Skip to main content

Surface ionization and novel concepts in nano-MOX gas sensors with increased Selectivity, Sensitivity and Stability for detection of low concentrations of toxic and explosive agents

Final Report Summary - S3 (Surface ionisation and novel concepts in nano-MOX gas sensors with increased Selectivity, Sensitivity and Stability for detection ¿ of toxic and explosive agents)

Executive summary:

The objective of S3 was to develop breakthrough technologies in gas sensing that provide higher sensitivity and selectivity at reduced cost and power consumption. This objective was pursued by bringing together excellence and complementary skills of European Union (EU) and Russian groups. Enhanced sensitivity and selectivity enabled toxic and explosive gases - target gas species will be nitrogen dioxide (NO2) and trinitrotoluene (TNT) for explosives applications, ammonia (NH3) and hydrogen sulphide (H2S) for industrial environment and safety- to be detected against a background of air constituents and ubiquitous air contaminants. This task was pursued by studying sensors and sensing principles based on semiconductor nanowire (NW)s molecularly engineered, in terms of doping level, used additives and /or functionalisation processes and HSs and deposited onto SiO2 / Si and / or alumina Microelectromechanical systems (MEMS) heater platforms.

These platforms were configured in innovative ways to obtain multiple signals from one and the same sensitive layer. Signals recovered included resistive, field-effect, ion emission and catalytic / thermal conductivity response of the NW materials. Low power operation and additional enhancements in selectivity were obtained through pulsed temperature operation and combined self-heated operation mode using dynamic and steady state responses and modulated optical excitation. Furthermore, the increased stability of NW-based sensing materials positively affected the reliability of the developed sensors. In order to meet application demands, S3 further explored novel concepts of sampling, filtering and preconcentration of target substances based on nanostructured filter and enrichment materials.

The development and the theoretical modelling of this new generation of nanostructured gas-sensing and ion emitting materials was supported by a wide range of morphological and physicochemical characterisation techniques.

The cooperation between EU and Russian groups was improved and strengthened by short and long term exchanges of researchers, the organisation of common workshops and tutorials and the establishment of Joint doctoral degree (JDD)s.

Novel approaches in gas and vapour sensing emerged from:

- materials and technologies of semiconductor metal oxide based NWs;
- catalysation of the surface with target specific receptor species;
- nanostructured filters for separation and enrichment;
- development of semiconductor hetero-structures;
- improved design of substrates for low power consumption: MEMS alumina membrane and self-heating;
- novel emerging approaches: combined resistive / ion emission and optical excitation;
- theoretical basis and modelling;
- experimental verification of sensor performance.

The project aimed at exploring novel emerging approaches in metal-oxide (MOX) semiconductor nano-structured gas sensor science and technology by addressing the theoretical basis, modelling, materials technology and experimental verification of the performance. Such developments generated impacts in the following areas:

1. Significant increases in the sensitivity, selectivity and stability of MOX gas sensors, including a reduction in the response and recovery time constants and a significantly decreased response to interferences (e.g. variable humidity) enabling the sensing of toxic and explosive agents in air or of health and safety hazards in critical environments.
2. Significant advances in the state-of-the-art in nano-structured sensor science and technology.
3. Increased cooperation between EU and Russian organisations.

Project context and objectives:

The objective of S3 was developing breakthrough technologies in gas sensing that provide higher sensitivity and selectivity at reduced cost and power consumption. This objective was pursued by bringing together excellence and complementary skills of EU and Russian groups. Enhanced sensitivity and selectivity enabled toxic and explosive gases - target gas species will be NO2 and TNT for explosives applications, NH3 and H2S for industrial environment and safety- to be detected against a background of air constituents and ubiquitous air contaminants. This task was pursued by studying sensors and sensing principles based on semiconductor NWs molecularly engineered, in terms of doping level, used additives and /or functionalisation processes and heterostructures (HSs) and deposited onto SiO2 / Si and / or alumina MEMS heater platforms.

These platforms were configured in innovative ways to obtain multiple signals from one and the same sensitive layer. Signals recovered included resistive, field-effect, ion emission and catalytic / thermal conductivity response of the NW materials. Low power operation and additional enhancements in selectivity were obtained through pulsed temperature operation and combined self-heated operation mode using dynamic and steady state responses and modulated optical excitation. Furthermore, the increased stability of NW-based sensing materials positively affected the reliability of the developed sensors. In order to meet application demands, S3 further explored novel concepts of sampling, filtering and preconcentration of target substances based on nanostructured filter and enrichment materials.

The development and the theoretical modelling of this new generation of nanostructured gas-sensing and ion emitting materials was supported by a wide range of morphological and physicochemical characterisation techniques.

The cooperation between EU and Russian groups was improved and strengthened by short and long term exchanges of researchers, the organisation of common workshops and tutorials and the establishment of JDDs.

Novel approaches in gas and vapour sensing emerged from:

- materials and technologies of semiconductor metal oxide based NWs;
- catalysation of the surface with target specific receptor species;
- nanostructured filters for separation and enrichment;
- development of semiconductor hetero-structures;
- improved design of substrates for low power consumption: MEMS alumina membrane and self-heating;
- novel emerging approaches: combined resistive / ion emission and optical excitation;
- theoretical basis and modelling;
- experimental verification of sensor performance.

