Medical Device Link.

DESIGN

Future diagnostics

Point-of-care diagnostics (POCT) promises reduced costs, small sample volumes, ease of use and the possibility of the continuous monitoring of patients. Simultaneous measurement of different proteins is necessary to cover the highly dynamic concentration range that is relevant for safe diagnostics. For these applications, microfluidic devices have become increasingly important during the past four years.²
The diagnostics market will be dominated in the future by disposable low-cost products that can be mass produced using industrial technologies. The most time- and cost-consuming steps of conventional laboratories for protein diagnostics, that is, dosing, centrifugation, and purification, must be replaced by integrated polymer microcomponents. As a result, the essential component of this diagnostic platform is likely to be a disposable chip, which includes the microfluidics for blood treatment and detection purposes. There exist many different sensing principles for protein analytics, based on enzyme-linked immunosorbent assays. Only optical detection principles are adequate for dynamic measurement down to nanomolar concentrations.

Operating principles

The modular diagnostic platform µTBC comprises three basic parts. The first part consists of a commercially available syringe for fluid handling, the second part is the disposable chip with integrated blood treatment and detection well, in which the surface coupling chemistry (antibodies) is immobilised. The third part is a compact miniaturised instrument that incorporates innovative laser-induced fluorescence reader technology, and a wireless data transmission module.

With a conventional lancet, the skin of the finger tip is penetrated. A conventional capillary syringe takes approximately 20 µL of blood from the finger and loads the chip using the syringe-pressure-driven flow (typical velocity is 2.5 m/s). Immuno-assay-based tests are performed on cell-free serum or plasma. Therefore, microassay devices for blood tests require integrated on-chip microfluidics for the separation of plasma or serum from whole blood. Standard technologies for removal of cellular components from blood are centrifugation techniques and membrane filtration.³ These separation techniques are time- and cost-intensive processes in daily medical laboratory work and are difficult to miniaturise.⁴

Figure 1: (click to enlarge) The microchannel bend structure for blood separation.

A novel separation unit has been developed that is based on a microchannel bend structure.⁵ It combines two separation mechanisms. The first one is a centrifugal force field, resulting from the bent microchannel. It produces different settling velocities based on density differences between particles and the surrounding fluid. The second separation mechanism is based on an effect called “plasma skimming.” This has primarily been observed in bifurcating capillary blood vessels.
If the flow rates in diverging bifurcations are significantly different, particles tend to enter the branch with a higher flow rate. This results in uneven haematocrit distributions in the daughter vessels.^6–8 In the developed separation unit (Figure 1), typical channel dimensions are 10 µm for the plasma channel width, and 100 µm for the cell channel. The height of all channels is 100 µm. The microchannel bend structure works by employing physical and hydro-dynamic separation mechanisms without the need for integrated actuators such as pumps or valves.

Figure 2: (click to enlarge) Reproducibility of RBC separation between two diferent chips with the same geometry.

Figure 2 shows the run-to-run and chip-to-chip reproducibility of the microchannel bend separation technique for red blood cells (RBC). Two different chips with the same geometry were compared in six consecutive test runs. These experiments were performed with diluted blood at a feed velocity of 2.7 m/s. The RBC separation efficiency varies between 98% and 100% for chip A and between 99% and 100% for chip B. The not shown white blood cell separation efficiency varies between 97% and 98% in all runs. The plasma yield of these chips was approximately 15%.⁹ This separation efficiency matches the demands of plasma generation for the immunoassay. The marked antibodies are immobilised on the surface of the plasma reservoir/well, and react with the proteins in the plasma. Compared with a standard commercial protein C assay, encouraging first results show a better performance and resolution by the µTBC assay.

Prototyping technologies

Investigations on conventional rapid prototyping methods, such as laser structuring and mechanical machining of polymers, highlighted the fact that geometrical limitations, poor surface roughness, and surface modifications by temperature limit the usability of these techniques. Therefore, hot embossing was chosen as a suitable method for fabrication of polymer chips. Although mechanical machining (milling) is simple and cheap, it is also problematic because of limited minimal tool size (40 µm hard tool, 200 µm diamond tool), burr formation, and a minimal average surface roughness of approximately 200 nm for hard tools. Lithographic-based methods are flexible, but expensive and of limited industrial application. To overcome the limitations of conventional tooling, a combination of different tooling methods (hybrid tooling) is a promising approach. High-speed precision milling with microtools used in combination with laser structuring as the finishing method is currently under industrial investigation. Electrochemical milling is a new method offering excellent potential for use in combination with mechanical milling,¹⁰ but it is not yet ready for industrial use. As a consequence, two different methods have been investigated. The first is a modified UV-LIGA (ultraviolet-lithography, electroforming and moulding) process, the second is a combination of mechanical machining with spark-erosion milling.

