SYSTEM SIMULATION FOR MICROFLUIDIC DESIGN AUTOMATION OF LAB-ON-A-CHIP DEVICES

N. Gleichmann, P. Horbert, D. Malsch, T. HenkelInstitute of Photonic Technology, IPHT, Germany

ABSTRACTThis paper introduces to the fast system simulation approach for droplet-based microfluidics and its validation. Full cus-

tom centric design means the integration and functional combination of microfluidic operation units, such as sample droplet generation, dosing or mixing to a complete Lab-on-a-Chip system. Comparable to large scale integration in EDA a fast simu-lation for droplet-based microfluidic systems is necessary to prevail in the face of their raising complexity by establishing mi-crofluidic design automation. The design process can be greatly enhanced by our fast network simulation that relies on beha -vioral models of the basic microfluidic operation units and their combined effect on the overall behavior.

KEYWORDS: System simulation, lab-on-a-chip, droplet-based microfluidics, custom specific development

Lab-on-a-Chip technology aims to bring complex laboratory workflows into small and portable devices that can be used independently from the laboratory infrastructure. This goal can be realized by the integration and functional combination of dedicated microfluidic operation units as for sample droplet generation, dosing, mixing, merging or aliquoting into custom ap-plication-specific Lab-on-a-Chip systems [1]. To cope with the increasing complexity of development accompanied by cus-tomized solutions [2], robust methodologies and processes developed by modern engineering have to be applied. Up to now iterative development is the most common approach in Lab-on-a-Chip technology. Considering that a single development cycle takes about three months, reducing the number of required iterations becomes a key issue for efficient development of Lab-on-a-Chip systems for custom laboratory workflows.

Figure 1: Development cycle for microfluidic Lab-on-a-Chip systems

The solution is to serialize Lab-on-a-Chip development into well-defined tasks as shown in fig. 1 with defined milestones and decision points. At the end of each task must be decided, whether an additional iteration is required or if the next task can be started. Primary tasks like the requirement analysis and the implementation of the process protocol demand special atten-tion. Failure in an early stage of development entails huge effort later. Efficient computer assistance tools like CAD tools like they exist for microfluidic bio chips can prevent this waste of resources [3].

In this paper we report on the implementation and validation of a software toolkit for system design and simulation of droplet-based microfludic Lab-on-a-Chip devices. The main field of application for the presented fluid network simulation toolkit “FluNet” is to assist the requirement engineering and the design stage of the development process shown as stage one and two in fig.1. It is primarily designed for setup, system simulation and interactive optimization of pressure driven droplet-based microfluidics. The selected software architecture takes care on portability and thus, the toolkit can be easily adapted to other microfluidic platforms with low effort.

MODELING OF COSTUM DROPLET BASED MICROFLUIDIC APPLICATIONSIn droplet-based microfluidics each droplet represents an individual sample that is processed according to a given laborat-

ory workflow. This workflow can be described as a sequence of unit operations. For each unit operation like dosing, mixing, splitting or merging droplet-based microfluidics is providing a microfluidic operation unit [4]. The laboratory workflow is im-plemented by appropriate interconnection of these operation units.

For the analysis of the requirements, depicted in fig. 1 as stage one, it is important to have a clear and unambiguous com-munication between the engineer and the prospective user to ensure the correct functionality of the developed chip. The mi-crofluidic network can be used to discuss the functionality and behavior. In summary, a microfluidic network formalizes the requirements for the lab-on-a-chip system. This kind of information organization is the best input for a computer simulation.

An example chip as shown in fig. 2 basically represents a network of channels, intersections and fluidic interfaces, thus a network of interconnected basic operation units. It is a Lab-on-a-Chip system developed for ultra sensitive in-line quantization of pharmaceuticals by Surface Enhanced Raman Spectroscopy (SERS). SERS has been proven to be a label-free technique with high specificity and sensitivity in the qualitative analysis of antibiotics or other drugs. Combining the application of iso-tope-edited internal standard with the advantages of the liquid-liquid segmented flow-based approach for flow-through SERS detection seems to be a promising means for quantitative SERS analysis [5]. A lubrication layer prevent the direct contact of the water-based SERS-active solution with the channel walls, so that no nanoparticle-aggregates stick to the glass surface. As in this case no memory effects occurs, unsupervised online measurements can be continued over a longer period of time. Fur -ther information on the SERS chip can be found in [5].

978-0-9798064-4-5/µTAS 2011/$20©11CBMS-0001 915 15th International Conference onMiniaturized Systems for Chemistry and Life Sciences

October 2-6, 2011, Seattle, Washington, USA

To set up the chip for simulation, every operation unit has a representative counterpart in the abstract microfluidicnetwork, mapping its functionality as well as geometrical and physical parameters. The microfluidic network map of the chip in fig. 2 is shown in fig. 3. The microfluidic network specifies only the necessary information to understand the behavior of a microfluidic design and to derive the fabrication layout. The full information, like the final placement of the operation units is defined in the additional documentation. Our toolkit provides tools to set up these networks and test their functionality.

IMPLEMENTATION OF THE MICROFLUIDIC NETWORK SIMULATION A software representation of a microfluidic network is a consistent model. It supports the functionality of basic operation

units and the transport of fluids in the channel. Implemented rules for the behavior of the basic operation units and the trans-port characteristics of the channels were deduced from experimental data and CFD simulations.

