BioMEMS: Marrying ICs and biotech
07/01/2002
By Andrew Campitelli, Els Parton, IMEC, Leuven, Belgium
Overview
BioMEMS (biological microelectromechanical systems), and biosensor systems in particular, will offer an important tool in the future of medicine, analytical chemistry, and environmental screening. At their heart is a biosensor module, or BioSiP (Biosensor-System-in-a-Package), in which microfluidics, transducer and a biological top layer are packaged. Together with the development of organic transducers and good immobilization techniques, this package concept is the key to the faster, smaller, cheaper biosensors the market is demanding for applications such as detecting ionic charge, neutral biochemical compounds, antigens, and DNA-sequences. Apart from pharmaceutical applications, biosensors will also be used for environmental screening, in the food and drink industry, as a follow-up to the fermentation process, in agriculture, and for military applications.
Techniques that are currently used to detect all manner of micro-organisms and genetic diseases require a multistep procedure with expensive reagents that can only be performed by well-trained personnel. Detection of a specific DNA sequence takes 5-6 hrs. Furthermore, detection of micro-organisms takes a few days because a culture must be set up to elevate the number of bacteria or fungi into a microscopically detectable amount.
The medical world is therefore anxious to find a way for fast, easier detection (see "Market considerations" on p. 88). By offering speed, low cost, and user-friendliness, biosensors promise an evolution in the analytical chemistry domain. Several characteristics of biosensors explain their low cost: 1) no expensive reagents are needed; 2) in some cases, the same biosensor can be used several times because attachment of the target molecules is reversible (desirable for diagnostic applications); and 3) biosensors may be fabricated with inexpensive polymer materials. The low-cost manufacturing of biosensors makes it possible to use them as disposable devices — an important issue in health monitoring where disposability reduces the risk of contamination.
Biosensors also have potential as field-portable devices to do measurements in situ and in real time. This can be important for the detection of plant diseases in fields or harmful biological agents in the environment in case of biological warfare. Another important advantage that biosensors possess over conventional analyzing methods is their multi-element character. Multiple screening for different components is possible when different arrays of sensors are used.
Nature on a chip
Besides the "natural" component of a biosensor (see "Biomolecules" on p. 88), a second constituent is the electronic transducer that translates the biological recognition reaction into a measurable electronic signal. Depending on the kind of transducer, one can differentiate between amperometric, conductimetric, potentiometric, optical, calorimetric, and acoustic devices. The big challenge with these revolutionary devices is connecting the natural and electronic compounds without influencing one another (in terms of their inherent performance and function). For this purpose, a linking layer is used on which the biomolecules are immobilized and which is attached on top of the transducer.
Biosensor-System-in-a-Package approach
At the heart of the BioSiP concept, or Biosensor-System-In-a-Package (Fig. 1), is a biosensor module in which novel microelectronic packaging solutions can integrate microfluidics, transducer, and biochemical interface components. Transducer configurations, organic surface chemistry, and optimized (specificity, coverage, density, orientation, etc.) biochemical probes are packaged in a durable, simple and inexpensive (often disposable) cartridge.
 Figure 1. Biosensor-System-in-a-Package approach. |
A major challenge in the development of biosensors is the attachment of the organic (biological) receptors to the inorganic transducer. Attention has been focused on affinity biosensors that use the binding of antibodies to antigens, cell receptors to their ligands, and DNA or RNA to nucleic acids with a complementary sequence. So-called immobilization techniques have to be developed to attach these biomolecules (antibodies, cell receptors, DNA, and RNA) to the transducer.
Very thin (nanometer scale) self-assembling monolayers (SAMs) of thiols or silanes on which to immobilize the biomolecules — on, respectively, gold and oxide (mainly SiO2 and Ta2O5) surfaces — were used for the research described in this paper. Different analytical techniques, such as contact angle goniometry, cyclic voltammetry, x-ray photo-electron spectroscopy, atomic force microscopy and scanning tunneling microscopy, and grazing angle fourier transform infrared spectroscopy were used to evaluate and optimize the quality of these so-called "linking layers." Furthermore, the density, distribution and activity of the biomolecules and the type and stability of binding between the coupled molecules and the gold or oxide substrate were investIgated with surface plasmon resonance (SPR) and a quartz crystal microbalance (QCM). It was observed that mixed SAMs show enhanced qualities for the immobilization of biomolecules. In some instances, a higher specificity and sensitivity can be achieved in comparison to commercially available layers.
