Search

This paper presents an equivalent electric circuit model that describes adsorption-desorption processes occurring on bio and chemical sensor surfaces under the Langmuir hypothesis and considers the following practical case: the pressure or concentration of the particles in the test chamber is not perturbed by these processes and keeps its initial value, as in the cases of relatively high pressure or concentration values with zero molecular flow, or in the presence of a molecular flow at any pressure or concentration value. It is also pointed out that the equivalent circuit for Langmuir adsorption is similar to the circuit proposed for enzymatic reactions. Even if this work essentially covers theoretic aspects, a way is suggested for the possible experimental determination of both adsorption-desorption parameters and adsorption-desorption site density.

In this paper the use of multi-transduction principle for sensing materials development is reviewed. In particular, the application of porphyrin-based films to a multi-transduction Electronic Tongue system for different analytical tasks is presented. The optical response of sensing films was registered by means of Computer Screen Photoassisted Technology (CSPT) that applies familiar devices, such as computer monitor screen and web-camera, as illumination light source and signal detectors. Simultaneously the electrochemical amperometric or potentiometric response of the same sensing material was measured. Data analysis combining both signals significantly improves the performance of the Electronic Tongue, thus opening new frontiers in application of such a system.

Sensor networks prompt a great deal of research interest within the computer and analytical sciences. To this regard, one of the most important issues is concerned with the interpretation of data that are collected by different sensors. Due to sensors non-reproducibility, this problem may also persist even when many replicas of the same sensors are considered. In this case additional calibrations may be required to use a common knowledge database. Noteworthy, the same problem arises in case of sensors replacement. In this paper we demonstrate that in case of optical chemical sensors drawing inspiration from the connectivity strategy of the olfactory bulb, this problem can find a straightforward solution when an image sensor is used to measure the optical properties of an extended sensing layer. If the sensing layer is formed by a number of spots of different indicators, it is demonstrated that a common data processing can be applied to any replica of the sensing layer even if the indicators are spotted with different geometries and in different quantities.

Gas sensors often suffer from signal drift and long response and recovery times, giving rise to problems to evaluate the steady state gas response. It is shown that these problems can be reduced by ...

Polymers are commonly used in optical chemical sensors as inert supports for indicator dyes keeping the molecules apart to allow for exposure to analytes and to maintain the peculiar optical properties of the sensitive dyes. However, the partitioning properties of polymers may play a significant role in determining the sensitivity of the sensors. Herewith, the properties of a single dye blended with different polymers have been studied through optical absorbance changes elicited by the exposure to organic vapors. Results show that the partitioning properties of polymers modulate the response of the dye and confer a combinatorial selectivity to the system, allowing for the recognition of vapors.

Electronic noses are object of research since more than two decades; nonetheless, the analogies between natural and olfaction systems are still limited to the selectivity properties of the receptors. The implementation of more sophisticated features such as the large number of receptors and the glomeruli layer have been hampered by technical difficulties related to the management of a large number of contemporaneous signals. As demonstrated in the past, optical imaging is a read-out technique for sensors development that can provide large sensor arrays. In this paper an artificial olfaction system based on the imaging of a continuous layer of chemical indicators is illustrated. The system incorporates an array of thousands of sensors, corresponding to the pixels of the image. The choice of Computer Screen Photoassisted Technology as a platform for optical interrogation of the sensing layer allows for the definition of a strategy for an automatic definition of a glomeruli layer based on the classification of the optical fingerprints of the image pixels. Chemical indicators were dissolved into a polymeric matrix providing the further property of odor diffusion mimicking the functions of the olfactory mucosa. The system has been tested in a simple experiment and data have been treated applying a lateral inhibition to the glomeruli layer resulting in a dynamic pattern resembling that observed in natural olfaction.

Rapid quantitative determinations of creatinine, potassium and glucose, all important parameters in routine medical diagnostics, are demonstrated using a computer screen photo-assisted technique (CSPT). CSPT uses regular computer screens as light sources and web cameras as imaging detectors for supporting optical evaluations of diverse character. The ubiquity and versatility of the setup, makes CSPT an attractive candidate for point of care determinations. A robust procedure for the automatic selection of experimental conditions in CSPT evaluation, including camera channels and illuminating colors that minimize the measuring time up to 5 times is described. The method utilizes the correlation between experimental conditions and classification scores, obtained under a generic 50 color training session, to extract measuring sequences as short as 9 s. The extracted measuring conditions automatically adapted to the different tests provided a general, practical and fast approach for CSPT optimization in real scenarios.

