Micro gas sensor array with neural network for recognizing combustible leakage gases

A micro gas sensor array, consisting of four porous tin oxide thin films added with noble metal catalysts on a micro-hotplate, was designed and fabricated. The micro-hotplate was designed to obtain a uniform thermal distribution along with a low-power consumption and fast thermal response. The sensing properties of the sensors toward certain combustible gases, i.e., propane, butane, LPG, and carbon monoxide, were evaluated. A multilayer neural network was then used to classify the gas species. The results demonstrated that the proposed micro sensor array, plus multilayer neural network employing a backpropagation learning algorithm, was very effective in recognizing specific kinds and concentration levels of combustible gas below their respective threshold limit values.     

Ionic liquid high-temperature gas sensor array

A novel sensor array using seven room-temperature ionic liquids (ILs) as sensing materials and a quartz crystal microbalance (QCM) as a transducer was developed for the detection of organic vapors at ambient and elevated temperatures. Ethanol, dichloromethane, benzene, and heptane were selected as representative gas analytes for various kinds of environmental pollutants and common industrial solvents. The QCM/IL sensors responded proportionately and reversibly to the organic vapor concentrations (i.e., ethanol, heptane, and benzene) in the gas phase from 0 to 100% saturation at room and elevated temperatures (e.g., 120 degrees C) but deviated from this linear relationship at high concentrations for dichloromethane, a highly volatile compound. Linear discriminant analysis was used to analyze the sensing patterns. Excellent classifications were obtained for both known and unknown concentrations of vapor samples. The correct classifications were 100% for known concentration samples and 96% for samples with unknown concentrations. Thermodynamics and ATR-FT-IR studies were conducted to understand specific molecular interactions, the strength of the interaction between ILs and organic vapors, and the degree of ordering that takes place upon dissolution of the vapors in ILs. The different response intensity of the QCM/IL sensors to the organic vapors depends on the different solubilities of organic vapors in ILs and varying molecular/ion interactions between each organic vapor and IL. The diverse set of IL studied showed selective responses due to structural differences. Therefore, a sensor array of ILs would be able to effectively differentiate different vapors in pattern recognitions, facilitating discrimination by their distinctive patterns in response to organic vapors in both room and high temperatures.     

Integrated sensor array for analysis of multicomponent gas mixtures

A new device has been proposed and tested experimentally for the discrimination of gases in “electronic nose” systems. The device consists of an array of surface-acoustic wave (SAW) sensors positioned on a single anisotropic substrate with a common gas-sensitive coating for each sensor. The specificity of the sensors is provided entirely by the elastic anisotropy of the single-crystal substrate: changes in the direction of propagation of the wave through the gas-sensitive film deposited on the anisotropic substrate are accompanied by changes in the partial components of the mechanical displacement of the wave and corresponding contributions to the resultant SAW “response.” Pis’ma Zh. Tekh. Fiz. 24, 40–45 (August 26, 1998)     

Development and calibration of field-effect transistor-based sensor array for measurement of hydrogen and ammonia gas mixtures in humid air

A sensor array for analyzing hydrogen and ammonia gas mixtures in humid air has been developed, built into a rugged system, and calibrated for laboratory testing. The sensor array is comprised of four chemically sensitive field-effect transistors (CHEMFETs). Chemically sensitive layers for the sensors were developed and tested using a Kelvin probe. A combination of catalytic and noncatalytic thin layers (palladium and polyaniline) was selected for the four-sensor array. The work function responses of the CHEMFET sensor array to mixtures of hydrogen, ammonia, and humid air were measured. Chemometric multivariate methods, linear and nonlinear partial least squares, were used for the calibration of the sensor array using gas mixtures in the concentration range from 0 to 10?000 ppm hydrogen and ammonia in humid air. The sensor array for ammonia showed good sensitivity, selectivity, response time, and stability and is recommended for field deployment. In contrast, the sensor array for hydrogen, though highly sensitive to hydrogen, demonstrated inadequate stability, requiring further development before deployment is recommended.     

A novel sensor array for field based ocular ultraviolet radiation measurements

The intensification of terrestrial solar ultraviolet radiation (UVR) due to the diminution of the ozone layer has promoted a variety of research into establishing the impact of this elevated potential dose of UVR on biological tissues. Certain anterior ocular tissues have been found to be susceptible to damage by incident UVR and potentially blinding diseases such as pterygium are thought to be a direct result of absorbed UVR at the nasal limbus. There is a need for more accurate quantification and localisation of incident UVR at the anterior ocular surface. A novel solar blind photodiode sensor array system has been designed, constructed and tested for this purpose. Initial measurements to quantify the irradiance across the anterior ocular surface within the latitudes known as the 'pterygium belt' provide us with a set of core data for different head orientations and tilt angles and indicate the accuracy and stability of the system.     