The project aimed at exploring novel emerging approaches in MOX semiconductor nano-structured gas sensor science and technology by addressing the theoretical basis, modelling, materials technology and experimental verification of the performance. Such developments generated impacts in the following areas:

1. Significant increases in the sensitivity, selectivity and stability of MOX gas sensors, including a reduction in the response and recovery time constants and a significantly decreased response to interferences (e.g. variable humidity) enabling the sensing of toxic and explosive agents in air or of health and safety hazards in critical environments.
2. Significant advances in the state-of-the-art in nano-structured sensor science and technology.
3. Increased cooperation between EU and Russian organisations.

The most important parameters describing the performances of gas sensors are the well-known 3 S, namely sensitivity, selectivity and stability. S3 significantly improved all three parameters. S3 led to a significant increase in sensitivity thanks to the innovations introduced in the preparation of sensing materials and in the final sensor platform. Besides, S3 intended to enter scientifically new territory by modifying the gas-surface interactions on MOX materials by grafting selectively binding molecular receptors onto the MOX surfaces. In going beyond the limits of the traditional MOX research, the S3 consortium benefitted from its previous achievements in the field of MOX-NW growth and the resulting excellent transducer functionalities of the MOX NWs. Part of the S3 approach was providing sensitivity and selectivity not only by intelligent materials design but also by interrogating one and the same sensor signal in physically different ways to obtain higher-dimensional sensor signals with enhanced information content. In order to arrive at the requested levels of sensitivity and selectivity, the S3 sensor systems also comprised sampling, filtering and preconcentration features, all of them building on novel materials with nanoscale morphologies. S3 also led to a remarkable improvement in the stability of the sensing materials. In this way S3 aim to remove one of the key factors that in the past have slowed down the commercialisation of MOX chemical sensors. This advance was enabled by materials engineering of the nanostructures and by the development of reliable electrical contacts to these structures. Long-term drifts in the electrical properties were firstly avoided by depositing almost mono-crystalline nanostructures, which do not suffer from nanostructure coarsening or phase change to the same extent as the currently available materials. The second measure was developing innovative high-temperature lift-off processes for the fabrication of reliable contacts to the NW sensor materials. The deposition of almost mono-crystalline materials also provided excellent control over the crystal growth direction. As a consequence, NW faces exposed to the gas ambient always looked alike. In this way a reproducible production of closely identical sensors becomes possible. This again is a key asset for commercialisation.

EADS, in particular, has a deep business interests in the field of homeland security. In this field, the 9 / 11 events caused a dramatic increase in the request for sensors and sensor systems that can detect rapidly, efficiently and at moderate cost trace explosives and a whole range of toxic substances at control points, inside Heating, ventilation and air conditioning (HVAC) systems and also in other confined spaces inaccessible to visual observation. Whereas control points are obvious spots of need, it should be considered that HVAC systems in aircrafts and in public buildings are efficient distribution systems for toxic materials. By providing more sensitive and selective sensor materials, S3 promoted progress in the widely open area between bulky and expensive high-quality analytic instruments and miniaturised low-cost electronic nose devices. S3 technologies, therefore, significantly contributed to increased safety and security in the fields of aeronautics and homeland security. As affordable high-performance light-weight instrumentation is not only of interest in aeronautics and homeland security, S3 is likely to trigger many downstream applications in the much wider field of chemical sensors and miniaturised analytical instruments. Nowadays, the main limitations for the sensing system networks are due to their power consumption that prevents fulfilling the rules of full autonomy in their working period. It obliges to perform their tracking and to control all the elements in short time periods, for example batteries and communication modules control, or just to have complex power supply installations that need continuously to be re-adapted to the requirements.

Monitoring means a high added cost, which presents a severe obstacle with regard to a wider development and an extension of the wide security network. Aside of these limitations about autonomy, other features such as size, weight, communication capability and power consumption level of the present-day systems must also be considered also as severe technical obstacles in the development of these network elements. S3 generated a large financial benefit as it allowed achieving important reduction in the power consumption based on the new substrates and/or in the self-heating NW platforms. It facilitated the establishment of networks without any pre-installation including mobile units. Thus, S3 results could extend the application for chemical product leakage control; dangerous matter control; pollutant and poisoning substance emission control and increase of the security control for harmful air-quality events.

S3 results facilitated outstanding technologies to have significant increase of safety and security in energy production plants, chemical industries and dangerous goods transports with potential risk on the ambient pollution. Size and power consumption reduction together with the increase of sensing performance reliability on the base of the use of well-tailored nanomaterial properties - mainly NWs - could launch innovative products and systems with very high impact on the market of low cost sensing systems and mobile networks with high added functional value.

The requirements and specifications of the sensors/platform to be developed are documented in deliverable D1.1 / milestone M1.1. Key target species are NO2, liberated thermally from explosive particle residue, NH3 and H2S for applications in air quality monitoring, fire gas detection and leakage monitoring, respectively.

Main topics are MOX sensors with resistive response (RR), surface ionisation (SI) and catalytic, heat conductivity (CH) read-out because of their low cost and small size.

Research activities have been performed to demonstrate the feasibility of the single-substrate (tasks 1.1 and 1.2). Key activities in the reporting period include gas measurements with different sensors on ceramic substrates that were configured to support three read-out options: resistive, ionisation current, catalytic / heat conductivity (SI, RR, CH) and the development and realisation of an electronic heater control circuit for ceramic multi-function heaters. One of the main objectives of the S3 project is to establish the integration programme, which aims to strengthen the collaboration between EU and Russian research groups. The instruments it uses are the long (months) and short (weeks) term researcher exchanges, the organisation of workshops and tutorials and the establishment of JDD. Research exchange between partners is proceeding. Many Russian young scientists move for short periods to EU partners for their research activities. The JDD framework agreement on scientific cooperation has been signed by the Rector of both University of Brescia and Moscow State University. In the meantime, the first JDD student exchange is initiated, with the first JDD student already at MSU.