A biological olfaction system is one of the highest performance chemical sensing systems for low-molecular-weight compounds. An ability of recognizing a various molecular properties based on their structures is essential for a biological odor recognition process. For the development of a sensor mimicking the olfactory system, we have formed benzene-patterned self-assembled monolayer (SAM) on the sensor surface using SAMs with nanostructure that serves as adsorption sites for benzene ring, which is representative substructure of aromatics. The structure of the benzene-patterned SAM surface was analyzed by contact-angle measurement, ellipsometry, cyclic voltammetry (CV) and atomic force microscopy (AFM). Moreover, the electrodes evaluated as sensor surfaces with cyclic surface-polarization impedance (cSPI) sensor that measures changes in impedance of the electrode surface by adsorption of substances. The results of cSPI indicated the benzene-patterned SAM surface could distinguish benzene by recognizing of the hydrophobicity and the molecular size.

The computer screen photo-assisted technique (CSPT) is a measuring method that uses regular computer screens as controlled light sources and web cameras as imaging detectors for optical evaluations of sensing substances, which optical properties respond to target analytes. The fact, that CSPT uses a globally distributed and familiar infrastructure as instrumentation and that is able to serve multiple targets with the same setup, makes it an attractive candidate for home testing. This article reviews the CSPT concept and focuses in two recent studies that demonstrate thorough CSPT determinations using a handy and eventually disposable assay format, with the potential to be as ubiquitous as the CSPT platforms themselves. The implications of such combination of technologies is discussed, and its potential to support the global monitoring of environmental and sanitary parameters, using Internet geography browsers, is analyzed.

The present contribution contains an overview of the development of gas sensitive field-effect devices in Linkoping during the last 25 years. It is completely centred to the work at the Laboratory ...

The computer screen photo-assisted technique (CSPT) is a method for the classification of colorimetric assays utilizing ordinary computer sets and web cameras as instrumentation. In CSPT measurements the web camera captures the image of the assay under the screen illumination, and typically a spurious spatial distribution of intensities is overlapped on the image. This issue is examined here, focusing on the effect of the sample and illuminating colors on the spatial modulation of intensity. A method for the selection of colors composing an illuminating sequence that minimizes the spatial variability is proposed. The approach is tested for the classification of different color substances showing improvements up to 53% of the intra/inter cluster distance ratio measured in a PCA space, when compared to randomly chosen colors.

An investigation to obtain reproducible measurements with a pulse voltammetric electronic tongue has lead to the development of self polishing device. A grit paper covered bar rotating over the wor ...

Computer screen photo-asssited detection of complementary DNA strands using a luminescent zwitterionic polythiophene derivative

The concept of electronic tongues or taste sensors has developed rapidly during recent years due to their large potential. They are based on electrochemical sensors combined with multivariate data analysis. Voltammetric electronic tongues have proven valuable in many applications. Due to their ruggedness and simplicity, they have been found especially suitable for on-line monitoring of industrial processes. A voltammetric electronic tongue, specially designed for use in the dairy industry is described. It consisted of four working electrodes (gold, platinum, rhodium and stainless steel), embedded in PEEK (TM). It was mounted in a housing of stainless steel, which was inserted in the process line for direct on-line measurements. The voltammetric electronic tongue was used to follow different sources of milk coming into the process and to monitor the cleaning process. (c) 2005 Elsevier B.V. All rights reserved.

The separation of emission from transmitted light for the fingerprinting of fluorescent substances using a computer screen photo-assisted technique (CSPT) is demonstrated. CSPT is a technique for optical evaluation using a simple cell with just a standard computer set and a web camera as instrumentation. It has been demonstrated to be a versatile system for colorimetric and fluorescent fingerprinting. Here the omnidirectional property of fluorescent emission is utilized to separate it from the background, using a simple optical arrangement compatible with CSPT purposes. This enhances the classification capabilities and makes classification at sub-μM concentrations possible.

