The 8th GOSPEL Workshop. Gas sensors based on semiconducting metal oxides: basic understanding & application fields GOSPEL is an international biannual research meeting (http://www.gospel-network.org, https://www.gospel-network.org/old/) dedicated to the R&D activities in the field of semiconducting metal oxide based gas sensors. The aim is to bring together academia and industry and help the latter to identify which of the new developments are the most relevant to it.
Chair: Eduard Llobet
Application runs the show: what can we learn about the future from the past?
Atomic layer deposition (ALD) is a chemical vapor deposition (CVD) deposition method in which high-quality, fine functional oxide films in the range of 10-1000 nm can be grown. The low deposition temperature and nonobligatory of heat treatment make it possible to obtain fine grains with high surface area. Despite the fact that few reports in recent years highlighted the importance of thickness/microstructure , the role of the defect sites, electronic and surface properties have not been well characterized with varying temperature and oxygen partial pressure , additionally understanding on chemisorbed oxygen species and oxygen vacancies (V_O^(°°)) which facilitates the replenishment of chemisorbed oxygen ions on the surface  are also lacking. This report focus on a correlation of surface electronic, structural and chemical properties with the surface reactivity and sensing behavior of the ALD layers of tungsten oxide (WO3), molybdenum oxide (MoO3) and tin oxide (SnO2) in the thickness range of 20-100 nm by employing advanced synchrotron based surface sensitive spectral-microscopic techniques; near ambient x-ray photoelectron spectroscopy (NAP-XPS), x-ray induced photoelectron emission microscopy (XPEEM), x-ray photoelectron spectroscopy (XPS), ultraviolet photoelectron spectroscopy (UPS) and low energy electron microscopy (LEEM). The characterization work is supported by extensive chemical sensor testing and electrical resistivity measurements.
Depletion layer, dissociative adsorption of oxygen, chemisorption and the continuous replenishment on the surface in the course of catalytic oxidation reactions are prerequisite for an explanation for sensing mechanism, indeed sustained operation for the metal oxide based chemical sensors. This work will focus on the stoichiometry, crystal structure, work-function (Ф), valance band, and surface defect analysis, oxidation state as a function of oxygen partial pressure and temperature and its effect on chemisorbed oxygen ion concentration, thin film processing conditions, microstructure/porosity, depletion layer and Schottky barrier height. Inorganic materials, particularly salts and oxides, are sensitive to photon induced damage, after prolonged high intensity x-ray exposure, on the other hand, it is required high enough intensity and spectral resolution (achievable in synchrotron sources 0.1-0.2 eV) to realize the proposed goals. PM4 with its low dose x-ray ability and SMART with its XPEEM capability end stations, at BESSY II, Berlin and unique NAP-XPS system at Charles University, Prague, enabled spectra-microscopic characterization with abovementioned techniques. The results are enclosed here briefly in Figure 1.
Figure1-a shows the valence spectrum of WO3 thin film extends up to the Fermi level starting after 250°C, indicating a change in the character of the sample surface together with electrical resistivity measurements. Figure 1-b shows bright-field image of the 700 °C deposited WO3 sample. As the figure reveals the formation of continuous thin film with homogeneous grain structure over the surface differentiating with the grain size as the thin film possesses larger grains and obviously higher surface roughness. In LEEM images contrast is determined by work function (Φ) variation over the surface. Surface topography, grain boundaries, porosity and stoichiometry can affect the work function (Φ). The variation in the work function (Φ) from grain to grain is higher in comparison to 600°C deposited sample. In order to have better understanding on this, XPEEM analysis was conducted (not included here). The Figure 1-c shows a valance band spectrum of defective MoOx taken with 65 eV photons with two different take-off angles. Efermi level followed by valance band minimum, and at larger binding energies, O 2s was seen in the figure. Another feature observable is the 4d and 5s of Mo blended in between 5-12 eV. It is detected that the band gap is not entirely free of states due to the nonstoichiometry of the MoOx thin film. The oxygen vacancies (V_O^(∙∙)) give rise the defect states located around ~3 eV below the Efermi level in 60° in comparison to the 0°. This is attributed to a change in occupancy of the Mo 4d and 5s bands due to reduction from Mo+6/Mo+5 to Mo4+. The valence band minimum is located at about 1.0 eV below the Fermi level.
The performance of a metal oxide gas sensor can be affected by different parameters as the fabrication process of the sensitive layer or the post-annealing treatment. Nevertheless, not only the way the material has been obtained is relevant for the sensor performance. The conditions under which the detection is carried out are of great importance, namely the temperature of the material during detection and the gas flow conditions, among others.
The way the flow arrives at the sensor surface is relevant for the sensor response, mainly for the velocity of the response, but also because it can modify the temperature of the sensor. In the experiments to characterize gas sensors, the position of the gas input and output inside the sensor chamber, as well as the magnitude of the gas flow, influence the way the concentration gradient reaches the sensing material.
Moreover, convective flows are generated inside the chamber due to the sensor itself, because it acts as a heat source. As the convective flows can generate random concentration gradients, the response of the sensor will vary depending on the position of the heat source, namely, of the sensor itself.
In this work, the influence of the sensor position inside the chamber and the gas flow are analysed, comparing experimental results with flow dynamic simulations.
Simulations and experimental
In order to investigate the effects of the position of the sensor and the flow of the gas inside the chamber during the sensing process, three cases were simulated and compared to experimental results:
i. Case I: sensor at the bottom of the test chamber and 200 sccm flow (Fig. 1 (a))
ii. Case II: sensor at the bottom of the test chamber and 400 sccm flow (Fig. 1 (b)).
iii. Case III: sensor in the middle of the test chamber and 400 sccm flow (Fig. 1 (c)).
The temperature of the sensor heater was set at 300 °C and the analyzed pulses were 5 ppm of NO2 for all the simulations and experiments.
Simulation: a finite volume method based on numerical fluid flow analysis was used to obtain decoupled solution for both the Navier-Stokes equations and the mass diffusion equations (convection-diffusion equations) in the absence of reaction-transport equations and involving a high degree of convection-diffusion discretization schemes, applied to solve the combined effects of both convection and diffusion. A mass flow inlet with a mass fraction of NO2 of 7.94418·10−6 (5 ppm of NO2) relative to air was set and atmospheric pressure was imposed in the outlet.
Experiments: sputtered WO3 gas sensor on alumina substrate were used for the experiments. The electrical measurements were performed inside a cylindrical sealed stainless steel chamber with a volume of 0.86 l.
The comparison of the simulations and the experimental results for the three different cases are shown in Fig. 1, where the simulated NO2 concentration arriving at the sensor (black) and the response of the sensor (blue) are plotted in the left and right axis, respectively.
Figure 1 – Simulated NO2 concentration arriving at the sensor surface and response of the sensor for 5 ppm of NO2 at 300 °C for: (a) at the bottom of the chamber and with 400 sccm of flow, (b) at the bottom of the chamber and with 200 sccm of flow and (c) at the middle of the chamber and with 400 sccm of flow.
For case I and II, the simulations fits the experimental data and when the flow is increased from 200 to 400 sccm, the time needed to reach the maximum concentration on the sensor surface also decreases by a factor of 2.