Three S3 schools have been organised in the three year of the project. The first one was carried out 2010 spring in Rimini (Italy), with the support of ISOCS. It was a complete success, and thus S3 propose the same program in the second one, held in the Russian Federation at Igora near St. Petersburg 2011, and in the third one, held in Tuebingen (Germany) 2012.

The first S3 workshop was organised in Barcelona during November 2010. Beside the contributions from the different groups, three external speakers were invited to participate in the event: Dr Ignasi Vilajosana, CEO of Worldsensing (see http://www.worldsensing.com/ online for further details) (Spain), Dr Barth from CNRS/Univ. Marseille (France), and Dr. Pecchia from CNR / ISMN (Italy). The second S3 workshop was a satellite event of the Eurosensors 2011 conference, held in Athens (Greece). The third S3 workshop has been held as a satellite event during the IMCS conference and Sensor&Test fair in Nuremberg 2012 as a continuation of the successful event during Eurosensors 2011 in Athens.

The Sensor&Test fair and the IMCS conference provided an excellent environment for the third S3 workshop, bringing together expert, stakeholders and users. Beside members of S3, INGENIOUS and Sawhot, S3 was able to encourage prospects from outside participating in this workshop.

One of the RF partners of the S3 project, Dr Nikolay Samotaev, Associate Professor at the Nano and Microelectronics Department NRNU MEPhI, was awarded the Vladimir Svorykin National Prise for Innovations. The award was handed over to Dr Samotaev by the President of the Russian Federation, Dmitry Medvedev (see http://eng.news.kremlin.ru/transcripts/1493 online for further details) at the occasion of the Skolkovo Meeting of the Commission for Modernisation and Technological Development of Russia's Economy on 14 December 2010. A video of an interview with Dr Samotaev in the Russian State TV is available under http://www.vesti.ru/videos?vid=311316

One of the EU partners of the S3 project, Dr Gerhard Müller, EADS Innovation Works, was elected as a new member of the EADS Hall of Fame in the category of Great Inventor for his contributions in the field of chemical and gas sensors. The award took place in the Dornier Museum Friedrichshafen on 9 February 2012.

Project results:

WP1: User requirements / specifications
WP leader: EADS
WP participants: EADS, IREC

WP2: Nano-materials for separation and enrichment
WP leader: IREC
WP participants: IREC, MSU

Task 2.1: Preparation of innovative nanomaterials for enrichment

The objective of task 2.1 was to prepare highly porous materials for enrichment and to define the setup to quantify the preconcentration ability of each material. Enrichment materials aim at selective enrichment of the target gas substances inside nanoporous materials to present - upon desorption - higher target gas concentrations to follow-on detectors.

For it, highly porous materials, mesoporous silica SBA-15 and KIT-6 were synthesised and loaded with several additives (Co, Cu, Fe, Mg, Ca and W) in order to modify their acid-base properties. The loading was in each case 2%wt of the metal additive. The found acidity was, as expected: WO3 > SiO2 > Fe2O3 > CuO > CoO > MgO > CaO.

Mesoporous SBA-15 and KIT6 were prepared and loaded with the selected additives. The materials obtained presented a high surface area, around 750m2 / g. It should be noted that for both mesoporous silica, Fe, W and Cu additives were better dispersed than the others, as the loss of BET area for these additives is less than 10 %. All these prepared materials were tested in an especially designed experimental set-up using a micro TD from Air-sense. This desorption unit allows to control the flux, temperature and time of adsorption step. Likewise, a precise methodology has been developed to calibrate the system for ammonia one of the targeted gases. The developed method allows quantifying low levels of ammonium, from 5 to 500 ppb. Higher levels will be quantified by dilution in DI water. An ISE electrode will be also tested in order to obtain quicker results. Thus, it allows quantifying the preconcentration ability of materials using highly diluted ammonia gas.

The capacity of ammonia adsorption, expressed in terms of weight of ammonia adsorbed per gram of pre concentrating material is plotted in the figures. Results were obtained after pre concentrating for a certain period of time, named flow time, a gas composed of a known NH3 diluted in synthetic air at a rate of 50 mL/min. In each experiment, around 100 mg of fresh pre concentrating material was used. In all graphs, the dotted line corresponds to 100 % desorption.

Comparing the results of both pure mesoporous silica with a standard sepiolite to serve as a reference, it seems that KIT-6 has a superior adsorption capacity due to its morphology.

In both cases, the addition of earth alkaline elements (Ca, Mg) greatly improves the adsorption capacity of the material. It has been attributed to the formation of OH bonds. Between metals, it seems that W which is the most acidic additive slightly improves the efficiency of pure mesoporous silica.

For W, there is a strong influence of the adsorption of the material with metal loading. It reaches a maximum at 6 %wt., with an adsorption capacity clearly bigger than in the case of Ca or Mg.

It should be also taken into account that the pre concentration efficiency for KIT6-6 %W is around 90 % and the minimum resolution is better than 0.1 mg / g. The expected pre concentration factor is higher than a conservative 40 ratio.

The proposed materials satisfy due to its own nature the requirements concerning thermal stability. They are totally stable up to very high temperature.