A study that investigates if it is possible to discriminate between the different rinses in a household washing machine with a voltammetric electronic tongue is concluded. The voltammetric electronic tongue applies a potential pulse train over two electrodes and measures the produced current. Multivariate data analysis is used to treat the data. In this paper, a simplified electronic tongue, with only 5% of the original current responses, is used. The rinses from 20 machine wash runs with four different prerequisites are investigated. Principal component analysis (PCA) and Soft-independent modelling of class analogy (SIMCA) are used in order to classify the rinses. In PCA, only one of the rinses is classified erroneous, and in SIMCA none of the rinses are classified only to the wrong class, although 38% of the rinses are classitied to more than one class. (c) 2005 Elsevier B.V. All rights reserved.

The influence of gate bias on the gas sensing properties of SiC-based field effect transistors with catalytic gate and a buried short channel has been studied. The drain current–voltage (Id–VD) characteristics of the device reveal non-saturation property, which is a consequence of the short channel design. The drain current is larger in hydrogen ambient than in oxygen ambient at the same drain voltage. The threshold voltage decreases with increasing positive gate bias, and increases with increasing negative gate bias. When a positive bias is applied to the gate, the Id–VD characteristics reveal a tendency to saturate. A positive gate bias increases the drain voltage response to hydrogen, as compared with a negative applied gate bias. However, a positive gate bias decreases the stability of the device signal. A change in the channel resistivity is the main reason for the change in the electrical properties when a positive gate bias is applied. A physical model that explains the influence of the gate bias has been studied, and the behavior of the barrier height in the channel was estimated by using the temperature dependence of the Id–VD characteristics.

The characterization of opaque color samples using reflected light is demonstrated with a computer screen photo-assisted technique (CSPT). CSPT utilizes standard computer screens as light sources and conventional web cameras as detectors to perform colorimetric or fluorescent evaluations that retain key spectral features of the tested substances. In this work, the fingerprinting of samples spectral reflectances is achieved by evaluating the reflected light under controlled computer screen illumination. Two different methods for processing of features are investigated, revealing complementary properties for classifications purposes. The attainable classification enables to distinguish different colors from just hues of a same color.

Fingerprinting of fluorescent substances for diagnostic purposes using computer screen illumination

The ability of the computer screen photo-assisted technique (CSPT) to evaluate color samples through its light transmittance profiles is demonstrated by using liquid crystal displays (LCD) as light sources. Transmittance profiles are more drastically affected by angular dependencies than those obtained with CRT illumination. A local light reference computation is introduced to counteract this effect yielding equivalent profiles for both kinds of displays thus allowing for a broader universe of platforms to perform CSPT experiments, including mobile devices.

A scanning light pulse technique (SLPT) operating in a totally unsupervised way suitable for chemical sensing and the efficient screening of new sensing materials is demonstrated. The procedure aut ...

Colorimetric evaluation of samples sorted in a two-dimensional (2D) array is performed with a computer screen used as a controlled light source and a web camera as detector. The transmittance profiles are predicted with a 12 parameters model within errors of 5 or 10% for qualitative or semi-quantitative purposes, respectively. The use of random colour illumination additionally enables robust acquisition with only 50 snap-shots (in less than 1min) that contains the information of the whole array, independently of the number of samples. © 2003 Published by Elsevier B.V.

Distinctive two-dimensional chemical images of hydrogen, ammonia and nitrogen dioxide have been obtained using a scanning light pulse technique (SLPT) on a metal-oxide-semiconductor (MOS) device. The sensor consists of an MOS capacitor with two continuous metal gates of Pd (40nm) and Au (50nm) arranged in a concurrent geometry, leaving a bare silicon dioxide region in between the biased gates. The present arrangement provides a new robust alternative to the existing strategies used to induce spatially distributed chemical responses, required for SLPT and allows to measure the response in a different place than where the biasing electrodes operate. Accordingly, the splitting of biasing and detection coordinates enables to optimize electrical and chemical functions separately providing a higher versatility in the choice of functional sensing materials or stable biasing electrodes. © 2003 Elsevier B.V. All rights reserved.