If the sensor is placed in the middle of the chamber and the maximum flow is used (case III), the simulations states that in less than one second more than 4.5 ppm of NO2 arrive at the sensor surface and then increases more slowly up to the maximum concentration introduced in the chamber (5 ppm). By contrast, from the experimental data, the needed time to reach the maximum response is ~ 500 s (the lowest time regarding the three different cases). This substantial difference between the simulation and the experimental results in the case III could be explained by the fact that this type of sensors have a minimum response time to reach the response value. Therefore, although the desired concentration reaches the sensor surface very fast, the sensors would need at least 8 min, in this case, to reach the maximum response, probably due to the time required for the diffusion processes.
Heterogeneous integration of metal oxides – towards a CMOS based multi gas sensor device
Consideration for oxygen adsorption species on SnO2 semiconductor gas sensors
Chair: Jong-Heun Lee
Gas sensor as a device composed of sensing material coupled with signal transducer, has been acknowledged as an analytical tool for detection and quantification of inflammable, explosive or toxic gases. The gas sensors based on nanostructured oxide semiconductor endowed with excellent sensing properties have exhibited great potential application in the fields of environmental monitoring, resource exploration, medical welfare, etc.. It is well known that the sensing mechanism of sensor employing oxide semiconductors is mainly that the interactions between the surface adsorbed oxygen species and target gases lead to a change in the electrical conductivity. Therefore, the gas sensing properties of oxide semiconductors are closely related with their composition, crystalline size, and microstructure. In this regard, design and preparation of oxides with novel architectures will be increasingly important in the construction of high performance gas sensors. Due to high specific surface area, low density, and good surface permeability, porous nanostructures oxide semiconductor sensing materials have attracted growing interest in recent years. In our work, we successfully prepared various porous nanostructures oxides and their composites to the construction of high performances gas sensors with enhanced sensitivity, selectivity, as well as lowered detection limit. The subsequent gas sensing measurements explicitly revealed that these oxides and composites manifested superior sensing behaviors (like much higher sensitivity and faster response speed), which can be ascribed to the porous architectures and the synergistic effects.
Hydrogen sulfide (H2S) is known as a colorless, flammable and highly toxic gas with strong odor. Therefore, to avoid the risk of its leakage, continuous monitoring of H2S is extremely important. Semiconductor gas sensor is promising for such applications because of its high sensitivity, selectivity, fast response, and good stability. It has been reported in many papers that CuO/SnO2 sensor shows a high sensitivity to H2S since the pioneering work by Tamaki et al. 1) It is believed that sulfidation of CuO with H2S degrades the PN junction between CuO and SnO2, resulting in a significant change in electrical resistance. However, the detailed mechanism remains unclear yet. In this study, we fabricated a PN junction film using p-type CuO nanocrystals and n-type ZnO nanorod films to make a sensor device and studied its diode properties under H2S atmosphere to reveal the sensing mechanism. It was expected that the combination of nanorods and nanocrystals increases the area of junction interfaces, leading to a pronounced change in electrical properties upon gas reaction.
ZnO nanorods (NRs) were grown on an ITO-coated glass substrate by a seeding method.2) As seeds, ZnO nanocrystals (NCs) were synthesized by a hot-soap method, in which zinc acetylacetone and 1,2-hexadecanediol were dissolved in oleylamine and heated at 220°C for 90 min. The synthesized ZnO NCs were deposited on an ITO-coated glass and annealed in air at 350°C for 30 min to remove surface capping agents. The ZnO NC film was dipped in an aqueous solution containing Zn(NO3)2 and hexamethylenetetramine (HMT), and then heated at 95°C for 5 h to synthesize ZnO NRs.
Cu2O nanocrystals (NCs) were also synthesized by a hot-soap method using oleylamine as a high-boiling-point solvent. Typically, Cu (II) acetylacetonate and 1,8-octanediol were added to oleylamine in a three-necked flask. The temperature was raised to 160°C under an Ar flow and kept at 160°C for 60 min. The surface ligands, olylamine, were replaced with 3-mercaptopropyl acid by dispersing NCs in a solution containing water, methanol, and NaOH with pH 12. The ligand exchange reaction was carried out for 30 min.
The Cu2O NCs were deposited onto the ZnO nanorod film by drop casting and annealed in air at 250°C for 30 min. Gold electrodes with 100 nm thickness were deposited on the film by thermal evaporation to fabricate a sensor device. IV curves of the device were measured at 50 to 200°C in air and air containing H2S with a Keithly 2400 source meter.
Results and Discussion
Fig. 1 shows a representative SEM image of ZnO NRs deposited on ITO. ZnO NRs grew perpendicular to the substrate to form a porous film. The diameter and the length of the NRs were estimated to be ca. 100 and 700 mm, respectively. Fig. 2 shows a TEM image of CuO/Cu2O NCs. The size of the NCs is 8 to 10 nm with a narrow size distribution. XRD results revealed that Cu2O was converted into CuO after ligand exchange with an alkaline solution.
Fig. 3 shows IV curves of the device, in which CuO NCs were deposited onto ZnO NRs, in air and air containing 8 ppm H2S at 150°C. The device exhibited a clear rectification behavior in air at 150°C, indicating that PN junctions were formed between CuO NCs and ZnO NRs. However, such a clear rectification behavior was lost under H2S atmosphere. The current in the forward and reverse directions significantly reduced after exposure to H2S. We have recently confirmed that CuO NCs-based sensors respond to H2S by an increase in electrical resistance due to reaction of adsorbed oxygen with H2S, representing a typical behavior of p-type semiconductors.3) Thus, the observed increase in resistance for the present device is possibly due to surface reaction of adsorbed oxygen with H2S and the resulting annihilation of holes. The rectification of the current recovered by reintroduction of air, which possibly increased the electron concentration in CuO. On the other hand, at higher temperature, the device showed a simple ohmic behavior (no figure), suggesting that complete conversion of CuO into CuS occurred. Thus, the sensing mechanism should be dependent on temperature. The above results demonstrate the feasibility of nanorod/nanocrystal-based PN-junction devices in constructing gas sensors. Currently, a more detailed mechanism is under investigation.
1) J. Tamaki, T. Maekawa, N. Miura, N. Yamazoe, Sens. Actuators B: Chem., 1992, 3, 197-203.
2) V. Gaddam, R. R. Kumar, M. Parmar, M. M. Nayak, K. Rajanna, RSC Adv., 2015, 5, 89985-89992.
3) K. Mikami, Y. Kido, Y. Akaishi, A. Quitain, T. Kida, Sensors 2019, 19, 211.
The modification of metal oxides by additives or dopants is well known for improving the performances of resulting chemoresistive gas-sensors. Common additives are transition metals, dispersed as ions in the oxide structure, or noble metals, usually present as nanoparticles dispersed among the oxide grains. Noble metals may induce electronic sensitization, based on the modification of the electronic properties of the host oxide, or spillover. As recently summarized: “In the case of the spillover mechanism, the target/analyte molecule is adsorbed onto the noble-metal-oxide cluster which leads to a weakening of its molecular bond. The adsorbate is transferred onto the support material where the reaction takes place. In the Fermi-level pinning mechanism, the gas detection reaction takes place on the surface of the noble metal cluster. The cluster electronically interacts with the base material and the contact pins the Fermi levels of both materials. If the work function of the noble-metal-oxide cluster is changed upon interacting with an analyte gas, the depletion layer in the base material caused by the contact is also affected.”  The same authors also note that clear evidence of these mechanisms is seldom presented, as expected from the complexity of the underlying phenomena and the difficulty of suitable experimental observations. The use of nanocrystalline metal oxides is another feature widely accepted in gas-sensing field as deeply beneficial for the performances of chemoresistive gas-sensors, after the early report by Yamazoe . Advanced synthesis techniques give the opportunity of combining the two approaches, by adding noble metal nanocrystals to nanocrystalline metal oxides, while at the same time retaining the nanocrystalline feature of the host oxide even at high operating temperatures. Nevertheless, if we refer to the above description of the possible interaction mechanisms between the noble metal and the host, we realize that the current description comprises an oxide host whose grain are implicitly assumed to be much larger than the noble metal guest structures. In the present work, on one hand the synthetic approaches will be described resulting in co-existence of nanocrystalline guest oxide and noble metal additive. The Pd-SnO2, Pt-TiO2 and Rh-TiO2 systems will be in particular reviewed. It will be confirmed that noble metal addition can effectively boost the sensor response with respect to the pure oxide. On the other hand, hints will be provided about additional issues with respect to “traditional” systems, for instance: i) possible doping of the metal host by noble metal cations, modifying the electronic properties of the oxide to a more relevant extent with respect to bulk materials; ii) necessity of considering structural modifications induced by noble metals additives on the oxide host. Above all, the necessity will be suggested of overcoming the traditional view of spillover and electronic sensitization, by finding a model capable of introducing a microscopic description of the interaction between the two components.