As summary, candidates for ammonia preconcentration were synthesised and characterised in terms of surface area, surface acidity and ammonia adsorption capacity. It was identified at least two promising materials: KIT mesoporous loaded with 6-8 % W or with earth metals, which are clearly better than the former silica mesoporous oxides.

Task 2.2: Preparation of innovative nanomaterials for separation

The objective of task 2.2 was to develop the pre-concentration system including the seolite-based adsorbents and oxidation catalyst for full extraction of NO2 molecules from air and alternatively NH3 molecules.

The sorbents for NO2 extraction and concentration have been prepared in the form of nanopowders by ion exchange procedure based on MFI seolites in Cu forms. To increase the effectiveness of NO2 extraction from air the Cu-MFI adsorbent was accompanied with MnO2/SiO2 catalyst.

The non-zero NO2 concentration in the gas flow during the adsorption stage (before heating) indicates NO2 breakthrough. During the thermodesorption the sensor signal presents two peaks. The first one is due to the oxidation of NO molecules formed during the decomposition of NO2 on the surface of adsorbent. The second peak corresponds to the desorption of NO2 from adsorbent. So, use of the catalyst revealed that the interaction with copper cations on the adsorbent surface can lead to the loss of up to half of the introduced NO2 in the form of NO which also can be captured by this adsorbent.

The obtained results indicate that increase of copper content leads to the increase of maximum NO2 concentration in gas flow after desorption. The NO2 loss can be decreased by combination of adsorption and oxidation functions in the same material. The use of such material allows to increase the efficiency of NO2 pre-concentration up to 100 times and to decrease the NO2 loss down to 10 %.

Likewise, catalytic NH3 conversion by applying MnO2 / SiO2 catalyst has been used. It allowed to increase the sensor signal towards NH3 using the catalytic oxidation of NH3 into NO2. For NH3 conversion the tube reactor with oxidation MnO2 / SiO2 catalyst is placed into furnace heated at 200 degrees of Celsius. After catalytic conversion the gas mixture was directed into the sensor chamber. To avoid the catalytic impact of sensitive material the sensor with pure SnO2 sensitive layer was used in such experiments.

The found results demonstrate the change of SnO2 resistance in the presence of gas mixture containing 10 ppm NH3 in air (black line) and the same mixture passed through catalytic tube (red line). The inversion of resistance change indicates the presence of oxidising gas (NO2) instead of reducing gas (NH3) in the gas mixture. So such conversion seems to be promising to increase the sensor response of SnO2-based materials to ammonia.

Activities developed in WP2 have selected of the most promising materials for enrichment and separation approaches. Whereas mesoporous silica with 6-8% of catalytic additives (W) has been proposed for the first one, modified SnO2 with MnO2 have been suggested for the second one.

WP3: Sensing nano-materials for surface activation and conductive detection
WP leader: UNIKO
WP participants: UNIKO, CNR, MSU

Task 3.1: Preparation of NWs by Physical vapour deposition (PVD)

Research activities in this task focused on the preparation of MOX nanostructures using physical methods, not involving the use of any chemical precursors. Three different techniques were developed or adapted to be used in the project's activities for the fabrication of metal oxide nanostructures based on vapour phase growth: a custom PVD in tubular furnace, direct thermal oxidation, and thermal evaporation in vacuum chamber. Custom PVD was historically the first system developed. It mainly consists in a Lenton tubular furnace, able to reach up to 1 500 degrees of Celsius, connected to a rotary vacuum pump, in order to control the pressure inside the alumina furnace's tube. Two MKS mass flow controllers inject the transport gases (argon or oxygen) inside the system. A home-made NI LabVIEW virtual instruments completely controls the system, in order to improve the reproducibility of the process almost neglecting the delays of a manual control. Metal oxides powder is placed on an alumina holder in the middle of the furnace tube, at a temperature high enough for evaporation. Target substrates for the condensation process are placed on another alumina holder close to the end of the tube, where temperature is lower. During warm up of the furnace the flow is kept in reverse direction, from the substrates to the powder, in order to avoid any undesired condensation at not-optimal temperature. When the furnace reaches the target temperature, the flow is put in direct direction, and the deposition starts. Deposition time depends on the desired amount of nanostructures, on the material itself and on the other growth parameters. During cool down, the flow is kept in reverse direction. It is possible to tune the morphology of the fabricated nanostructures by changing the condensation temperature, pressure inside the alumina tube, carrier gas flow and composition, deposition time, and catalyst on the target substrates. In the frame of the S3 project ZnO, SnO2 and CuO were successfully synthesised on various substrates, depending on applications: silicon for morphological and structural characterisation, 2x2 mm alumina, alumina and silicon membrane for functional measurements as chemical sensors. Many different catalysts were used to promote the growth of NWs, like Pt, Pd, Au, Nb and Ru. Direct thermal oxidation of metals is another technique that was successfully used to prepare metal oxide nanostructures. Thermal oxidation has many advantages: it works in atmospheric pressure, without the need of vacuum equipment; it is scalable, in a view of large-scale production; it is possible to control the growth pattern, via, for example, shadow masking; it has a high yield, with no risk of cross contamination, due to growth patterning. The major disadvantage is the time required for the growth, usually in hours, depending on the material. Growth process consisted into two main steps: metallic layer deposition and thermal oxidation. Deposition of a metal layer is required if the target substrate is not of metal itself, like for example in chemical sensors. Many different techniques could be used, i.e. thermal evaporation, electro-deposition and magnetron sputtering. The latter was used in all of our experiments. After the patterned metal deposition, samples were placed in tubular furnace in presence of an oxidising atmosphere, usually consisting of a mixture of oxygen and argon. According to the furnace temperature, atmosphere composition, deposition time and gas flow it is possible to control the morphology and the density of the desired nanostructures. Using this technique large quantity of ZnO and CuO NWs were prepared.