Neuro-fuzzy network for flavor recognition and classification

This paper presents the neuro-fuzzy Takagi-Sugeno-Kang (TSK) network for the recognition and classification of flavor. The important role in this process fulfills the self-organizing process used for the creation of the inference rules. The self-organizing neurons perform the role of clustering data into fuzzy groups with different membership values (the preprocessing stage). Applying the automatic control of clusters, we have the optimal size of the TSK network. The developed measuring system has been applied for the recognition of flavor of different brands of beer. The fuzzy neural network is used for processing signals obtained from the semiconductor sensor array. The results of numerical experiments have confirmed the excellent performance of such solutions.     

Contribution of electron tunneling transport in semiconductor gas sensor

It was found that thin film devices derived from SnO₂ sols by spin-coating method showed unique thermal behavior of electric resistance in air involving a temperature region where resistance was independent of temperature. The temperature independent resistance region extended up to 400 °C, replacing a region of temperature-conventionally dependent resistance, as film thickness increased. Such unique behavior of resistance was observed also for a brush-coated device but not for screen-coated thick film devices or disk-type device, suggesting that the absence of mechanical forces applied during device fabrication favored the occurrence of the unique behavior. It was shown that the unique behavior could be well accounted for by postulating a combination of electron tunneling transport and conventional migration transport. Calculation of tunneling probability based on a simple model allowed estimating that electron tunneling transport can take place between oxide grains with a probability of 0.01 or larger if a gap in between is narrower than 0.01 nm.     

Improved probabilistic neural network algorithm for chemical sensor array pattern recognition

An improved probabilistic neural network (IPNN) algorithm for use in chemical sensor array pattern recognition applications is described. The IPNN is based on a modified probabilistic neural network (PNN) with three innovations designed to reduce the computational and memory requirements, to speed training, and to decrease the false alarm rate. The utility of this new approach is illustrated with the use of four data sets extracted from simulated and laboratory-collected surface acoustic wave sensor array data. A competitive learning strategy, based on a learning vector quantization neural network, is shown to reduce the storage and computation requirements. The IPNN hidden layer requires only a fraction of the storage space of a conventional PNN. A simple distance-based calculation is reported to approximate the optimal kernel width of a PNN. This calculation is found to decrease the training time and requires no user input. A general procedure for selecting the optimal rejection threshold for a PNN-based algorithm using Monte Carlo simulations is also demonstrated. This outlier rejection strategy is implemented for an IPNN classifier and found to reject ambiguous patterns, thereby decreasing the potential for false alarms.     

基于非线性 PCA的微气体传感器阵列信号处理

在线性叠加模型基础上提出了气体传感器对混合气体的非线性叠加模型,并引入了非线性主成分分析(Nonlinear Principal Component Analysis,NLPCA)法对微传感器阵列的信号进行处理。使用该模型对由4个微热板式气体传感器组成的阵列的信号进行了分析,对照基于线性叠加模型的主成分分析法(Principal Component Analysis,PCA)的识别结果,说明该方法能够提高对混合气体识别和量化的准确度。     

Metal oxide-based monolithic complementary metal oxide semiconductor gas sensor microsystem

A fully integrated gas sensor microsystem is presented, which comprises for the first time a micro hot plate as well as advanced analog and digital circuitry on a single chip. The micro hot plate is coated with a nanocrystalline SnO2 thick film. The sensor chip is produced in an industrial 0.8-microm CMOS process with subsequent micromachining steps. A novel circular micro hot plate, which is 500 x 500 microm(2) in size, features an excellent temperature homogeneity of +/-2% over the heated area (300-microm diameter) and a high thermal efficiency of 6.0 degrees C/mW. A robust prototype package was developed, which relies on standard microelectronic packaging methods. Apart from a microcontroller board for managing chip communication and providing power supply and reference signals, no additional measurement equipment is needed. The on-chip digital temperature controller can accurately adjust the membrane temperature between 170 and 300 degrees C with an error of +/-2 degrees C. The on-chip logarithmic converter covers a wide measurement range between 1 kOmega and 10 MOmega. CO concentrations in the sub-parts-per-million range are detectable, and a resolution of +/-0.1 ppm CO was achieved, which renders the sensor capable of measuring CO concentrations at threshold levels.     