 A. Staerz, I. Boehme, D. Degler, M. Bahri, D. Doronkin, A. Zimina, H. Brinkmann, S. Herrmann, B. Junker, O. Ersen, J.-D. Grunwaldt, U. Weimar, N. Barsan, Rhodium Oxide Surface-Loaded Gas Sensors, Nanomaterials, 8 (2018) 892.
 N. Yamazoe, Sens. Actuators, B 5 (1991) 7-19.
Hierarchical core-shell (C-S) heterostructures composed of a NiO shell deposited onto the stacked-cup carbon nanotubes (SCCNTs) were synthesized using atomic layer deposition (ALD). By ALD technique, a precisely controlled film of NiO particles was uniformly deposited onto the inner and outer walls of the CNTs. Indeed, by varying the number of ALD cycles from 25 to 700, NiO coating of 0.80 and 21.8 nm were deposited on SCCNTs. The morphological, microstructural and electrical characteristics of the as prepared NiO-SCCNTs C-S nanostructures were thoroughly investigated. The electrical resistance measurements highlighted the large influence of the NiO thickness on increasing of many order of magnitude the baseline resistance of NiO-SCCNTs C-S nanostructures with various thicknesses of the NiO shell layers, suggesting that the conductivity of the sensors is dominated by Schottky barrier junctions across the n(core)-p(shell) interfaces.
The behavior of NiO-SCCNTs sensors was investigated for low concentrations of volatile organic compounds (VOCs) such as ethanol and acetone. The gas sensing response of the NiO-SCCNTs heterostructures towards acetone and ethanol showed a strong dependence on the thickness of the NiO shell layer. The remarkable performance of NiO-SCCNTs sensors benefits from the conformal coating by ALD, large surface area by SCCNTs and the optimized p-NiO shell layer thickness which regulate the modulation of the electron-depletion region in the NiO shell layer. So, optimizing the NiO layer thickness, NiO-SCCNTs sensors display a response about 6 times higher than pristine SCCNTs at the operating temperature of 200 oC. On the basis of the morphological, microstructural and electrical characterization and sensing results, the sensing mechanism which account for the marked variation in the resistance during the interaction of the target gas molecules has been here discussed.
Reliable and real-time measurements of gaseous pollutants (e.g., NO2 and CO) in both indoor and outdoor air are required to implement worldwide air quality legislation designed to protect human health and the environment. While electrochemical sensors are popular for air quality monitoring due to their fast and linear response, low power consumption and excellent selectivity, some applications operate in environments that extend beyond their capability. Gas sensors based on semiconducting metal oxides (MOx) technology offer advantages such as high sensitivity, low manufacturing cost, miniaturization potential and long lifetime. Commercially available MOx sensors are typically based on n-type SnO2, WO3 or versions thereof modified by the presence of precious metal catalysts such as Pt or Pd. The shortcomings of these materials i.e. baseline drift, humidity interference and cross-sensitivity to nuisance gases, are well-known. Moreover, exposure to oxidizing and reducing gases have reverse effects on a MOx electrical conductance, governed by its semiconducting characteristics. This introduces a key challenge in interpreting the response of a single MOx sensor exposed to a mixture of oxidizing and reducing gases.
To address the aforementioned shortcomings, Alphasense and partners have adopted a “Multi-MOx” array approach, where p-type and n-type metal oxide sensing elements are combined on a single ceramic chip. A Platinum heater on the underside of the chip heats the sensor to the desired operating temperature. The sensor discussed here is comprised of:
1. p-type CTO (titanium-doped chromium trioxide), a ternary oxide which provides a stable baseline, minimal humidity interference and high sensitivity to reducing gases such as CO, and
2. n-type WO3, a binary oxide with excellent sensitivity to oxidizing gases such as NO2 and O3.
In the case of p-type CTO, exposure to CO causes a decrease in the charge carrier (hole) concentration in the near-surface region and a decrease in the measured conductance. Whereas, the measured resistance of n-type WO3 increases in exposure to NO2 due to an increase in the density of charge carriers (electrons) trapped at the oxide surface (see Fig. 1). The use of different MOx materials in conjunction with operating temperature modulation and advanced on-chip filtering can be used to reliably measure both oxidizing and reducing gases. This report summarizes our recent work on the p-type/n-type Multi-MOx gas sensor platform.
Chair: Kuniyuki Izawa
NH3 is an irritant gas with a unique pungent odor; sub-ppm-level breath ammonia is a medical biomarker for kidney disorders and Helicobacter pylori (H. pylori) bacteria-induced stomach infections.[1,2] The humidity varies both in an ambient environment and exhaled breath and thus humidity dependence of gas sensing characteristics is a great obstacle for real-time applications. Herein, flexible, humidity-independent, and room temperature ammonia sensors are fabricated by the thermal evaporation of CuBr on a polyimide substrate and subsequent coating of a nano-scale moisture-blocking CeO2 overlayer by electron-beam evaporation.
CuBr sensors coated with a 100 nm-thick CeO2 overlayer exhibits an ultrahigh response (resistance ratio) of 68 to 5 ppm ammonia with excellent gas selectivity, rapid response, reversibility, and humidity-independent sensing characteristics at room temperature. In addition, the sensing performance remains stable after repetitive bending and long-term operation. Moreover, the sensors exhibit significant response to the simulated exhaled breath of patients with H. pylori infection; the simulated breath contains 50 parts per billion (ppb) NH3. The sensors thus show promising potential in detecting sub-ppm-level NH3, regardless of humidity fluctuations, which can open up new applications in wearable devices for in situ medical diagnosis and indoor/outdoor environment monitoring.
The baseline resistance and gas response remained nearly the same in different bending modes which is very suitable for wearable devices.
The CeO2 layer plays the role of blocking the interaction between moisture and CuBr.
 Di Natale, C.; Paolesse, R.; Martinelli, E.; Capuano, R. Solid-State Gas Sensors for Breath Analysis: A review. Anal. Chim. Acta 2014, 824, 1-17.
 Krishnan, S. T.; Devadhasan, J. P.; Kim, S. Recent Analytical Approaches to Detect Exhaled Breath Ammonia with Special Reference to Renal Patients. Anal. Bioanal. Chem. 2017, 409, 21-31.