A second point of interest during the research was the development of novel transducers and transducer applications based on microelectronic technology. The applications of the developed biosensors are sought in a variety of fields, from medical diagnostics and analytical chemistry to environmental monitoring and industrial process control. Specific applications are being developed, such as a polymer-based ChemFET (chemically modified field effect transistor) for pH monitoring and the detection of neutral compounds, an acoustic-based biosensor for the detection of very low concentrations of the prostate specific antigen (PSA), and an interdigitated electrode-based sensor for the detection of specific DNA sequences.
Polymer-based ChemFET
The use of organic materials for microelectronic applications leads to considerably lower costs due to the simplicity of the fabrication process. Conventional ISFET (ion selective field effect transistor) devices are manufactured by CMOS technology using silicon wafers as substrate material; the problems related to long-term stability correlated with the fabrication costs are major limiting factors for practical applications. For health-related applications, however, single-use microsensors are highly desirable because of safety requirements and the reduced lifetime of the biocomponent. For this purpose, organic materials present several advantages, in terms of simpler fabrication techniques compared to silicon processing, compatibility with plastic substrates, and biocompatibility. Appropriate modifications of the organic transistors enable the development of detectors for different analyte species, since both charged and uncharged chemical species can be monitored by the field-induced current variation in the transistor channel if an appropriate recognition layer is added to the transistor to provide chemical specificity.
In order to demonstrate the principle, a pH-sensitive and a glucose-sensitive field-effect transistor have been realized using the polymer poly(3-hexylthiophene) as the active semiconductor material. The sensitivity of the organic transistor toward protons is achieved by directly exposing the dielectric (previously Si3N4, recently Ta2O5) to the aqueous solutions. Changes in the chemical composition of the solution modify the surface potential of the organic semiconductor, and therefore the density of the accumulated charge in the transistor channel.
 Figure 2. Structure of the polymer-based ChemFET. |
In the first design, the organic transducer was fabricated using a silicon wafer as a support. This is only a mechanical supporting structure that does not perform an active role in the device operation. In the future development of the transducer, the silicon will be replaced with a plastic substrate for low cost. For the charge sensitivity, in the region defined by the channel area, the silicon was completely removed. The gold source and drain interdigitated electrodes were deposited by thermal evaporation and patterned by a lift-off technique directly on the gate dielectric. They form ohmic contacts on the conduction channel, and their geometry was chosen to account for the low conductivity of the organic semiconductor. In the final step, the electrodes were covered with a layer of poly(3-hexylthiophene) regioregular (P3HT), by spin-coating from a chloroform solution. The device structure is shown in Fig. 2. Figure 3 illustrates the drain current variation when the device is immersed successively in different pH buffers.
 Figure 3. Current response of ChemFET after immersion in different pH buffers. |
In a second design, a pH detector with improved performance was obtained using electron-beam evaporated Ta2O5 as a H+-sensitive dielectric. E-beam evaporation, perfectly compatible with plastic technology, reduces the complexity of the process flow. The use of Ta2O5 as gate dielectric material is advantageous from both the sensing and transistor point of view. The large number of surface sites that can bind protons makes it the best choice for an insulating material for sensing, whereas its high dielectric constant enables device operation at very low voltages. The devices made with Ta2O5 have shown significantly improved characteristics and a larger output current sensitivity towards pH changes, due to larger transconductance compared to similar Si3N4-based devices.
Field-effect-based detection is generic and can be used in biosensing systems based on enzymes or cells by adding a layer with a specific functionality onto the organic transistor. In this way, an organic-based ENFET (enzyme field effect transistor) for glucose detection was developed by anchoring a layer of the enzyme glucose oxidase onto the gate dielectric.