Simulation as a virtual experiment with the model determines the behavior of droplets in a microfluidic network that is de-rived from an application specific Lab-on-a-Chip device concept. Pressure drop and volume flow are the main state variables from which the droplet displacement and the fluid sequence in the channels can be calculated by taking into account the hy-drodynamic parameters. With respect to the implemented microfluidic organization based on functional nodes that are linked by connectors its overall behavior can be calculated with some general rules.

A central ingredient of the calculation algorithm is the transport model describing the correlation between pressure distri-bution, volume flow rate and the hydrodynamic resistance in the channels. It was derived from experimental data, obtained by measurement of pressure throughput characteristics in test channels with a systematic variation of droplet distance and flow rate. The pressure drop in a channel with a given volume flow rate depends on the count of droplets, the distances between them, the geometry of the channel and the used fluids.

To obtain a mathematical form suitable for the solution algorithm, the simulation network is converted to a matrix repres-entation. This matrix is based on the application of the Kirchoff's circuit law as a system of equations. The values for resist -ance are calculated with the help of the transport model and the operation unit rules are applied.

The simulation is fast enough to enable the engineer to test a micro channel network design within minutes what otherwise would require days or weeks. As a result the simulation during early design stages gives certainty, because their results can be often compared with the requirements. These facts prove the practical importance of the development toolkit.

VALIDATIONTo validate the software implementation and the transport models, a circular micro channel system with two T-shaped

junctions for inflow and outflow of droplets is used for the benchmark as microfluidic state machine. The system was previ -ously described for investigation of droplet transport characteristics [6,7]. The flow rate in each of the branches depends on the number and distance of the droplets inside. A droplet, arriving at the loop is directed to the branch with the higher flow rate. So, transport of droplets continuously changes the system state. These changes are measured by digital imaging and com-pared with the simulated system evolution.

Touchstone of the network simulation algorithm is its ability to handle complex network architectures with integrated flu-idic loops correctly. The experimental verification leads to the result shown in fig. 4. It reflects general effects of droplet-based flow in channels. Change of the droplet count results in a change of the measured or calculated volume flow. To set up the simulation all geometrical and hydrodynamic parameters are adapted to the experimental setting.

The comparison between the simulation and the experiment exposes the nearly identical behavior of both. The differences in droplet timing entering or leaving the loop is based on the high sensitivity to small differences in the experimental settings.

Figure 3: The microfluidic network map is an abstrac-tion of the chip to the level of interconnected operation

units. Driven by tetradecane as continuos medium, the first injector forms droplets consisting of analyte, water and in-ternal standard. Then silver colloid and activation agent

are added at the second injector. Readout is performed after a given residence time in the dwelling channel.

Figure 2: A microfluidic chip fabricated in all-glass technology for ultra sensitive in-line quantization of

pharmaceuticals by Surface Enhanced Raman Spectro-scopy (SERS). The corresponding network map depict-

ing the process protocol is shown in fig. 3.

916

Figure 4 Comparison of experimental results (upper diagram) with simulation results (lower diagram). The pictures on the right side show the generated droplet patterns in the experiment. The experimental un-

certainties can lead to a minimal delay in the droplet timing for a split second.

The droplet count in a channel changes the droplet velocities in both branches. In consequence, the choice of the droplets influence the moment of the next leaving droplet, which changes the system behavior again. The simulation results were ex-actly reproduced for each periodical droplet pattern. So the simulation algorithm, the underlying models and the droplet ad-ministration were validated.

CONCLUSIONSThe development of droplet based microfluidic lab-on-a-chip systems needs a fast simulation ability to conquer the raising

complexity of user-defined Lab-on-a-Chip systems. The introduced fast system simulation approach provides such a solution.

ACKNOWLEDGEMENTSWe gratefully acknowledge the German BMBF, the Free State of Thuringia, the European Regional Development Fund

(EFRE) and the European Union in the projects DiNaMiD signature 03144916, BioOptiSens signature 2008FE911, Mik-roplex signature PE-113-1, Photonics4Life grand agreement number: 224014

REFERENCES[1] T. Henkel, Chip modules for generation and manipulation of fluid segments for micro serial flow processes, Chemical

Engineering Journal 101, pp. 439-445 (2004).[2] A.B. Theberge, F. Courtois,Y. Schaerli, M. Fischlechner, C. Abel, F. Hollfelder and W.T.S. Huck, Microdroplets in Mi-

crofluidics: An Evolving Platform for Discoveries in Chemistry and Biology, Angew. Chem. Int. Ed., 49, pp. 5846 – 5868 (2010).

[3] K. Chakrabarty and F. Su, Design Automation Challenges for Microfluidics-Based Biochips, Proc. IEEE Design, Test, Integration and Packaging of MEMS/MOEMS (DTIP), pp. 260-265, (2005).

[4] S. Haeberle, R. Zengerle, Microfluidic platforms for lab-on-a-chip applications, Lab Chip, pp. 1094-1110 (2007).[5] A. März, R.K. Ackermann, D. Malsch, T. Bocklitz, T. Henkel and J. Popp, Towards a quantitative SERS approach

online monitoring of analytes in a microfluidic system with isotope-edited internal standards , J. Biophoton. 2, No. 4, pp. 232–242 (2009).

[6] D. Link, S. Anna, D. Weitz and H. Stone, Geometrically Mediated Breakup of Drops in Microfluidic Devices, Phys. Rev. Lett. 92 (2004).

[7] M.J. Fuerstman, P. Garstecki and G.M.Whitesides, Coding/Decoding and Reversibility of Droplet Trains in Microfluidic Networks, Science 315, pp. 828-832 (2007).

917