The parameters used to rate acoustic wave-based chemical sensors are sensitivity and selectivity. While sensitivity can be improved by selecting the appropriate sensor device, selectivity still remains a major concern. This is primarily because the analytes under consideration often belong to the same group/class of chemicals. In such cases, the sensor output does not provide enough information to reliably identify, estimate and/or classify the analytes being investigated. As a result, data analysis techniques are used to extract selectivity from the sensor. An approach to analyze sensor signal data using statistical pattern recognition techniques such as principal component analysis and nearest neighbor algorithm is presented.

Mass sensitive devices such as QCMs and SAWs coated with molecularly imprinted layers are suitable for chemical recognition and characterizing chemical changes in complex mixtures. This strategy makes it possible to e.g. monitor degradation processes in automotive engine oils.

In this contribution, a flow-through potentiometric micro sensor is described which is based on semi-permeable tubing. Basically the proposed ion selective electrodes are of the liquid membrane type having an internal electrolyte. Sensors were constructed by guiding 0.3 mm diameter dialysis tube from an artificial kidney through a cavity, precision machined in Perspex^TM.

Microarrays represent a new approach to the rapid detection and identification of analytes. Studies to date have shown that the immobilization of receptor molecules (such as DNA, oligonucleotides, antibodies, enzymes and binding proteins) onto silicon and polymeric substrates can result in arrays able to detect hundreds of analytes in a single step. The formation of the receptor/analyte complex can, itself, lead to detection, or the complex can be interrogated through the use of fluorescent, chemiluminescent or radioactive probes and ligands.

In this work, micromachining and planar processing have been used to produce gas sensing devices with lower power consumption at lower cost. The small size brings new advantages for chemical selectivity as well: multi-element arrays whose time-varying signals can be interpreted using pattern recognition methods. The device platform is a `microhotplate,' consisting of a built-in heater, thermometer, and electrodes to probe the sensing films. Microhotplates are fabricated using CMOS-compatible technologies, enabling on-chip circuitry for multiplexing and signal amplification.

Microfluidic components, especially micropumps, are essential for miniaturized dosing systems, Micro Total Analysis Systems and miniaturized labs for medical diagnosis. A variety of micropumps for these different applications have been developed at the Institut fuer Mikrotechnik Mainz GmbH.

An automatic sensor system with chemical sensors for water analysis is presented. It is part of a groundwater monitoring network which allows frequent measurements at low cost. The sensor system has an active fluidic system that combines micromachined and non-micromachined components to take water samples automatically. Three different types of sensors are included in the system, two ion selective sensors and an optofluidic microsystem that works as a microphotometer. It is shown that photometric analysis of different substances in water can be carried out in the microfluidic system using commercially available photometric reagents. Finally, an integrated microsystem for dosing and mixing of sample and reagent and photometric measurements on one chip is presented.

Automated microfluidic analysis has historically been carried out by flow injection analysis techniques. Sequential injection analysis represents a more versatile method for automated fluid handling. We have explored the use of sequential injection analysis for performing microcolumn separations. These separations can be used as part of a microanalytical procedure, or for sample preparation. In addition, with detection of retained species on the microcolumn, sequential injection separation represents a technique for sensing. Recently, it has been demonstrated that sequential injection separation can be carried out with renewable separation columns--the beads with interactive surfaces can be delivered to the microcolumn, used for processing the sample, and discarded after each measurement. Delivery of new beads for each measurement provides a method for renewable surface separation and renewable surface sensing. Applications in environmental analysis and bioanalytical chemistry will be presented.

Surface-enhanced Raman spectroscopy (SERS) promises to be one of the most sensitive methods for chemical detection and in recent years SERS has been used for chemical, biochemical, environmental, and physiological applications. A variety of methods using various media (electrodes, colloids, and substrates) have been successfully developed to enhance Raman signals by six orders of magnitude and more. However, SERS has not become a routine analytical technique because these methods are unable to provide quantitative measurements. This is largely due to the inability to fabricate a sampling medium that provides reversible chemical adsorption, analysis-to-analysis reproducibility, unrestricted solution requirements (reagent concentration and pH) or sample phase (liquid or solid). In an effort to overcome these restrictions, we have developed metal-doped sol-gels to provide surface-enhancement of Raman scattering.

We describe a new approach to fluorescence sensing based on fluorescence polarization measurements. The sensing device consists of an analyte sensitive fluorescence probe and a reference fluorophore which is not affected by the analyte. Combined emission from probe and reference passes through two adjacent orthogonally oriented polarizers and is viewed with second analyzer polarizer. Changes in the probe intensity result in changes in the polarization of the combined emission. The analyzer polarizer is rotated to yield equal intensity from both sides of two orthogonally oriented polarizers. So constructed sensor earlier was used for manual visual detection of RhB in intralipid and to measure pH using 6-carboxyfluorescein. A sensor equipped with the simple electronic detection system of a dual photocell and a Watson bridge improves the accuracy. We used this device with UV hand lamp, electroluminescent light source or LED to detect pH, oxygen and calcium. This sensing method is generic and can be used with any fluorophore which displays an analyte-dependent change in intensity.