With the fabrication of a modified organic-based field-effect transistor with proton sensitivity, a low-cost, disposable transducer for ionic charge detection is being developed. The sensitivity of this sensor towards protons is achieved by directly exposing the dielectric (previously Si3N4, recently Ta2O5) to the aqueous solutions. Changes in the chemical composition of the solution modify the surface potential of the organic semiconductor and therefore the density of the accumulated charge in the transistor channel. Figure 3 illustrates the drain current variation when the device is immersed successively in different pH buffers. In order to display directly the pH response, the organic transistor is connected to a specially developed amplifier system. This detection principle is generic and can be used in biosensing systems based on enzymes or cells by adding a layer with a specific functionality onto the organic transistor. In this way, a bioFET was developed by adding the enzyme glucose oxidase on top of the dielectric.
 Figure 4. Structure of the surface acoustic wave-based biosensor for the detection of prostate specific antigen. |
SAW-based sensors
Another biosensor system that IMEC developed is based on an electrochemical wave traveling along the sensing surface of the sensor (surface acoustic wave or SAW) (Fig. 4). This surface is covered with antibodies that exclusively bind prostate specific antigens (PSA). The antibodies are attached to a gold plate by means of self-assembling monolayers (SAMs) of thiols or silane to immobilize the biomolecules on, respectively, gold and oxide (mainly SiO2 and Ta2O5 surfaces).
Interdigitated structures are then positioned on both sides of the gold plate to generate and detect the electromechanical wave. When PSAs attach to the biosensors' antibodies, a difference in mass loading, electrical properties, viscoelastic and mechanical properties of the antibodies can be detected with high sensitivity. This allows the determination of an elevated concentration of PSA in the blood — an indication for prostate cancer. Both single crystal piezoelectric materials (quartz and LiTaO3) and thin piezoelectric films (AIN) are being investIgated.
Novel IDE structures for nucleic acid hybridization detection
An advantage of biosensors in comparison to traditional analyzing methods is the possibility of single-step identification of multiple sequences in, for example, nucleic acid target molecules. This is illustrated in the multi-element device that Innogenetics, IMEC and IMM (Institüt für Mikrotechnik Mainz) developed for the detection of DNA hybridization or immuno-affinity binding of antibody-antigen combinations. Each sensor site comprises many interdigitated electrode structures (IDE) on which a specific DNA probe or antibody is attached (Fig. 5). Binding of the complementary DNA strand or antigen changes the electrical properties in the solution directly above the electrodes. Detection of this impedimetric change is the basis for this biosensor. When different DNA sequences or antibodies are used on top of the different biosensor sites, a multi-element biosensor is available. Prototype polymer arrays consisted of a series of 30 IDE sensor structures with chip dimensions of 22 x 10mm.
 Figure 5. Structure of an interdigitated electrode-based biosensor. |
Miniaturization techniques are very important for impedimetric biosensors in terms of sensitivity and signal to background discrimination. Reduction to micron and even submicron electrode sizes and inter-electrode spacing focuses the interrogated sample-area around the target molecules. For this purpose, deep-UV lithography and metal lift-off techniques are used to make submicron electrode array structures.
SAW-based sensors are manufactured with microelectronics techniques because they can make devices that are relatively small. The process is also 'batch' production, so many devices can be produced at the same time with high reproducibility and precision. Of course, SAWs fabricated this way are not in the smallest dimensions possible in today's microelectronics.
A similar situation exists for IDEs, but the sensing space — the volume in which the electric field expands — is directly related to the dimensions of the electrode fingers (width and spacing). The electric field goes about as far into the measuring solution as the spacing between the electrode. Since our biomolecules are submicron in dimension, we need small electrodes — also submicron — to sense them with an electric field, as close as possible, for high sensitivity.
An important achievement in this European Brite-Euram project was the development of a novel production technique that allows the low-cost production of arrays of interdigitated electrode structures. The new technique is based on a combination of plastics precision molding and oblique evaporation of thin metal layers.