Thermal evaporation is the most recent technique introduced. It mainly consists in the thermal evaporation of source metal powder in a stainless steel vacuum chamber, connected to Edwards turbo vacuum pump able to reach very low values of pressure. A heating element able to reach up to 900 degrees of Celsius is located in the middle of the chamber, powered by a DC power supply. Two MKS mass flow controller inlet argon and oxygen inside the chamber, locally oxidising the evaporated source material before condensation on target substrates. Because temperature is not uniform above all the heating plate, and the highest temperature is in the centre, substrates are placed close to edges. The temperature gradient enables the condensation of oxidised metal powder on the substrates. With this technique it is possible to tune the pressure and the atmosphere composition inside the stainless steel chamber, more precisely then in PVD. The biggest disadvantage is the maximum temperature reachable on the hotplate, which is much lower compared to alumina furnace. Promising nanostructures were fabricated with this technique, such as ZnO NWs and nanotubes. The oxidising temperature, atmosphere composition, pressure and gas flow, together with catalyst material used, allow tuning the morphology of fabricated nanostructures.

Efforts were performed in the preparation of NWs functionalised with other materials, via both wet (see task 3.3) and dry methods. In the latter case, nanoparticles of noble metals were deposited on SnO2 NWs via magnetron sputtering in a random pattern. The sputtering parameters and the amount of noble metal are calculated based on the quantity of the SnO2 material, rate condition and the requested ratio between SnO2 and noble metal (from 0 to 3 % ideally). Platinum was used as noble metal for functionalisation.

After the synthesis, samples were delivered to partners involved in WP5 for morphological and structural characterisations, and ones involved in WP6 for functional characterisations, to evaluate the performance of the devices towards target chemical species.

Task 3.2: Preparation of NWs by Chemical vapour deposition (CVD)

In this task we focused on the preparation of nanostructured materials via chemical vapour deposition. Metal oxide nanostructures were grown in a horizontal cold wall CVD reactor in which a high frequency field was used for heating silicon or alumina substrates inductively by placing them on a graphite susceptor to the decomposition temperature inside a quarts glass tube. The desired precursor was introduced in the reactor through a glass flange by applying dynamic vacuum (10-4 mbar) and heating the precursor reservoir to the evaporation temperature. The precursor flux was regulated following the feedback of the pressure measurement in the reactor during the CVD process. Given the delicate interplay of precursor chemistry and material properties, reproducible CVD processes demand suitable precursors possessing adequate vapour pressure and stability in the gas phase. For the deposition of MxOy structures metal alkoxides are used as a common precursor class. However, these compounds are extremely sensitive towards air and moisture. To optimise the process parameters novel precursor class was developed and successfully applied for the deposition of MxOy nanostructures and for their functionalisation with different metals / metal oxides. The description of molecular precursors and detailed report on the process conditions could be found in previous reports. Both thin films and NWs could be grown on the multifunctional substrates via CVD technique. The growth of NWs is following vapour-liquid-solid (VLS) mechanism, where different metals (Au, Cu, Pt) or metal oxides (SnO2 QD) are used as a catalyst.

To enhance the selectivity surface of SnO2 NWs could be modified using different metal/metal oxides. This procedure was done using two-step CVD technique. In the first step of CVD process SnO2 NWs were grown of the Al2O3 substrates covered with gold and then in the second step functionalised with metal particles. This method allows to synthesise high crystalline NWs covered with particles of controlled size. The size of the particles is strongly dependent on the deposition temperature. Cu and Ni were used for surface functionalisation; these samples were sent to the partners for the further characterisation. SnO2-Cu and oxidation of Cu to CuO led to the formation of p-CuO / n-SnO2 HS which found to be highly sensitive and selective to H2S gas.

Task 3.3: Chemical modification of NWs

On the basis of sensor response patterns towards CO, NH3, H2S and NO2 we have selected the most suitable material for SnO2 modification for the purpose of maximising sensor response to any particular gas.

Nanostructured materials (powders) were obtained by wet chemistry synthesis (M18). The synthesised materials were characterised by a complex of complementary physicochemical methods to obtain information on the elemental and phase compositions, particle sizes, and the specific surface area.

SnO2-Sb sample was obtained by coprecipitation of stannic acid and Sb2O3*xH2O followed by heat treatment.

For the gas sensing experiments, the materials were deposited in form of thick films over functional substrates, which consists of a 2mm - 2mm - 0.25 mm alumina substrate, provided with Pt contacts on the front side. A Pt-meander that acts both as heating element and temperature probe is deposited on the back-side. Measurements have been carried out by flow through technique in a temperature-stabilised sealed chamber (volume of 1 l) at 20 degrees of Celsius under controlled humidity, working with a constant flux of 300 ml / min. During the measurements the substrate temperature was maintained constant in the range 200-350 degrees of Celsius. During each measurement cycle, devices were exposed to a constant concentration of gas for 25 minutes and then to the background atmosphere for 1 hour. DC volt-amperometric measurements (U = 1 V) have been carried out to monitor samples electrical conductance (G) during exposure to CO (40 ppm), acetone (130 ppm), NH3 (20 ppm), H2S (2 ppm) and NO2 (3 ppm). The sensor signal values have been calculated as G / G0 for reducing gases (CO, acetone, NH3, H2S) and as R / R0 for oxidising gas (NO2). Modified materials show strikingly different sensor response to target gases. The obtained results allowed us to plot the patterns of sensor response towards different gases. The selectivity of the response to gas molecules is greatly affected by chemical composition of sensor materials as well as and by the sensor working temperature.