We present flexible chemo-resistive sensors based on AACVD grown tungsten trioxide (WO3) nanowires. The sensor response to gases, before and after a 50-cycle bending test, is reported. Thus, proving that reliable gas sensors, able to withstand repeated bending, have been achieved. Moreover, their integrity and durability have been tested under harsh bending conditions until break down.
Flexible sensors are a promising technology for personal environmental monitoring, sports or healthcare and medicine. These applications involve using wearables, which are attached to the body or clothes to sense different variables, including gases. The use of flexible substrates improves the performance in these systems . However, usually, little or no information is given on reliability. Here, metal oxide nanowire gas sensors over polymeric foil, have been developed and their performance after repeated mechanical stress test has been evaluated.
Materials and Methods
The sensor architecture consists of one electrode and one coplanar heater over a polymeric substrate (Kapton® 50.5 µm thick) as used in . There is a layer of WO3 nanowires coating the active area (Fig 1). Electrode and heater patterns were stenciled using silver ink as reported in . The WO3 layer was grown directly on the flexible substrate via an aerosol assisted chemical vapor deposition in a hot (350 ºC) wall reactor. Precursors used were tungsten hexacarbonyl dissolved in a mixture of acetone and methanol, as reported in .
Figure. 1: Device layout. Left: electrode and heater. Right: active layer over both elements. Device size 14×14 mm2. Figure 2: Universal testing machine grips holding the sensor. d = maximum deflection distance.
We made a controlled bending test using an electromechanical universal testing machine (Shimadzu AGS-X 10 kN). Sensors were strained up to 15% and the maximum deflection due to buckling was d=3.23 mm. The test consisted of 50 continuous moves of the upper grip. Each move was 2 mm down and up, at 20 mm/min, producing a curvature radius of approx. 3.1 mm (Fig. 2). Meanwhile, the electrical resistance of the active layer, the stroke and the force applied were measured.
The AACVD process resulted in the direct growth onto the transducer of WO3 nanowires (150 nm in diameter and 10 microns in length). Sensors were tested against H2 before and after the bending test. In both cases, three cycles of three H2 concentrations (250, 500 and 750 ppm) were tested, at a heater mean temperature of 150 ºC. Sensor response after the bending test (Fig. 3) shows small changes (after bending test baseline electrical resistance increased by 5.5 % Fig. 4, which can be easily calibrated). An additional test was carried out until the sensing layer was damaged: the electrical resistance increased significantly at zero stroke. This occurred after 200 bending cycles: 100 under compressive strain and 100 under tension strain, with a shift of 4 mm and d=4.3 mm of maximum deflection (curvature radius 3.2 mm).
Figure 3: a) Current through active layer at different concentrations of H2 (250, 500 and 750 ppm H2) and b) Sensor response to H2 concentration, before and after 50 repeated bending tests
Figure 4: Percentage of sensor resistance increase during 50-cycles bending test.
We have proved that is possible to produce reliable flexible sensors with a very affordable technology. After a 50-cycles bending test, under tensile strain, sensor response remains almost unchanged. Moreover, the limits of the physical system have been tested under harsh bending conditions. The sensors could withstand up to 200 bending cycles before losing functionality. This is being developed further via the design of in-house made metal oxide inks for achieving fully printed functional devices. Characterization results will be presented at the conference.
 A. Nag, et alt., IEEE Sensors Journal , 17, 3949, 2017.
 J. L. Ramírez, et al. Sensors Actuators, B Chem., vol. 258, pp. 952–960, 2018.
 M. Alvarado, et al. Sensors, vol. 18, no. 4, p. 999, 2018.
 S. Vallejos, et al. J. Nanosci. Nanotechnol., vol. 11, no. 9, pp. 8214–8220, 2011.
By laser micromilling technology it is possible to fabricate custom MEMS microhotplate platform and also SMD package for MOX sensor, that gives complete solution for integration in mobile devices - smart phones, tablets and etc. The 3D design and fabrication of MEMS microhotplates and packages products occurs simultaneously that give opportunity for ultra-fast time making unique solutions for MOX sensors (number of microhotplates, hot spot size on microhotplates, diameter holes in package cap and etc.) without looking at standard solutions (primarily the package type).
The main idea of our developed technological flow based on laser micromilling is wide flexibility in developing of MEMS and SMD structures. Using of equipments only widely presented on the market and refusing of technological steps needs a clean rooms support. Only semi custom 3D printing type software is especially developed product for laser micromilling system needed for successful development and production of MEMS and SMD structure during our experiments. Software is needed for translation CNC code to 4-axis laser micromilling setup and online measuring of geometrical parameters of MEMS and SMD structure for corrections micromilling procedure during automatic production.
During our work, an Ytterbium pulsing 20 W fiber laser with a wavelength of 1.064 μm and tunable pulse duration from 50 to 200 ns is used. This laser emitter is installed on the four-coordinate portal complex, which allows the laser scanner to be moved across wilde field. The processing of ceramic substrates is carried out in a snap-in fixed in a rotational device, which allows processing of flat substrates on both sides, cylindrical substrates over the entire surface area. Currently, fiber markers are most often used in industry for marking various types of products and are not intended for 3D laser milling, despite the fact that the technical capabilities of any laser marker allow it by using of our developed software. Fabrication of MOX sensors includes the following main steps:
• MEMS microhotplate modeling and both bottom and top parts of the SMD package (Fig.1) in 3D CAD programs with output file in STL format and also 2D modeling of MEMS and SMD metallization topology in DXF format;
• Optionally MEMS microhotplate parameters could be simulated in COMSOL program, which allows to predict approximate thermal characteristics of the MOX sensor;
• 4-axis laser facility is used for monolithic ceramics laser micromilling with help of 3D models of bottom and top parts of the SMD package and MEMS microhotplate;
• Platinum metallization deposition according with 2D model of topology, metallization annealing paying attention to specification on jet or screen-print platinum materials (Fig.3);
• Optionally the metallization can be processed with laser according with 2D model;
• MOX gas sensitive layer deposition and annealing on the MEMS microhotplate;
• Assembling separate parts of sensor into one SMD package and adhesion with special glass (Fig.2)
Using described tech flow, experiments were carried out to fabricate a possible minimum size of MEMS microhotplate from Al2O3 ceramics. The minimum size of the manufactured microhotplate with 250 mW power consumption at 450°C with track width was 30 μm and 20 μm thickness in SMD SOT-23 package type (3.0x1.4x1.0 mm with max dissipating power at 20°C - 350 mW) were achieved. Tests of fabricated MEMS microhotplate present in work 1.
Advantage of ceramic using as a material for laser micromilling is extension of MOX sensor working temperatures range up to 1000°C compare with typical 700°C for silicon technology. Also useful advantage of fully ceramics based MOX sensor is long term stability against harsh environmental conditions including extreme temperature and acid or alkaline gases.
This research was funded by Ministry of Science and Higher Education of Russian Federation under grant number 14.584.21.0030 from 22.11.2017, unique identifier RFMEFI58417X0030.