Making the prototypes

Prototypes of a blood-analysis chip have been fabricated from cyclo-olefin copolymer (COC) (TOPAS 5013, Ticona GmbH, Germany, www.ticona.com), using a UV-LIGA process and hot embossing. From the beginning, 2-mm thick, polished brass wafers were used as the substrate because they are low-cost products with low average surface roughness (Ra <15 nm); other metals such as stainless steel are also suitable. The brass wafers were shaped out of a brass plate using water-jet cutting according to the commonly used 100 mm wafer format. The fabrication process of the mould insert is a modified UV-LIGA process, in which a thick film photo-resist material (NANO SU 8 2100, Microchem Inc., Newton, Massachusetts, USA, www.microchem.com) works as a micromould for the electroplating process with nickel. A typical cycle time for this type of process sequence is approximately 15 hours.¹¹ This flexible processing technology allows rapid optimisation of microfluidic devices by practical experiments, rather than through the (nearly impossible) simulation of a complex fluid-like blood in microcapillaries.

Microelectrodischarge milling (µEDM-milling) or spark-erosion milling is an advancement of the standard electrodischarge machining (EDM) process and similar to numerically controlled milling. It uses a universal tool-electrode of simple shape, for example, spherical or cylindrical, which is moved along a complex path to machine hard materials such as tool steels. EDM is a technology that continuously removes a small amount of work-piece material by controlling the discharge energy applied between tool electrode and work piece.¹² Both electrodes are flushed in a dielectric fluid such as deionised water or oil, which electrically insulates and cools the electrodes as well as removes the particles caught in the sparking gap. A µEDM-milling machine (SX-100, Sarix SA, Losone, Switzerland, www.sarix.com) was used for this work.

Figure 3: µEDM-milled steel tool with multiple plasma channels.

Figure 3 shows first results of a µEDM-milled microchannel bend structure with multiple plasma channels. Good average surface roughness of 100 nm and defined taper angles for low demoulding forces were achieved. Minimising thermal burr formation and further improved geometrical accuracy with lower tolerances are currently under development.

High-volume production

Figure 4: Injection mould with exchangeable mould insert.

For true high-volume production of the disposable chip under industrial conditions, injection moulding is the method of choice. The prototyping methods described above achieved an optimised design (separation efficiency and plasma yield) with six parallel plasma channels. A special mould with exchangeable mould insert was designed (Figure 4). The runner and mould design was supported by process simulation software (Plastics Insight, Moldflow Corp., Charlotte, North Carolina, USA, www.moldflow.com). The UV-LIGA-made mould inserts were used for material and process investigations. These inserts are of limited life span, because the brass and nickel material is brittle in the long-term as a result of stresses induced by the cyclic temperature variations. The typical life of this type of insert is approximately several thousand injection cycles. With these inserts, process optimisation was performed on the COC material. Results of these tests were cycle times of 12 s and improved replication in terms of low shot-by-shot variations of geometrical tolerances (Figure 5). No active tool temperature control was used; imperfect filling of thin-wall cavities therefore resulted in lightly, but tolerable, rounded corners (Figure 6). For economical mass production, mould materials with higher mechanical strength are required.

Figure 5: µTBC disposable chip made by injection moulding COC with an EDM-milled tool steel mould insert.

Future work

Figure 6: Cross-section of plasma channels, detail of Figure 5.

Technology platforms need complete process chains for prototyping and production. So far, within the POCT µTBC platform, processes for prototyping and industrial high-volume production of microfluidic devices for whole blood–protein diagnostics have been developed. Other important elements for POCT are fluid handling, interfacing and the sealing microchannels. Various new sealing methods based on thermal bonding and lamination (reel-to-reel), or laser welding are under further investigation. An interesting new method for sealing disposable polymer chips is a combination of mechanical joining and laser welding.¹³ Real POCT done by the patient is still a vision, but the technology base for new ideas in this field is starting to work now with the µTBC platform.

Acknowledgements

The authors thank the project partners Dr P. Kerth, µTBC-Preventor GmbH (www.preventor.de) and Dr F. Lewitzka, Optimare GmbH (www.optimare.de) for their co-operation. Additional thanks go to Dr S. Bassus and T. Scholz, Deutsche Klinik für Diagnostik Wiesbaden (www.dkd-wiesbaden.de) for measurement support and helpful discussions as well as Dr W. Koester, University Hospital Freiburg, Department of Clinical Chemistry (www.uniklinik-freiburg.de) for establishing and supporting the analyses of the blood specimens.

Dr A. Schoth1* is Group Leader, R. Jurischka1 is a PhD student, Dr C. Blattert1 is a Postdoctorate, I. Tahhan1 is a PhD student and Professor Dr H. Reinecke^1,2 is Head of Laboratory for Process Technology and member of the executive board of HSG-IMIT.

1. Department of Microsystems Engineering IMTEK, Laboratory for Process Technology, University of Freiburg, D-79110 Freiburg, Germany, tel. +49 7612 037 355, e-mail: schoth@imtek.de
2. HSG-IMIT, D-78052 Villingen-Schwenningen, Germany

* To whom all correspondence should be addressed.