Tin dioxide modification by RuOx brought about a drastic increase of sensor signals to NH3 with the maximum sensitivity at 200 degrees of Celsius. The sensors based on SnO2 / RuOx allow for reliable detection of as low NH3 concentration as 400 ppb (Fig. 2), which is close to the maximum permissible concentration (300 ppb).

WP4: Low power substrates and packaging
WP leader: EADS
WP participants: EADS, RRC, MEPhI, NIIET

WP4 deals with the development of heater platforms that allow multiple sensor signals to be extracted from one and the same kind of functional nanomaterial (see D9.4 M18).

The signals to be extracted are the following: i) the RR of a MOX sensing layer towards specific target gases: ii) the SI response of a MOX sensing layer towards specific target gases; iii) the temperature change (delta T) of a MOX sensing layer that arises out of catalytic interactions of specific target gases with the sensor surface and/or due to heat conductivity changes in the gas ambient.

Studying such 3-parameter sensors is attractive for two reasons: (i) all three kinds of sensor signal readout can be realised using the same base technology, (ii) all three kinds of sensor readout feature dissimilar cross sensitivity characteristics.

Both features together make the 3-parameter readout scheme attractive for arriving at a higher level of gas selectivity at a minimum level of technological complexity.

Additionally temperature controlled operation instead of the usual heat power controlled operation needs to be employed to attain reliable RES and SI gas signals. Both Res baseline and SI background follows ambient temperature changes and this produces without temperature control fake gas sensor signals (see M36 periodic report WP4).

Task 4.1: Multifunctional substrates for sensing layer interrogation

MEMS substrates: EADS works has been concentrated in the second report period on the development of simpler SI devices having regard to MEMS miniaturisation and commercialisation.

The activities revealed the unforeseeable result that SI process in the planar mode creeping ions and SI in the conventional vertical mode flying ions are totally different processes with different selectivity criteria. Only the vertical mode provides the high amine selectivity which is necessary to realise an illicit drug detector. Due to this finding the implementation of the flying ions SI process in a planar MEMS chip approach has caused a considerable amount of additional work and has made the integration of SI readout in commercial available MEMS hotplate chips more difficult than expected (WP6 an D6.5). Presently, obtaining SI readout requires positioning a counter electrode opposite to the sensing layer surface. Ion extraction over a 1mm air gap requires bias voltages up to 1 000 V. In order to arrive at lower extraction voltages, experiments at different planar readout structures have been made. Planar arrangement experiments provide only ceeping ions without the high amine selectivity. In order to provide flying ions using the planar arrangement IREC has performed FIB cuts on AS micro heaters, thereby separating the pre-deposited interdigital Pt contacts by an additional trench into an anodic ion emission and a cathodic ion extraction part. Also there the high amine selectivity was not observed.

It can be summarised for all experiments:

1. All planar readout configurations failed to provide high amine selectivity as it is with the vertical readout configurations where an external collector electrode exhibited significant amine selectivity necessity of air gaps breaking of adsorbate bonds.
2. Cross talk between MEMS integrated heaters and SI readout electrodes (Si3N4 membrane thickness 1µm) happened with all used MEMS hotplates external heating of MEMS structures with ceramic heater substrates was necessary.

These two limitations are based on quite recently attained measurement results (see D6.5). Device implementation of these new findings requires a new kind of sensor architecture.

Additionally we found that both RES and SI detection need temperature controlled hotplates because of the temperature dependence of the baseline resistivity and the background current (see D9.6 M30 report, task 4.2).

- Clear complementary MOX semiconductor (CCMOS) hotplates are equipped with a temperature control diode; available without MOX and ready for custom-specific MOX deposition.
- Within the past few days AS has equipped its MEMS MOX sensors with constant heater resistance control circuits to compensate environmental changes.
- UST announces TRIPLE sensors with temperature control circuits.

Task 4.2: Packaging

The package of choice consists of a sensor element housed in TO-type of header and a printed circuit board with the corresponding hard- and software. This kit allows to operate the hotplate either with constant heating power or with constant heat resistance and the respective other value can be monitored, thus environmental influences like temperature and pressure changes on the sensor signal can be avoided. EADS cooperates with AS to include an additional SI read out function.

WP5: Characterisation and modelling at nanoscale
WP leader: IREC
WP participants: IREC, CNR, EKUT, UNIKO, MSU, MEPhI, RRC

Task 5.1: Morphological, physical and chemical characterisation

Main activities in this task were oriented towards sensing material selection based on the development of tool's development able to analyse and identify sensing mechanisms. Efforts have been focused on the more outstanding nanostructures selected from the consortium from the initial works for fulfilling the target of the project mainly for NH3 and H2S gas detectors besides the more standard CO and NO2 sensors. Likewise, emphasis has been addressed too on the very promising core / shell structures as well as on HSs metal / MOX (Cu / CuO) with the supporting (SnO2) NW or, just simply, those catalytic clusters with the supporting nanoparticles (Pd, Ru). Morphological, structural and functional characterisations have been achieved. It has allowed a more straightforward understanding of these nanomaterials, as well as their functional modelling.