As a consequence of the formation of water droplets in the car engine at cold start, the fragile ZrO 2 λ sensor cannot be heated until the engine is sufficiently warm. A possibility to shorten the time before closed loop λ control would decrease the exhaust emission. As a solution to this problem, the metal insulator silicon carbide field effect transistor (MISiCFET) sensor, which presumably is more thermo shock resistant than the ZrO 2 sensor, could be used at cold start. The requirements for a cold start sensor are, among others, sensitivity to λ (air to fuel ratio) close to stochiometry, selectivity to λ and high speed of response. In this communication, the possibility of using the MISiCFET sensor at cold start is treated. The sensor consists of a SiC based MOSFET device with a buried channel design and a catalytic gate metal of 10 nm TaSi x and 100 nm Pt. The response depends linearly on λ at 500 °C. The sensitivity of the device has been tested both in artificial atmospheres and in an engine. Two-level factorial designed experiments showed a high selectivity to λ compared to other gases such as CO, hydrocarbons, NO x and H 2 . The response time was found to be

This communication presents a modulated mode of operation for MOSFET gas sensors. A low-power micromachined device allows pulsing the temperature of MOSFET gas sensors with a time constant less than 100ms. Modulating the temperature during the gas exposure modifies the kinetics of the gas reactions with the sensing film. The way the sensor response is modified by the temperature modulation depends on the sensor "history", on the nature of the surrounding gaseous atmosphere, and on the type of materials used as catalytic sensing film. Pulsing the temperature up just after the gas exposure can reduce the recovery time for specific applications, such as for hydrogen detection. Cycling the temperature can allow the discrimination between different gas mixtures. Discrimination was shown for gaseous mixtures of hydrogen and ammonia in air. The results obtained indicate that a "smart" combination of sample and temperature profile could be used to expand the information content in the sensor response. © 2003 Elsevier Science B.V. All rights reserved.

The metal oxide silicon carbide field effect transistor (MISiCFET) sensor has several possible car engine applications, such as an ammonia sensor in selective catalytic reduction (SCR) systems or as a lambda-sensitive device for enhancing catalytic converter efficiency. Both these applications involve closed loop control of the engine and thereby require fast sensors, that is why it is important to investigate the speed of response of the devices. The sensor consists of a SiC-based MOSFET device with a buried channel design and a catalytic gate metal, which makes it sensitive to a wide range of different gases. The selectivity and sensitivity of the sensor to a specific gas depends mainly on the choice of gate metal, its structure and the operating temperature. In this presentation, the speed of response of MISiCFET devices with many different gate metals at several operating temperatures are compared. The tests have been performed in the laboratory using the moving gas outlet (MGO) equipment. The equipment allows two gas outlets to move back and forth under the sensor, which makes it possible to change the atmosphere surrounding the sensor from synthetic air to the test gas quickly. The method is verified by changing the temperature of the device and frequency of the moving gas outlets. The test gas is either ammonia or hydrogen. The time constant of the sensors is shown to be very small

The interaction of chemical species with molecular films of porphyrins causes variations of the work function of the film itself, as it has been recently demonstrated by using the Kelvin probe technique. This characteristic makes porphyrins films suitable to be used as sensitive layers in ChemFET sensors. In this paper, we present a preliminary report about the fabrication and testing of such gas sensitive devices. The technological solutions towards an optimised device are also illustrated and discussed. © 2001 Elsevier Science B.V.

Future legislations for car emissions make direct measurements in exhaust gases of hydrocarbon (HC) as well as CO and NOx interesting. Robust sensors that can stand the high temperature and rough environment in the exhaust gases are needed. Silicon carbide has the advantage of being a chemically very inert material, which, due to its high band gap, is a semiconductor even at temperatures around 800°C. Catalytic metal insulator silicon carbide Schottky diode sensors respond to gases like H2, HC, NOx in exhaust gases. The choice of catalytic metal, structure of the metal, and the operation temperature determines the response pattern to different gases. Here we will demonstrate that an array of different MISiC sensors to some extent predicts the HC concentration in gasoline exhaust gases. Chemometric methods are used for the evaluation of the signals. © 2001 Elsevier Science B.V.

Field effect chemical sensors, utilising silicon carbide as semiconductor, can be operated at high temperature and in rough environments. Gas sensitive field effect transistors, MISiCFET, are now d ...

A new method for the evaluation of fast gas sensors is described. By using moving gas outlets, we can quickly change the ambient around a sensor. Different platinum-insulator-silicon carbide (MISiC) structures are investigated. Their sensor response contains fast components, which respond within milliseconds to a change in the ambient from a reducing gas to an oxidising gas and vice versa. Cylinder specific monitoring of car engines with the MISiC structures is discussed.