1 N. Samotaev, K. Oblov, and A. Ivanova, “Laser micromilling technology as a key for rapid prototyping SMD ceramic MEMS devices”, MATEC Web of Conferences, 2018, vol. 207, article number 040034
Chair: Kengo Shimanoe
Luminescence Probing of Surface Adsorption Processes Using InGaN/GaN Nanowire Heterostructure Arrays
Adsorption phenomena lie at the heart of understanding semiconductor gas sensors. The most widely studied kind of gas sensors are metal oxide sensors which respond via resistance changes to changes in the ambient gas concentration. A key problem in the analysis of gas sensor behavior is that the effects of adsorption are not readily reflected in the experimentally observable resistive gas response as the electronic transport through such porous and nanocrystalline sensing layers can depend in manifold ways on the stoichiometry and the morphologies of the sensing layers. In an attempt at providing a much more direct view onto the effects of gas adsorption on semiconductor surfaces we have studied the photo-luminescence (PL) response of InGaN/GaN nanowire arrays (NWA) with naturally oxidised surfaces. Such NWAs exhibit an efficient photoluminescence which persists to temperatures of 200°C and beyond, which additionally exhibits chemical sensitivity as the NWAs are exposed to gases or liquids [1-5]. Gas sensing tests on such arrays reveal both quenching and enhancing PL responses, which indicates that the native PL response of such NWAs is controlled by native defects at the nanowire surfaces which become modified as reactive gases adsorb on these defects. A very interesting observation is that the concentration dependence of the PL response of all kinds of analytes investigated so far (O2, NO2, O3, H2O, EtOH) is that it consistently follows Langmuir-type isotherms which are easy to interpret regarding adsorbate-specific adsorption energies. A surprising finding is that best-fitting adsorption energies do not reveal as species-dependent constants but rather as linear functions of temperature with species-dependent slopes. We show that this behaviour can be explained by a competition of test gases and background gases for common adsorption sites on the NWA surfaces. A particularly interesting gas species is water vapour, which naturally forms quenching adsorbates which transform into enhancing ones as water vapor exposures are prolonged under conditions of moderate surface temperature and under intense UV light illumination. Reducing gases such as H2, aliphatic hydrocarbons and alcohols do not exhibit any intrinsic ability of modifying the PL yield. Such species, rather, seem to be detected in an indirect manner by consuming quenching oxygen adsorbates and by forming enhancing H2O ones as these interact with oxygen species co-adsorbed in reactive backgrounds of ambient or synthetic air. Regarding ongoing research, a very promising aspect is that PL probing of surface adsorption phenomena forms a natural complement to existing technologies of in-operando spectroscopies which are increasingly employed in gas sensor research. Such research is likely to reveal microscopic information on surface adsorption processes which is important for optimizing the operation of gas sensors, catalysts and photoelectrochemical solar cells.
Keywords: III-nitride semiconductors, nanowires, photoluminescence, gas adsorption
Understanding the mode of operation of metal-oxide gas sensors (e.g. SnO2, In2O3) is of great scientific and economic interest. Such a knowledge based approach requires the development and application of spectroscopic tools to monitor the relevant surface and bulk processes under working conditions (operando approach). In this contribution, we will present recent operando results on ethanol and CO detection using undoped and Ag doped In2O3 gas sensors, demonstrating the advantages of (i) operando Surface Enhanced Raman Spectroscopy (SERS) to monitor the metal oxidation state, and (ii) extending our operando Raman / gas phase FT-IR setup by UV-Vis spectroscopy to reveal the degree of In2O3 reduction.
Aim of this work is to compare the electrical responses to 100-400ppb NO2 gas concentrations of WO3 electrospun nanofibers both activated by thermal (in the temperature range 25-100°C) and/or visible light at different wavelengths (Red λ=670 nm, Green λ=550 nm, and Purple-Blue λ=430 nm). WO3 nanofibers were prepared by mixing a W-O sol-gel transparent solution with a polymeric solution made of PVP and DMF, electospun and subsequently annealed at 450°C. Regarding gas sensing measurements, Purple Blue light resulted the most effective light source as respect to the others. Light illumination at room temperature revealed to improve both base line recovery and response time, whereas temperature enhances relative response, with a maximum at 75°C. Light-radiating room temperature gas detection yields a satisfactory response notwithstanding a slight reduction of sensor gas sensitivity. Light induced electrical response mechanisms is presented and discussed.
Chair: Geyu Lu
The full abstract is in the attachments.
Oxide Semiconductor Gas Sensors with Nanoscale Catalytic Overlayer: Toward Highly Selective and Sensitive Gas Detection using Bilayer Design
Jong-Heun Lee1, Seong-Yong Jeong1, and Hyeon-Mook Jeong1
1 Department of Materials Science and Engineering, Korea University, Seoul 02841, Republic of Korea
Oxide semiconductor gas sensors have been widely used to detect hazardous gases and their applications are being expanded to wide range of applications including disease diagnosis, indoor air quality monitoring, fruit quality control, and the discrimination of different smells/odours. Although various noble metals or oxide catalysts have been loaded or doped to oxide semiconductor chemiresistors in order to enhance the gas selectivity and sensitivity, gas sensing materials are limited and their sensing characteristics are still insufficient to cover numerous chemicals. Moreover, the catalyst loading often leads to the significant increase of sensor resistance by oxygen spillover or charge transfer from sensing to catalytic materials, which hampers the measurement of sensor resistance using conventional electronic circuit. In the literature, relatively thick configuration of catalytic overlayers have ever been coated on the sensing film in order to filter interference gases. However, the enhancement of response to less reactive analyte gases via gas reforming using nanoscale catalytic overlayer has been barely investigated. In this presentation, we will demonstrate that the bilayer sensor design can provide not only the separation of catalytic and sensing reaction but also the tuning of both gas selectivity and sensitivity. For instance, two different bilayer sensor designs, Pd-SnO2 thick film coated with nanoscale Co3O4 overlayer and Co3O4 thick film coated with nanoscale TiO2 (or SnO2) overlayer, are suggested to show highly selective and sensitive detection of benzene and methylbenzenes, respectively. The increase of less reactive analyte gases and the decrease of highly reactive interference gases could be explained successfully by gas reforming and oxidative filtering, respectively. And gas sensing mechanisms of bilayer sensors were discussed and explained in relation to the material and thickness of catalytic overlayer, sensing layer, and sensing temperature. The bilayer sensor design can provide a new and versatile platform to control gas selectivity and sensitivity, which will make gas sensing library more abundant.
 S.-Y. Jeong et al. J. Mater. Chem. A 5, 1446-1454 (2017)
 H.-M. Jeong et al. ACS Appl. Mater. Interfaces, 9, 41397-41404 (2017)
Metal oxides represent a vast class of materials of interest for various scientific communities, ranging from physics to chemistry, from material science to engineering. The majority of metal oxides are in principle sensitive to gases. However, in order to be performing materials for the fabrication of chemical sensors, they need to fulfill specific requirements such as sensitivity, selectivity, stability, fast response and recovery time.
In the field of chemical sensing, metal oxides have been investigated in different forms, starting from thick to thin films. In recent years, nanostructured metal oxides, and in particular one-dimensional (1D) nanowires, have attracted a wide attention, especially in the field of conductometric sensing devices. These devices are among the most performing, smal size, easy readout and cheapest devices that can be integrated, for example, into portable detection systems. In order to increase the sensor response and selectivity, various strategies have been employed, such as the modulation of the sensing temperature,  morphological changes,  catalyst doping/loading,  and catalytic filtering of interference gases . Another effective strategy to enhance the sensor response and selectivity is to construct heterojunctions between two different oxides that enables the control of conductivity at p-p, p-n, and n-n interfaces, and synergistic catalytic effects between different materials.