Studies of size and distribution of the additives have been done as well as significant information has been obtained about interface areas between additive clusters and new HSs based on quantum dots or coaxial / radial NWs. Proposed additives for controlling acidity or basicity at the surface (Pd, Au, Ru) have been studied in correlation with the technology procedure.

Proposals for NH3 and H2S detectors have been elaborated. Likewise, detailed analyses have been done for explaining the role of the surface catalyst with the oxygen interchanges through the surface or just to explain how the NH3 molecule is acting on the SnO2 surface and how it transfers chemical information to electrical signal justifying the response time dependence on the NH3 concentration and operation temperature. These results constitute a significant advance towards the development of improved sensors on the base and understanding of the basic mechanisms taking place on the nanostructure surfaces. Detailed morphological, structural and functional characterisations of nanostructures have been carried out using the developed tools and procedures during the project.

Special interest has been paid on the interface of the nanoheterostructure defined by CuO nanoclusters onto the surface of a SnO2 supporting NWs. The interaction of H2S with the CuO induces the sulphurisation of the cupper which becomes a CuSx compound with metallic character. It involves a strong modification of the work function from CuO to CuSx with the consequent changes of the heterojunction properties and modification of the associated charge space zone. It means that the available conduction channel inside the NW changes radically and the resistance can be modified in orders of magnitude. These mechanisms have been analysed and determined, concluding in what conditions this systems can be a linear gas sensor or in what conditions it can define a highly selective and feasible gas alarm.

In parallel and as continuation of the analysis of the role played by the catalytic clusters detailed studies have been performed about the structure, composition and interface of the catalytic clusters. Different additives used for doping of for catalyse the nanostructures have been investigated from HRTEM and associated techniques, as well as from fine XPS analysis for determining the composition changes taking place on these structures which play a basic role in the sensing mechanisms. Nevertheless, more experiences about the surface of these clusters as well as about their interface need to be done.

In parallel, other different tentative materials developed for achieving the project objectives have been analysed in order to contribute to their improvement and optimisation according to the different WP targets. In this context, we have analysed the new CuO p-type NW material as well as new nano structures of ZnO NWs.

All of these information have been feedback to determine the condition definition for the best material to be assigned to obtain the foreseen improved gas sensor devices targeted in the project CO, NO2, H2S and NH3.

On the other hand, core / shell nanoheterojunctions have also been analysed as very interesting alternative for new ideas about advanced design of improved nanogas sensors. Consequently, activity and effort have been addressed on this kind of coaxial nanostructures.

One of the advantages for these nanoheterojunctions is the possibility for doping the core independently of the electrical characteristics of the shell or coated layer. This one can be selected among different materials for offering better interaction with the target gas or for offering different surface properties or presenting different characteristics such as that of module the surface electrical field. In this way, the charge transfer mechanisms taking place from the surface to the inside of the sensing material could be modulated. As defect at the interface can block these functionalities we have paid our attention on analyse and define criteria for shell material with a epitaxial growth on the core at the same time that electrical conductivity, optical band gap and oxygen chemical potential are considered as parameters for design and improved core / shell systems.

As summary, in this task adequate nanometrology tools have been facilitated and used to determine significant data and information to understand and justify the sensing behaviour obtained in the sensing devices fabricated using the prepared nanostructured materials and their modification using catalytic additives as well as new HSs.

Task 5.2: Modelling of bulk, surface, adsorption and charge transport

Chemical-to-electrical transduction mechanisms between NH3 molecules and SnO2(110) has been examined by Density functional theory (DFT) calculation and confronted to experimental results obtained with individual NW devices.

The surface of SnO2(110) has been considered to be described by four kinds of atoms: Sn5c, Sn6c, bridging O2c and in-plane O3c. Sn5c is a Lewis acid site able to withdraw electrons from an electron rich atom like nitrogen in the NH3 molecule. NH3 transformation reactions were calculated on both the stoichiometric and reduced SnO2(110) surface.

Experimentally, the interaction mechanism of NH3 and SnO2 has originally been studied using individual NW in order to facilitate the interaction between the theoretical simulation and the sensing data. It has been found that it is a thermally activated process with a maximum response between 215 and 250 degrees of Celsius. In this temperature range, pulses from 500 ppb to hundreds of ppm were detected without evidence of signal saturation at the highest concentrations, in good agreement with previous reports. Response times of these devices to NH3 followed an Arrhenius law with an activation energy of Ea equal to 0.5 eV and dependence on the ammonia concentration fitted well to (NH3)-1/2. These results have also directly related to the NH3-SnO2 reaction DFT model proposed.

According to the proposed model, both the regular and the oxygen deficient SnO2 surfaces present relatively reasonable scenarios to understand the activity in NH3 detection. In the case of the reduced surface, the sensing path description is true only if the O2:NH3 ratio is reasonable and the concentration of oxygen vacancies is not extremely high. In the stoichiometric scenario, which is difficult to be fully representative of real SnO2 devices, partially hydrogenated species are highly mobile on the surface and thus can generate large molecules with more than one N atom that subsequently loose H atoms through the O2c. Upon this step, water molecules are generated at these positions that evolve to the gas-phase releasing electrons in the wires. This leads to an effective reduction of the electrical resistance.