Catalytic metal insulator silicon carbide, MISiC, Schottky diodes are promising devices for on board exhaust diagnosis in cars. These sensors show a direct or indirect sensitivity to gases like H2, CO, HC (hydrocarbons) and O2. The catalytic reactivity of the sensor will effect the gas sensing conditions. In some situations knowledge about the reactivity of the catalytic surface may give more information about the exhaust gas composition. For instance, the sensor signal normally moves to a lower voltage in an ambient containing H2 and HC, however, under certain conditions when exposed to rich gas mixtures, the HC response is opposite the one for H2. Measurements performed by the MISiC sensors on simulated exhaust gas mixtures, either rich or lean, are shown here. Some fundamental studies of the HC response have been performed. Reaction limitation conditions are suggested as an explanation for the response of HC opposite the one of H2.

The combination of an electronic tongue and an electronic nose for classification is described. The `electronic nose' consists of an array of gas sensors with different selectivity patterns, signal handling and a sensor signal pattern recognition and decision strategy. The `electronic tongue', which was developed for the taste analysis of liquids is based on pulsed voltammetry. Measurement data from the artificial smell and taste sensors are used to produce sensor-specific opinions about these two human-like sensing modalities. Using principal component analysis, it is shown that both the electronic nose and the electronic tongue alone are able to discriminate reasonably between experimental samples. When combining information from both the electronic nose and the electronic tongue, however, the classification properties are clearly improved.

A distributed chemical sensor system consists of a flow cell with a catalyst and a number of gas sensors mounted along the flow direction in the proximity of the catalyst. In such a system, the catalytic combustion is independent of the chemical sensing. The reaction rate on the catalyst varies for different molecules, which might be utilized for increasing the information about the gas mixture. In this work, the properties of such a system were studied, especially regarding the possibility of gas identification/quantification. It was shown that different molecules gave different response patterns from the sensors in the cell. The components in a binary mixture, hydrogen, and ethanol, could be quantified applying the sensor responses from the distributed chemical sensor system as inputs to an artificial neural network. It was thus found that by employing a number of gas sensors in combination with a catalytic surface, the ability to discriminate between different molecules could be increased.

The effects of different components in simulated car exhaust gases on silicon carbide based field effect sensors are studied using a two-level factorial design. Strong effects are observed for H-2, ...

High temperature sensors, Schottky diodes and capacitors, based on catalytic metal-oxide-silicon carbide devices are investigated. Reducing gases like hydrogen and other hydrogen containing gases, ...

When a gas mixture is passed over a catalytic surface its composition will change due to the catalytic activity of the surface. Different molecules have different reaction rates and this fact can be used in combination with distributed chemical sensors to increase the discriminating capability of the sensors. In a measurement cell (palladium covered or non-covered) a row of identical Pd-MOSFET sensors was mounted along the flow direction. The catalytically active measurement cell changed the concentration of detected molecules (hydrogen or ethanol) considerably compared to the non-catalytic cell. It was observed that even catalytic combustion in the lid 1 mm above the sensor row influenced the sensor responses significantly. When comparing hydrogen and ethanol it was found that the two molecules were influenced to different degrees when passing the catalytic surfaces. Hydrogen was consumed at a much faster rate, which could be seen when comparing the sensor responses along the row. A distributed sensor system thus shows an increased discriminating power. It was also found that chemical sensing was much more temperature dependent for ethanol than for hydrogen.

A halothane vapor sensor based on hexamethyldisiloxane (HMDSO) plasma polymer films have been studied using an ellipsometric technique. The sensor signals, i.e. the ellipsometric parameters ψ and Δ of a HMDSO film change reversibly upon exposure to halothane vapors. The sensor shows very good repeatability and stability. Our study also revealed that the sensitivity and response time depend strongly on the film deposition conditions and a film with a high refractive index has low sensitivity to halothane and longer response time.

Field effect devices based on catalytic metal-oxide-silicon carbide (MOSiC) structures can be used as high temperature gas sensors. The devices are sensitive to hydrocarbons and hydrogen and can be ...