The main idea behind this work is to bring together the properties of two different nanostructure materials into a single sensing platform by using a common, simple, low cost and high yield growth method. Herein, we report on the novel preparation and characterization of NiO/WO3 and NiO/ZnO branched 1D-1D nano-heterostructures, consisting of inner NiO nanowires  and outer WO3 and ZnO nanowires obtained through the same vapor-phase method. The surface morphology of the nanowires was investigated by using scanning electron microscopy (SEM) while, for structural characterization GI-XRD, the transmission electron microscopy (TEM), and Raman spectroscopy were performed. The structural characterizations shows the presence of a ternary nickel tungstate (NiWO4) phase in NiO/WO3 nanostructures, while for NiO/ZnO no new phase was detected and the lateral growth of ZnO on NiO resulted almost epitaxial. Finally, NiO nanowire and heterostructure based conductometric gas sensing devices have been fabricated and tested towards different gases spices such as (NO2, H2, CO, VOC’s) and their sensing performances have been compared. Interestingly, both NiO/WO3 and NiO/ZnO heterostructure based sensing devices shows superior performance compared to NiO sensors.
 E. Comini et. al., Sensors Actuators B Chem. 179, 3 (2013).
 N.Kaur et. al., Sensors Actuators B Chem. 262, 477 (2018).
 L. Wang et. al., Mater. Sci. Eng. C 32, 2079 (2012).
 S.-Y. Jeong et. al., J. Mater. Chem. A 5, 1446 (2017).
 N. Kaur et. al., Nanotechnology 27, 205701 (2016).
One dimensional (1D) nanostructures of metal oxides and their derivatives are very useful for the gas sensing applications due to their high sensitivity, quick response, and responsiveness to a wide range of target gases. However, their controlled integration on low-power and ultra-compact MEMS micro-heating platform has been very challenging due to the difficulty of their handling and manipulation, as well as poor controllability and low throughput of device integration. We have developed a novel method for the direct synthesis and facile in-situ integration of 1D metal oxide nanostructures and their derivatives via localized hydrothermal synthesis, selective surface modification, and liquid phase deposition for the chemical conversion [1-3]. In this talk, we would like to present two recent achievements based on this technology. First, a low-power chemoresistive MEMS gas sensor array consisting of four suspended MEMS strip type microheaters and locally synthesized 1D nanomaterials (ZnO nanowires, ZnO/SnO2 core-shell nanotubes, Pt coated ZnO nanowires, and Pt coated ZnO/SnO2 core-shell nanotubes) is explained. By taking advantage of ultra-small thermal mass of sensing region and high surface area of sensing materials, low-power (sub-5mW) and highly selective multiplexed gas sensing for H2 and H2S gases has been realized. Second, we have developed a catalytic combustion sensor array consisting of suspended MEMS strip type microheaters and locally synthesized 1D nanomaterials (ZnO nanowires and Pt nanotubes). By taking advantage of high catalytic activity of Pt nanotubes and ultra-small thermal mass, we could realize a low power, quickly responding, and selective catalytic combustion sensor for H2 gas.
As a semiconductor metal oxide with perovskite structure, LaCoO3 is of interest for chemical sensors. The hole-type conduction occurs via Co-O framework. The surface of LaCoO3 nanostructures exhibits different adsorption sites (La3+ and Co3+) and active sites (chemisorbed oxygen, lattice anions) for gas molecules reception. The sensing mechanisms with LaCoO3 and its nanocomposites are unclear. In this work we obtained nanocrystalline LaCoO3 modified by Ag nanoparticles with improved sensitivity and selectivity to H2S, characterized the microstructure and surface sites of materials, and proposed the sensing routes during gas-solid interaction.
Nanocrystalline LaCoO3 with particle size 30-80 nm (Fig. 1) and specific surface area 5-10 m2/g was obtained by sol-gel synthesis using ethylenediamine as a coordination ligand. The samples were impregnated by Ag nanoparticles with the size increasing in the range 30-60 nm on increasing silver percentage 2-5 wt.%. XPS spectroscopy demonstrated the presence of La3+, Co3+, O2- ions in the bulk along with a large fraction of chemisorbed oxygen species. Metallic Ag nanoparticles were observed by XPS and XRD. The DC-resistance increased in presence of Ag due to electrons donation into p-type LaCoO3. The Ag/LaCoO3 nanocomposites demonstrated higher sensitivity to 0.2-5 ppm H2S at 200 ºC, in comparison to pure LaCoO3 (Fig. 2). Cross-sensitivity tests showed about 10-times higher sensor response of Ag/LaCoO3 to 2 ppm H2S, as opposed to 20 ppm CO and NH3 (Fig. 3). On DRIFT spectra of the samples Ag/LaCoO3 exposed to H2S at 200 ºC the evolution of peaks was observed relevant to adsorbed H2S, Ag2S and SO42- groups (Fig. 4a). Thus, the sensing process occurred via H2S adsorption favored by Ag nanoparticles and oxidation to sulfur oxide and sulfate species on the LaCoO3 surface. The reaction products, except SO42-, disappeared during further exposure in air, which accounts for sensor recovery (Fig. 4b). The persistent sulfate species were likely inactive by-products that did not affect the sensors behavior.
Herein, formaldehyde sensors based on gallium-doped In2O3 inverse opal (IO-(GaxIn1-x)2O3) microspheres were purposefully prepared by simple ultrasonic spray pyrolysis method combined with self-assembly sulfonated polystyrene spheres template. The well-aligned inverse opal structure, with three different-sized pores, plays dual roles of accelerating the diffusion of gas molecules and providing more active sites. The Ga substitutional doing can alter the electronic energy level structure of (GaxIn1-x)2O3, leading to the elevation of Fermi level and the modulation of band gap closed to a suitable value (3.90 eV), hence, effectively optimizing the oxidative catalytic activity for preferential CH2O oxidation and increasing the amount of absorbed oxygen. More importantly, the gas selectivity could be controlled by varying the energy level of adsorbed oxygen. Accordingly, the IO-(Ga0.2In0.8)2O3 microspheres sensor showed high response toward formaldehyde with fast response and recovery speeds, and ultralow detection limit (50 ppb). Our findings finally offer implications for designing Fermi level-tailorable semiconductor nanomaterials for the control of selectivity and monitoring indoor air pollutant.
In this work Ga doped ZnO thin films have been deposited by RF magnetron sputtering onto a silicon micro-hotplate and their structural, microstructural and gas sensing properties have been studied. ZnO:Ga thin film with a thickness of 90 nm has been deposited onto a silicon based micro-hotplates without any photolithography process thanks to a low cost and reliable stencil mask process. Sub-ppm sensing (500 ppb) of NO2 gas at low temperature (50 °C) has been obtained with promising responses R/R0 up to 18.
Chair: Veronica Sberveglieri
The analysis of volatile organic compounds (VOCs) as disease biomarkers released by the urine, it permits an early and non-invasive diagnosis of Urinary Tract Infections (UTI) . For this purpose, an instrumental method like the electronic nose composed by micromachined metal oxide gas sensors has been taken under consideration. Escherichia coli (E.coli) is the pathogenic microorganism responsible for up to 80% of theUTI and it is here chosen as benchmark bacterium . The purpose of this research work is to test the capability of the electronic nose approach to recognise the presence of E.coli, identificative of a possible UTI disturb , in urine samples.