Facile hyphenation of gas chromatography and a microcantilever array sensor for enhanced selectivity

The very simple coupling of a standard, packed-column gas chromatograph with a microcantilever array (MCA) is demonstrated for enhanced selectivity and potential analyte identification in the analysis of volatile organic compounds (VOCs). The cantilevers in MCAs are differentially coated on one side with responsive phases (RPs) and produce bending responses of the cantilevers due to analyte-induced surface stresses. Generally, individual components are difficult to elucidate when introduced to MCA systems as mixtures, although pattern recognition techniques are helpful in identifying single components, binary mixtures, or composite responses of distinct mixtures (e.g., fragrances). In the present work, simple test VOC mixtures composed of acetone, ethanol, and trichloroethylene (TCE) in pentane and methanol and acetonitrile in pentane are first separated using a standard gas chromatograph and then introduced into a MCA flow cell. Significant amounts of response diversity to the analytes in the mixtures are demonstrated across the RP-coated cantilevers of the array. Principal component analysis is used to demonstrate that only three components of a four-component VOC mixture could be identified without mixture separation. Calibration studies are performed, demonstrating a good linear response over 2 orders of magnitude for each component in the primary study mixture. Studies of operational parameters including column temperature, column flow rate, and array cell temperature are conducted. Reproducibility studies of VOC peak areas and peak heights are also carried out showing RSDs of less than 4 and 3%, respectively, for intra-assay studies. Of practical significance is the facile manner by which the hyphenation of a mature separation technique and the burgeoning sensing approach is accomplished, and the potential to use pattern recognition techniques with MCAs as a new type of detector for chromatography with analyte-identifying capabilities.     

Independent component analysis applied on gas sensor array measurement data

The performance of gas-sensor array systems is greatly influenced by the pattern recognition scheme applied on the instrument's measurement data. The traditional method of choice is principal component analysis (PCA), aiming for reduction in dimensionality and visualization of multivariate measurement data. PCA, as a second-order statistical tool, performs well in many cases, but lacks the ability to give meaningful representations for non-Gaussian data, which often is a property of gas-sensor array measurement data. If, instead, higher order statistical methods are considered for data analysis, more useful information can be extracted from the data. This paper introduces the higher order statistical method called independent component analysis (ICA) as a novel tool for analysis of gas-sensor array measurement data. A comparison between the performances of PCA and ICA is illustrated both in theory and for two sets of practical measurement data. The described experiments show that ICA is capable of handling sensor drift combined with improved discrimination, dimensionality reduction, and more adequate data representation when compared to PCA.     

Results of network analyzer measurements with leakageerrors-corrected with direct calibration techniques

Exact calibration procedures to correct two-port measurements with leakage errors are developed for a network analyzer having three or four measurement channels. Many measurement systems including open air devices, such as MMIC wafer probes, contain numerous crosstalk paths which are included in the full models of the direct calibration procedures. These novel 15-term and 22-term procedures are based on closed solutions and need five and six successive measurements of completely known calibration standards. First, both procedures solve these general error models exactly. The reduction of the common 22-term model to a 10-term model describes a network analyzer with three measurement channels in a simple and common manner. In this way one may create a great number of new calibration procedures. The sensitivity and accuracy of a direct 15-term and a direct 22-term error correction procedure is demonstrated by experimental results     

Magnetically actuated complementary metal oxide semiconductor resonant cantilever gas sensor systems

In the present paper, an electromagnetically actuated resonant cantilever gas sensor system is presented that features piezoresistive readout by means of stress-sensitive MOS transistors. The monolithic gas sensor system includes a polymer-coated resonant cantilever and the necessary oscillation feedback circuitry, both monolithically integrated on the same chip. The fully differential feedback circuit allows for operating the device in self-oscillation with the cantilever constituting the frequency-determining element of the feedback loop. The combination of magnetic actuation and transistor-based readout entails little power dissipation on the cantilever and reduces the temperature increase in the sensitive polymer layer to less than 1 degrees C, whereas previous designs with thermally actuated cantilevers showed a temperature increase of up to 19 degrees C. The lower temperature of the sensitive polymer layer on the cantilever directly improves the sensitivity of the sensor system as the extent of analyte physisorption decreases with increasing temperature. The electromagnetic sensor design shows an almost 2 times larger gas sensitivity than the earlier design, which is thermally actuated and read out using p-diffused resistors. The gas sensor is fabricated using an industrial complementary metal oxide semiconductor (CMOS) process and post-CMOS micromachining.     

Precise measurements of small gas flows—metrological basis for the calibration of leak detectors

This article discusses technical aspects of the calibration of leak detectors by using a series of artificial control flows. Results are presented from the calibration of several mass-spectrometric leak detectors. Procedures and methods of analyzing measurements obtained in the calibration of such flows are discussed, and several approaches to the metrological certification of halogen and electron-capture leak detectors are described. Translated from Izmeritel'naya Tekhnika, No. 2, pp. 20–23, February, 1997.     

基于时延冗余测量的阵形校准

海底柔性水平阵的阵形校准,可以利用两个校准声源测量阵元间的声传播时延差来求解得到。传统算法只利用了相邻阵元信号测量的时延差,但误差会随噪声增大而出现累积。在信噪比降低时,所提方法通过冗余时延测量有效地减少了误差累积,从而提高了阵形校准的精度与稳定性。数值仿真与实验结果都验证了该方法的有效性。