Sensor drift is addressed as one of the most serious impairments afflicting chemical and biochemical sensors. One possible solution to this problem is to view sensor arrays as time-varying dynamic systems, whose variations have to be tracked by adaptive estimation algorithms. A theory of hidden variable dynamics for the rejection of common mode drifting of sensors has previously been developed and is here coupled with a recursive least squares algorithm. In Section 7 a smart solution is provided for a difficult vapour recognition problem vexed by drift that have failed with traditional pattern recognition techniques. Among the many advantages we distinguish that model adaptation to changes in the sensor array makes lifelong calibration possible without interrupting the operation of the array.

Most gas sensors do not give stable responses over a long period of time. This drift causes electronic noses to classify gases wrongly after some period of time. In this paper this problem has been studied, and two methods for counteracting drift are proposed. One is based on reference patterns that are updated when new measurements are made, and the other uses models known from system identification theory. The two methods are suited for different experimental situations, and both give a better result than a back-propagation neutral network classifier when they are used under the right conditions.

A brief survey of different approaches to gas sensing with solid state devices is given. It is pointed out that many physical phenomena can be coupled to chemical reactions on the surface of or within the sensing material. One of the new directions in solid state based chemical sensing is then described in some detail. It concerns the development of fast sensors, e.g. for combustion monitoring at high temperatures. This is first exemplified by the development of semiconducting metal oxides, like strontium titanate, for high operating temperatures. Sensors based on silicon carbide devices with catalytic metal platinum) gates are described in ome detail. A general discussion is given about the fundamental limits of the time response of chemical sensors. It is concluded that, from a chemical point of view, time constants much smaller than presently observed experimentally should be possible.

Both artificial neural networks (ANNs) and the abductory induction mechanism (AIM) have been proven to be suitable for the evaluation of signals from chemical sensors with strong interactions of gas components on the catalytically active surface. The algorithms allow the calculation of calibration curves for a multisensor. AIM yields a good mean approximation within a very short time; ANN covers a broader concentration range with an adequate approximation. The signal evaluation of a set of two pellistors and another set of six MOSFETs is used for illustration.

A hybrid intelligent gas array sensor (or electronic nose) has been constructed, which comprises 10 CHEMFET devices, four Taguchi gas sensors (TGS), one infrared CO 2 sensor and a microcomputer in order to examine the odours from five cardboard papers from commercial manufacturers. Four of the papers came from two different production lines of the same Swedish manufacturer; the duplicate paper from each line was processed further to reduce the odour from this packaging material. The fifth paper came from another manufacturer. The sensor array data was screened using both principal component analysis (PCA) and cluster analysis (CA), and predictively classified using a back-propagation neural network. It was discovered using PCA/CA that only four of the 15 sensors were necessary to discriminate totally between the five classes of paper when air, which was initially classified separately, was used as a reference. Thus we have shown that the olfactory quality of cardboard papers can be recognized using a simple hybrid CHEMFET/TGS electronic nose.

Some of the ongoing studies at our laboratory of gas-sensitive field-effect devices with catalytic metal gates are reviewed. More particularly, we discuss the use of such devices in so-called electronic noses due to the possibility of changing the selectivity patterns of the devices by the choice of catalytic metal and operation temperature. Several examples of the application of electronic noses consisting of field-effect devices in combination with metal oxide-based sensors are given. Finally, a summary is given of some remaining scientific problems and studies related to the understanding and development of gas-sensitive field-effect devices.

Field effect sensors based on metal-oxide-silicon carbide (MOSiC) devices are used as high temperature gas sensors. They are sensitive to, for example, saturated hydrocarbons and hydrogen and can be operated up to at least 800 degrees C, which make them suitable for several types of combustion control. A metal gate with two layer platinum and a buffer layer of tantalum silicide in between gave a large increase in the long term stability of the sensors. At temperatures below 600 degrees C, the response to ethane in oxygen was shown to have a threshold at a ratio of about 0.38 for the ethane-to-oxygen concentrations. Below this ratio, the surface can be considered as mainly oxygen covered and the response is small. Above this ratio the metal surface is probably mainly hydrogen covered and the response is considerably larger.

Conducting polymers appear very attractive as sensor materials either as the gas-sensitive component or as a matrix for easy immobilization of a specific substrate. The planar Schottky barrier diod ...

Field-effect devices with a catalytic metal gate are operated as gas sensors over a large temperature range by the use of 6H-silicon carbide (bandgap 2.9 eV) instead of silicon (1.1 eV) as the semi ...