Materials and Methods
In the research’s work, a device named miniMOx (JLM Innovation) has been involved. It is equipped with two micromachined metal oxide gas sensors (MOX): TGS8100 (Figaro) and CSS801 (CCMOSS). The MOX are capable to work with custom temperature modulation protocols controlled though their embedded heaters. This modulation periodically activates and freezes the interaction between gaseous molecules and the metal oxide surface, producing a periodic resistance vs. time curve as a response. In particular, a square wave of a 20 seconds period was applied. A warm semi-period was settled at voltage of Vheaters: 2.31 V for 10 seconds while the cold one at the voltage of Vheater = 1.65 V for the same amount of time. The resistance vs. time curves obtained were described through the ΔRcold-hot, ΔRcold and ΔRhot parameters. The ΔRcold-hot represents the subtraction between the sensor’s resistance measured at the end of the cold period and the resistance measured at the start of the warm period after 0.2 seconds. ΔRcold signifies the difference between the sensor’s resistance measured at the end of the cold period and after 0.2 seconds or at the beginning to the same period. ΔRhot respects the warm period. In the end, a Principal Component Analysis algorithm (PCA function on Matlab) was used to elaborate the data acquired with the described parameters. Three representative samples were taken under consideration: urine, urine contaminated with a pathogenic microorganism (Escherichia coli) and sterilized water as a control. The analysis’ procedure provided to place in contact the miniMOx for a time of 5 minutes with the head-space released from the samples, interspersed with 10 minutes for the sensors’ recovery in ambient air. In parallel, bacterial counts were performed to monitor the Escherichia coli concentration during the whole analysis.
Results and Discussion
The resistance vs. time curves was acquired with the two micromachined metal oxide gas sensors during the exposition at the VOCs released by uncontaminated urine and urine inoculated with E. coli at the initial concentration of 104 CFU/ml. The resistance values are lower during the warm semi-period (200-300KΩ with urine samples contaminated by E.coli, 100-200KΩ with urine samples uncontaminated) and larger during the semi-cold one (600-1000KΩ, with urine samples contaminated by E.coli, 250-650KΩ with urine samples uncontaminated), mainly due to thermal effect on the MOX semiconductor. The shape of these curves is sensitive to the surrounding atmosphere, with differences that can be properly resumed in terms of ΔRcold-hot and ΔRcold.PCA algorithm applied to the parameters explained before. The PCA Score Plot represents a scenario with three separated cluster, each one representative for sterilized water, urine and urine contaminated with Escherichia coli. Therfore, there is a separation between the two urine’s samples. Since the difference between the two urine’s samples is the E. coli presence, potentially the pathogenic microorganism is the responsible to the separation itself.
The custom measurement protocol developed with the commercial electronic nose miniMOx revealed suitable to discriminate between water, urine and urine with E. coli through the analysis of the VOCs released by them. Since E. coli causes different kind of diseases in the human body, an early detection of this pathogenic microorganism into the urine could prevent the illnesses development. In conclusion, the miniMOx could be an easy-to-use, low-cost device for the pre-screening diseases through the VOCs released by urine.
 Mills G. A., Walker V., “Headspace solid-phase microextraction profiling of volatile organic compounds in urine: Application to metabolic investigation”, Chromatogr. B Biomed Sci. Appl., 259-668, 2001.
 Persaud K.C., Pisanelli A.M., Evans P., Travers P. J., “Monitoring urinary tract infections and bacterial vaginosis”, Sens. Actuator B Chem., 116-120, 2005.
 Bernabei M., Pennazza G., Santonico M., Roscioni C., Paolesse R., Di Natale C., D’Amico A., “A preliminary study on the possibility to diagnose urinary tract cancers by an electronic nose”, Sens. Actuator B Chem., 1-4, 2007.
Electronic Olfactory Systems, also called Electronic Noses, are instruments designed to mimic the sense of smell. This is obtained by using an array of different gas sensors, whose signals are collected and elaborated by a processing unit; the measured signals are then compared to a pre-determined odour training set, in order to obtain odour recognition and quantification.
In SACMI Electronic Noses, an array of six different semiconducting metal oxide gas sensors is used as the sensing element. These sensors are obtained by the deposition of a 1x1 mm$^2$ layer of sensing material on a 2x2 mm$^2$ Aluminium Oxide substrate, and are realized through sputtering deposition, with a thickness of 10-100 nm, or by serigraphic deposition, with a thickness of about 10 µm.
Even though this technology offers a high sensitivity to many odorous compounds and a good stability with time, several problems related to the sensor output need to be addressed to achieve a stable and reproducible instrumental response, namely:
Many hardware and software solutions have been implemented in SACMI Electronic Noses to overcome these limitations:
As a result, these instruments can be used reliably both in laboratory for quality control measurements and outdoor for continuous environmental monitoring of odour nuisance. In the past years, SACMI Electronic Noses have been applied in food quality control (coffee, olive oil, tomato, etc.), packaging quality control, and monitoring of industrial activities (refineries, waste treatment plants, chemical plants, etc.).
An array of 8 MOX gas sensors coupled with artificial neural networks has been used to assess different parameters of grated Parmigiano Reggiano cheese.
In this study, the aim was to find characteristic volatile organic compounds (VOCs) of C. jejuni in order to detect its presence with an array of metal oxide (MOX) gas sensors in an in-vitro test.
Chair: Lionel Presmanes
In this presentaiton, a refreigerant gas sensor will be introduced with its development background.
Takuya Suzuki1,2, Andre Sackmann1, Alexandru Oprea1, Udo Weimar1, and Nicolae Barsan1
1Institute of Physical Chemistry, Universi ty of Tübingen, Tübingen, Germany
2Corporate R&D Headquarters, Fujielectric Co. Ltd., Hino-city Tokyo, Japan
Obtaining low cost, simple, compact and good performance chemoresistive CO2 gas sensors has the potential to be a game changer in the field of indoor air quality monitoring as well as the agricultural and food businesses. Rare-earth oxycarbonates Ln2O2CO3 (Ln = La and Nd) have been proposed as promising chemoresistive materials for CO2 sensors 1 2. In this contribution we present the results of a broad investigation focused on selecting the best candidates in the rare-earth compounds and, in the case of the best performing material, preliminary results dealing with the understanding of sensing by the operando methods 3.
Rare-earth oxycarbonates and rare-earth oxides (rare-earth element = La, Nd, Sm, Gd, Dy, Er, Yb) were produced by the heat treatments of the oxalate hydrate or the acetate hydrate in a flow of ambient air at temperatures between 450°C and 550°C for 18 or 72 hours. The powders after the heat treatment were mixed with propane-1,2-diol. The resulting pastes were screen printed onto alumina sensor substrates (provided with Pt interdigitated electrodes and Pt heater). The substrates were dried and then heated at the same temperature as its heat treatment.
Figure 1 shows the comparison of sensor signals to 1,000ppm CO2 under standard humidity and operation temperature conditions (20°C50%rh, 300°C) for all (11) sensors. The sensor signal is defined as the relative change of the resistance with respect to the resistance in air (CO2=0ppm). Every sensor, excepting the CeO2 and Nd2O3 based, was sensitive to CO2.
Additional investigations of selectivity and stability indicated that hexagonal La2O2CO3 possesses the best properties for a CO2 sensor so far. The detailed performance is shown in Figure 2.
To reveal the sensing mechanism, we started by investigating the transduction by focusing on the conduction through the sensitive layer, with the help of operando AC impedance spectroscopy, and the effect of humidity, with the help of operando work function changes measurements; these investigations will be complemented by operando DRIFTS (Diffuse reflectance infrared Fourier transform spectroscopy) experiments; the operando stands for actual gas sensing conditions (e.g. at an
operation temperature of 300°C, with or without gas exposure, humid or dry atmosphere).