A new lead sulfide integrating detector assembly has been developed for applications in NIR spectroscopy in the 1 - 3 micron spectral region. The assembly contains a 128 element, thermoelectrically cooled, multiplexed detector array sensor, a driver control unit, and a software interface for signal digitization and processing. The system has selectable single scan integration times of 1 ms - 25 ms, and a signal-to-noise dynamic range of greater than 1,000 at 10 ms integration time. Dark referencing is used to control dark signal drift and to maintain pixel-to-pixel offset uniformity to within 1%. The PbS detector assembly has been interfaced with an optical fiber, a dispersion grating, and a laptop computer, to produce a compact 1 - 3 micron spectrophotometer with a per pixel resolution of 12 nanometer.

The fabrication and performance of a micro-sensor for NMR- spectroscopy of nanoliter-sample volumes is presented. On both glass and GaAs-substrate. Planar coils with inner diameter from 50 micrometers to 400 micrometers including a coplanar wave-guide leading to the bonding pads were fabricated. A chamber for the liquid samples of 200 - 500 micrometers diameter was etched isotropically on the backside of the substrate, located under the coil. In initial experiments, the spectrum of a 20 - 50 nl-volumes of pure silicon-oil is analyzed in a ¹H-NMR experiment in a 11T spectrometer (500 MHz). The microcoil serves as a receiver, while the RF-power was transmitted by a macroscopic coil perpendicular to the receiver coil. We observe characteristic lines from the silicon-oil spectrum which clearly indicates the high sensitivity of the microcoil. Additional signal from different materials in the experiment are suppressed by gradient fields and an adequate design of the sensor.

The sensing range of surface plasmon resonance (SPR) refractometry is greater than the thickness of most thin films of interest. Therefore, an SPR sensor will also respond to changes in the refractive index (RI) of the bulk analyte adjacent to the thin film, caused for instance by variations in analyte composition or temperature. These changes in bulk RI degrade the quality of SPR sensing data. One solution to this problem is simultaneously to measure both the SPR response and the bulk RI of the analyte and correct the SPR response for bulk RI variations. We present a simple implementation of this approach which uses critical angle refractometry. Our sensor is based on Texas Instruments' Spreeta^TM SPR sensor. The gold is removed from the portion of the sensor surface which corresponds to angles less than the critical angle. The modified sensor delivers a composite spectrum which may be used for measurements of both the critical angle edge and the SPR dip. Theory of critical angle compensation is presented, and calibration and data analysis issues are outlined. Critical angle compensation for temperature and concentration induced bulk RI changes is demonstrated in detergent adsorption and antibody binding experiments.

Chemosensory devices with self organized structures and artificial receptors are developed for a wide field of applications. Supramolecular hosts, highly ordered liquid crystal phases and even Langmuir Blodgett films are promising recognition elements. Cage compounds such as tert- butyl-calix[n]arene show high preorganization due to their rigid walls and form highly symmetrical cavities suitable for host guest inclusion of analytes. Disturbance of the highly ordered cholesteric and nematic phases influences the optical and dielectric properties of these materials.

The lateral d.c. resistivity of thin metal films with layer thicknesses of less than 30 nm is increased due to the adsorption of certain particles and is decreased by their desorption. The contribution of the adatoms to the film resistivity can be understood similarly to the effect of foreign atoms in a bulk metal. The magnitude of the resistivity increase is related to the surface coverage of the thin metal film. Using thin metal films of gold as working electrodes in a conventional three-electrode arrangement, a novel electrochemical microsensor, based on the described mechanism of the surface resistivity changes has been developed. The thin film sensor has been prepared by means of process steps of silicon planar technology. With this sensor the trace analysis of heavy metals, such as cadmium, lead, nickel, thallium, and zinc ions as well as cadmium-EDTA complexes in aqueous solutions is possible. The different species could be distinguished from each other due to their characteristic stripping potentials. For the investigated species a linear signal relation has been obtained over a wide range of concentrations from several ppb to some ppm.

The application of sensitive layers for chemical microsensors consisting of multicomponent compositions and dielectric materials requires specific deposition techniques, since the different chemical and physical properties of the respective components can be significantly disturbed during the deposition process. To avoid this drawback, the pulsed laser deposition technique is suggested as a novel thin film preparation method for such sensor devices.

A chemical sensor system based on arrays of surface acoustic wave (SAW) delay lines has been developed for identification and quantification of volatile organic compounds. The individual SAW chemical sensors consist of interdigital transducers patterned on the surface of an ST-cut quartz substrate to launch and detect the acoustic waves and a thin film coating in the SAW propagation path to perturb the acoustic wave velocity and attenuation during analyte sorption.

