoric or chromic and other corrosive agents such as bromine [6,7].
Glass-like carbon electrodes are derived via pyrolysis, heating to high temperatures in an inert atmosphere, of a previously shaped organic polymer in a process known as Carbon MEMS (C-MEMS) . Carbonizable polymers are widely available and high-quality ones are typically much less expensive than metals such as gold and platinum used in thin film and electroplating metal electrode fabrication. The polymer can be shaped using any suitable low cost technique such as photolithography, machining, molding and embossing. No expensive and complex equipment such as metal evaporators or metal sputter coaters is required. The use of SU-8 photolithography in particular allows for the patterning of very narrow gaps between high-aspect-ratio structures featuring vertical walls.
Not surprisingly, 3D carbon electrodes are not the perfect solution for every application. A potential disadvantage of carbon-DEP is the electrical resistivity of glass-like carbon (~1 X 10-4 Ω·m ) which is four orders of magnitude greater than gold. The voltage loss that develops from the Ohmic resistance in the narrow leads connecting the base of the electrodes and the function generator makes it necessary to use higher voltage levels than those used in metal-electrode DEP but still much less than those use in insulator-based DEP. A voltage in the range of 20 volts has been demonstrated to be sufficient to create a suitable DEP force to manipulate eukaryotic cells when using carbon electrodes [8, 10, 12]. Current efforts are on the patterning of carbon electrodes on top of metal leads to further lower the voltage requirements. Another potential disadvantage in carbon-DEP is the restriction on the kind of substrate used for fabrication. Few materials survive the high temperatures (>900 °C) required during the carbonization process and they can be expensive. For example, the fabrication of carbon electrodes on transparent substrates currently requires the use of fused silica or quartz. However, if a transparent substrate is not required, carbon electrodes can be fabricated in relatively inexpensive substrate such as silicon and silicon oxide. On the positive side, a single substrate, i.e. a 4” wafer, can lead to several experimental devices and the impact of the high cost of the substrate can be minimized.
The fabrication details of carbon-electrode DEP chips featuring transparent and opaque substrates can be found in the works by Martinez-Duarte and colleagues [8, 10]. Carbon-electrode DEP has so far been used for the manipulation of S. cerevisiae [8,10], Drosophila melanogaster , E.coli , DNA and M. smegmatis (data not yet published). The incorporation of carbon-DEP in a centrifugal microfluidics platform has also been demonstrated  towards an automated and portable DEP platform. Modeling and simulation work of carbon-DEP has been carried out by different authors [13,14]. An overview of the use of carbon electrodes in other applications has been presented by Martinez-Duarte et al. .