Out of the results of AC impedance spectroscopy, presented in Figure 3 as Cole-Cole plots, one can derive an equivalent circuit, see Figure 4. In it, there are two contributions that describe space charge regions – comprising parallel resistive and capacitive contributions. They can either describe electrode contact and intergranular contributions or heterogeneous intergranular contributions. In series, one finds an additional resistive contribution, which could describe the grains bulk. In DC conditions, the resistive contributions that are describing space charge regions, dominate and will show an exponential dependency on the surface barriers, which vary with ambient conditions. The changes of resistive contributions (Rc + Rgb) are correlated with the changes in the surface barrier height ΔVs as in equation (1).
(Rc + Rgb)0 / (Rc + Rgb)gas = exp (-qΔVs/kT) (1)
where (Rc + Rgb)0 and (Rc + Rgb)gas are the values at 0 ppm and at a certain concentration of CO2, and q is elementary charge respectively.
The inputs from the AC impedance spectroscopy are allowing to separate the contribution of electron affinity Δχ and band bending qΔVs to the work function changes ΔΦ as in (2).
ΔΦ = qΔVs + Δχ (2)
Figure 5 show the preliminary results in the case of the hexagonal La2O2CO3 based sensor operated at 300°C in 20°C10%rh. In this case, the work function changes more than 0.6 eV at 4,000ppm CO2 and the contribution of electron affinity Δχ is larger than that of band bending qΔVs.
The electron affinity mainly depends on the surface dipoles which are caused by surface adsorbents such as hydroxyl groups. We will identify the surface adsorbents by operando DRIFTS experiments.
1 I. Djerdj, A. Haensch, D. Koziej, S. Pokhrel, N. Barsan, U. Weimar, M. Niederberger, Chem. Mater. 21, 5375-5381 (2009)
2 A. Haensch, I. Djerj, M. Niederberger, N. Barsan, U. Weimar, CO2 sensing with chemoresistive Nd2O2CO3 sensors - Operando insights, Procedia Chem. 1 (2009) 650–653.
3 N. Barsan, D. Koziej, U. Weimar, Metal oxide-based gas sensor research: How to?, Sensors Actuators, B Chem. 121 (2007) 18–35.
Acetone was one of the volatile organic compounds present in respiration, and acetone contained in the exhalation of diabetic patients was found to be a combustion metabolite of body fat. Degradation of acetyl-CoA due to the metabolism of fatty acids in diabetic patients increase the concentration of acetone in the blood. Acetone in the blood is excreted as urine or breath. It has been studied that acetone released from breathing is 0.3 to 0.9ppm for healthy people and 1.8ppm or more for diabetic patients. Therefore, a variety of studies have been conducted to monitor diabetes by measuring the acetone gas released from breathing. Methods for measuring the amount of acetone in the exhalation using a GC-MS, an electrochemical sensor, and a method using an array of gas sensors based on metal oxide types were studied.
In this paper, we have been developed an E-Nose system using a metal oxide sensors array and measured the expiration of the normal and diabetic groups to distinguish diabetic patients from normal subjects. And blood samples from those peoples were analyzed to compare the exhaled breath test results using an E-Nose system.
The E-nose system is composed of sensor array, data acquisition and processing, and clustering part. The sensor array shown as figure 1 was fabricated as one chip by depositing indium and tungsten with electron beam applying glancing angle deposition method at Korea Institute of Science Technology (KIST), Korea. A chamber was used to maintain the stable operating temperature of the sensor array and solid phase microextraction (SPME) fiber was used for the transfer of the measurement gas. Figure 2 is shown full system which has been used for experimental work.
The subjects were divided into controls and diabetes group, and 12 samples for controls and 11 samples for diabetics were selected. The collection and measurement of expiration and blood test were conducted in Dongsan Medical Center after approval of the Institute Review Board (IRB). The Clinical data for this study was summarized at Table 1.
The PCA results for these data are shown in Figure 3. As shown in Fig. 3, diabetic patients and controls are distinguished, but some samples were displayed in different areas. In the blood test, Blood Sugar Test (BST), glucose, and HbA1C were given more information for classification. Throughout the primary results for comparative analysis between blood test and breath analysis using a sensors array, we confirm the clustering between controls and diabetics is possible, but we need more specific blood test information to confirm accuracy of breath analysis.
Valeriy Krivetskiy1, Matvey Andreev1, Alexander Efitorov2
1 Department of Chemistry, M.V. Lomonosov Moscow State University, Leninskie gory 1/3, 119234, Moscow, Russia
2 Skobeltsyn Institute of Nuclear Physics (SINP MSU), M.V.Lomonosov Moscow State University, 1(2), Leninskie gory, GSP-1, Moscow 119991, Russian Federation
Detection of pipeline hydrocarbons leakage is a valid industrial demand . A deployed network of autonomous miniature micromachined metal oxide semiconductor gas sensors with low power consumption possess a great perspective of practical use in this regard [2,3]. The main obstacle of their high cross sensitivity can be overcome by the implementation of sensor arrays or working temperature modulation in combination of signal processing and nonlinear calibration [4,5]. In this work we demonstrate stable selective detection of propane and methane in low concentrations in the real urban ambient air by the SnO2-based semiconductor gas sensor.
Nanocrystalline SnO2 gas sensitive material has been synthesized by flame spray pyrolysis technique. Gas sensors were fabricated on the basis of 2x2x0.15 mm alumina micro-hotplates with the use of α-terpineol as a binder. Measurements were carried out in a flow-through sensor cell with the use of outdoor air with the admixture of methane and propane from certified gas bottles. Gas concentrations varied from 40 to 200 ppm. Sets of data were collected during a series of 24h experiments with variable air temperature and humidity. The measurements were conducted through 2 consecutive months in order to determine the stability of sensor performance. Collected 17 data sets were divided to 10 sets, used for model training and calibration, and 7 sets, used for motel testing. The details of sensors working temperature cycle and gas sensor setup are given on fig. 1
Figure 1. (a) gas sensor setup (b) metal oxide gas sensor working temperature and sensitive layer resistance profile.
The obtained gas sensor resistance profiles, recorded during temperature cycles, demonstrate considerable variance due to effects of ambient air humidity and temperature changes. The application of principal component analysis (PCA) to the raw sensor data did not allow to distinguish between methane, propane and air in any acceptable extent (Fig. 2a). The use of data pre-processing, represented on fig. 2b (baseline cut-off, data scaling, extraction of data points only from 300-500 oC working temperature region), in combination with machine learning algorithm (artificial neural network with 50 neurons in hidden layer, dropout regularization and sigmoidal activation function) allowed to achieve 86% accuracy of identification of methane vs. propane vs. air in real urban air in 40-200 ppm concentration range.
Figure 2. (a) PCA score plots for raw sensor data (b) raw sensor data preprocessing, used for machine learning algorithm.
Data preprocessing allows for compensation of metal oxide gas sensor drift effects during operation in real urban air, caused by variations of weather conditions. Application of machine learning algorithms, based on the artificial neural network approach gives the possibility of selective detection of air pollutants even of very close chemical nature. The presented approach demonstrates the applicability of MOX sensors for application in industrial safety tasks, related to flammable and explosive gases leakage.
The work was funded by Russian Science Foundation grant № 17-73-10491.
1. Mujica, L.E. et al., Struct Hlth Monit 2015, 2350-2357.
2. Santra, S. et. al., Nanotechnology 2016, 27.
3. Guha, P.K. et al., Sensor Actuat B-Chem 2007, 127, 260-266.
4. Krivetskiy, V. et al. Sensor Actuat B-Chem 2018, 254, 502-513.
5. Collier-Oxandale, A.M. et al., Atmos Meas Tech 2019, 12, 1441-1460.