The whole process starts with the realization of the positive silicon copy of the final IDE-device. High-resolution UV lithography and dry etching are necessary to make canals and hills in silicon. This silicon wafer with canals and hills serves as a master for the fabrication of mold inserts. After the injection molding process, the plastic chips are mounted on a special holder for the e-beam evaporator. The interdigitated finger electrodes are formed in one evaporation process step. The shadow of the hills is used to create the interdigitated features. The height of the hills and the angle of the evaporation are selected in such a way that the shadowing effect of the hills provides separation and, as such, an electrical insulation of the even- and odd-numbered electrode fingers from each other.
This manufacturing process will bring the submicron resolution of microelectronics processing to the low-cost production of disposable plastic devices because sophisticated and expensive microelectronics manufacturing techniques, such as photolithography, can be omitted.
Microcalorimetry for the screening of drugs
A fourth kind of biosensor developed at IMEC is a differential microcalorimetry tool in microplate format. This generic technology allows measurement of temperature changes that can be used commercially for the screening of drugs. The system comprises miniaturized calorimeters positioned on a single substrate with integrated temperature sensors by means of microsystem technology. The integrated temperature sensors measure the temperature change between the test well with the potential drug, and the reference well, with a reference sample. A change in temperature between the neighboring wells indicates drug activity. A spin-off initiative called Vivactiss was started to commercialize this technology further.
Conclusion
Biosensors incorporate the unique combination of nature's selectivity and microelectronics' speed and user-friendliness. The driving force for the development of biosensors originally came from the healthcare industry because of its need for rapid diagnosis and treatment. However, the application domains are diverse: environmental screening, food and drink quality control, follow-up of the fermentation process, plant disease detection on the field, detection of harmful biological agents in the environment during wartime, etc. A chemFET-based sensor for low-cost pH detection and two kinds of affinity-based biosensors that either use electromechanical waves or impedimetric changes for the detection of antigens or DNA sequences are just a few of the devices produced by our research.
Andrew Campitelli is group leader of the Biosensor Group at IMEC. He received his PhD in electronic engineering, specializing in sensor technology, from the RMIT, Melbourne, Australia. Previous organization affiliations include LPMO-CNRS, Besançon, France. IMEC, Leuven, Belgium; ph 32/16-281-907, fax 32/16-281-501; [email protected].
Els Parton is scientific editor at IMEC. She has an engineering degree and PhD in applied biological sciences from the KULeuven. Kapeldreef 75, 3001 Leuven, Belgium; ph 32/16-281-467, fax 32/16-281-637, e-mail [email protected].
Market considerations
Despite the numerous advantages and the unique concept of biosensors, these devices are not commonly used in practice. Market acceptance seems to be a major limitation in the overall successful commercialization of biosensor devices. The customer market that is targeted for biosensors is completely different from the traditional one for the IC industry. Indeed, pharmacists, medical professionals, and biologists may have little regard for the underlying technology that is vital to biosensor development. To become competitive with existing laboratory-based assay techniques, sensors need to have considerable advantages concerning costs per test, speed, and ease of manipulation, etc.
At this moment, 90% of the market for biosensors is taken by glucose biosensors. The reason for this is the importance of managing diabetes and the stability of the enzyme used as a biomolecule in the recognition layer. Implantable glucose biosensors can provide a continuous measurement of blood glucose. Ultimately, this biosensor will be linked to an implanted insulin pump, which can guarantee a constant glucose level like a kind of mechanical pancreas.
Synergistic partnerships between engineering and pharmaceutical industries express the strong need to bring biosensors to the level of a whole, complete system. This would facilitate rapid and easy assay investigations, providing a greater access and acceptance of the underlying novel technology.
Biomolecules
A biosensor combines the extreme selectivity and specificity of nature with the speed and user-friendliness of microelectronics. The natural component of a biosensor is a layer of biomolecules that is used for the recognition of other biological molecules to which they react or attach.
The first group, called catalytic biosensors, use enzymes as their recognition element. These are proteins that catalyze specific chemical reactions in our body. A second group is based on the specific attachment between antigens and antibodies (components of our immune system) or complementary DNA strands (our genetic material). In all these cases, natural molecules are preferred because it is impossible to acquire the same level of selectivity with synthetic molecules.