We report on results achieved with three different types of polymer-coated chemical microsensors fabricated in industrial CMOS technology. The first and most extensively studied transducer is a microcapacitor sensitive to changes in dielectric properties of the polymer layer due to analyte absorption. An on-chip integrated (Sigma) (Delta) -converter allows for detecting the minute capacitance changes. The second transducer is a resonant cantilever sensitive to predominantly mass changes. The cantilever is electrothermally excited, its vibrations are detected using a piezoresistive Wheatstone bridge. In analogy to acoustic wave devices, analyte absorption in the polymer causes resonance frequency shifts as a consequent of changes in the vibrating mass. The last transducer is a microcalorimeter consisting of a polymer-coated sensing thermopile and an uncoated reference thermopile each on micromachined membranes. The measurand is the absorption or desorption heat of organic volatiles in the polymer layer. The difference between the resulting thermovoltages is processed with an on-chip low-noise differential amplifier. Enthalpy changes on the order of (mu) J have been detected.

In this paper we will demonstrate chemometric approaches that can be applied to data from a well-understood polymer- coated acoustic wave vapor sensor array to extract information about the properties of detected vapors, whether that vapor was in the training set or not. Derivation of the approach and simulation using `synthetic' data are presented.

There is a clear need for a portable analytical system with high sample throughput, which yields rapid and clear results at a field level for color-based assays. ELISA is a widely used technique for environmental and clinical analysis. However it is time consuming, non-portable, requires expensive plate reading equipment and skilled analysts. An alternative system has been developed in our laboratories using a digital camcorder as a detection method. Preliminary results generated in our laboratories using this approach to screen antibody-antigen reactions have been very promising.

The task of optimization of gas sensor characteristics such as absolute value and temperature range of sensitivity, time response and selectivity is one of the most important problems of gas sensor design. Unfortunately, at present time the decision of this problem has empirical character, which is not effective, because of multi-factor task. We propose another approach to GS optimization, which is based on the theoretical modeling of gas sensing characteristics in the framework of chemisorptional views. Such approach is more effective as it permits both to understand and to predict the influence of surface and bulk parameters of metal oxide films on sensing characteristics.

A novel approach to multichannel SPR sensing based on encoding information from separate sensing channels into a single optical spectrum is presented. A dual channel SPR sensor using this approach is demonstrated. Attention is given to exploitation of the dual-channel SPR sensor for compensation for background interference and non-specific adsorption of the biomolecules to the surface of the SPR biosensor. Experimental results indicate that background refractive index changes were compensated with accuracy better than 8 X 10^-5 RIU (refractive index unit); the effect of a temperature change of 3.6 K was reduced by a factor of 13 by the dual-channel sensor. SPR biosensor-based detection of monoclonal anti-dinitrophenyl antibody (a-DNP) with compensation for non-specific adsorption is demonstrated.

The rising degree of miniaturization in sensor technology and the efforts to make industrial use of it require an adequate solution for coating of sensors with membranes needed for various applications. A fully automated dispensing device has been developed which is capable of dispensing droplets in nanoliter range with high accuracy and reproducibility. The device combines a three axles positioning system with a pattern recognition system and a dispensing value and is suited for industrial mass production of sensors. Up to 150 droplets per minute are possible. Positioning accuracy is below three micrometer and standard deviation of the dispensing process is 2% or lower. The reproducibility of the process is independent from properties of the medium to be dispensed such as viscosity or solvent and shows no dependence on dispensing parameters such as needle diameter or dispensing time. The measurement of dissolved oxygen in a liquid solution serves as application example to show the practical suitability of the dispensing device.

In this work, an identification system based on an array of semiconductor tin dioxide gas sensors has been developed. This system has proven a good success rate in the discrimination of carbon dioxide, forane or their mixtures without a sensor dedicated to carbon dioxide. After a characterization of the five sensor array, pre-treatment was tested on the collected data to select the well representing parameters. Then the information contained in the resulting look-up table was analyzed with PCA, but no significant result was observed. Discriminant factorial analysis was then used and has shown a better separation of the different clusters and unknown data were taken to validate the classification. So our results show that a reliable system can be designed using non-dedicated chemical semiconductor gas sensor.

In air conditioned atmospheres, relative humidity (RH) is generally fixed in the 30% - 60% RH range to be comfortable. But tin oxide gas sensors, which are generally used in electronic nose applications, are sensitive to the humidity. So, we have studied the influence of the relative humidity rate variation on the response of a Tagushi type sensor array used in an environmental electronic nose application. In this paper, we first summarize the sensor response distribution in terms of conductance dynamic slope or steady-state conductance values. Afterwards, two pattern recognition methods, Principal Component Analysis and Discriminant Factorial Analysis, are successively applied. We show that we are able to identify the target gas, a refrigerant gas Forane R134a, and also to well discriminate an unknown case by using only the conductance dynamic slope even if the relative humidity rate varies.