In contrast, for the reduced surface the situation substantially changes. As the real mixture in the sensing process contains SA and thus the oxygen content is about 20 %, O2 molecules might heal the oxygen vacancies on the surface. The calculations show that no NH3 molecules are attracted to these vacancy sites. Moreover, O2 dissociation also occupies active positions for NH3 as one of the liberated oxygen atoms from the former O2 molecule ends up on top of an under coordinated position, Sn5c. This reaction is endothermic by about the same amount as the barrier that is experimentally found, Ea equal to 0.5 eV. Therefore, we propose the oxygen adsorption as the rate determining step for this particular path and the main responsible of the activation energy retrieved in the experiments. The resulting O atoms in non-lattice positions are more likely to attract the H atoms from neighbouring NH3 molecules and as result, water formation is rather easy. As discussed before, the N2 formation is concomitant to water formation and its evolution to the gas-phase. This last step releases a pair of electrons in the SnO2 for each water molecule created at lattice positions, and it is considered the ultimate responsible for the resistance reduction upon exposure to NH3. Again, sensing reactions cannot be described as non-catalytic in nature, since a different state for the surface is retrieved after two ammonia molecules are consumed.

Understanding the role of water in the sensing performance of SnO2 NWs is essential to evaluate the full potential of these devices. The interfering effect of moisture with target gases has also been investigated. Nevertheless, a direct quantification of the cross sensitivity between NH3 and H2O has not been reported before. H2O sensing by SnO2 NWs is a completely reversible process, which temporarily enhances the surface electrical conductivity. The associated electrical resistance changes are usually related to the formation of hydroxyl groups at the MOX surface. At normal humidity concentrations (from 25 to 100 % of RH at room temperature), the sensor response saturates. This is an indirect indication of the high affinity of water for the SnO2 surface and results in its high coverage by water-moieties. The H2O-SnO2 interactions turned out to intrinsically interfere with the sensing response of this metal oxide towards typical NH3 traces. This result is easily explained with the help of the theoretical modelling of the NH3-SnO2 interactions:
1. Our calculated water adsorption energy (Eads 1.2-1.3 eV) in the molecular or dissociative form is slightly more effective than that of ammonia at Sn5c sites. As a result, a high surface coverage of water molecules and related species is expected under normal humid conditions due to thermodynamic contributions.
2. In addition, typical water concentrations in humid environments when compared to ammonia levels (hundreds of ppm) are exceedingly high, resulting in preferential adsorption of water by pressure effects. Hence, the overall water coverage partially blocks ammonia adsorption onto SnO2 in mildly humid conditions.
3. Finally, the presence of water in the gas-phase (excess of product molecules) might push the equilibrium of N2 formation towards reactants. The displayed interference of water implies a serious drawback for real applications in which fluctuating humidity level in the environment may lead to false NH3 concentration read-outs. For this reason, the design of future devices should take into account this intrinsic effect. To minimise it, the use of external filters might be considered an effective and straightforward solution.

In parallel, a general modelling based on a charge carrier self-exhaustion theory, CCSE has been developed.

For it, the called Double Shottky barrier model (DSBM) theoretical approach (S.R. Morrison, N. Yamasoe, K. Shimanoe) has been analysed as starting point. This approach is the traditional viewpoint on the origin of the sensitivity to gases considering that it results from the formation of the potential barrier on the boundary between neighbouring grains. This barrier arises due to the granule surface charging during the electronegative gas molecule adsorption. In our approach, we have shown that the DSBM approach permits the satisfactory explanation of some of the found the dependencies, but the results of the measurements of the conductivity in presence of low concentrations of target gas in inert atmosphere cannot be explained in the framework of this approach.

The main difference of our model from DSBM consists in the taking into account superficial Tamm states (TS) at the calculation of electrophysical properties of the material both in presence and in absence of adsorbed molecules. These states always exist at the grain boundaries. They are characterised by rather large surface density MTR 1 014 cm2. Consideration of this fact allows essentially generalise the standard approach.

The main conclusions of the proposed model are:

1. CCSE theory was extended to the case of small grain sise, when the superficial hopping conductivity becomes important. It was shown that, if particle sise a < aSH (5 nm for SnO2), the mechanism of hopping conductivity dominates. The sensitivity of the NSSM disappears in this case.
2. The results of the analysis of starting sensitivity S0 based on the CCSE theory are formulated in the form of theorem of sensitivity, which enables the choose an optimal prospective material for the sensing layer of gas sensor.
3. On the bases of the physical concept, the analysis of published experimental results is carried out. This analysis shows that the concept developed in CCSE theory makes possible the explanation of the experimental data on sensing of both electronegative and electropositive gases.
4. It is shown that it is necessary to take into account superficial Tamm states not only at the calculation of electron balance, but also at the description of adsorption mechanisms, where these states play a key role.
5. It is shown that the DSBM approach (S.R. Morrison, N. Yamasoe, K. Shimanoe) has a limited area of application, because it neglect Tamm states. It is applicable for the description of the electrophysical properties of NSSM only at high humidity of the atmosphere in absence of high concentrations of oxygen. The idea of double Shottky barrier between grains becomes incorrect at the grain size a < 7 nm, when the grain contains less than one donor.
6. It is shown that the Debye length is not a proper parameter of the theory used for the calculation of NSSM conductivity. The comparison of Debye length with the rain size, which is often used for the analysis of electrophysical properties of the materials, is not a correct procedure.
For the adequate formulation of the criterion of the sensitivity, it is necessary to use the value of aCR instead of RD.
7. It is shown that in almost all practically important cases, for the materials with high gas sensitivity (a < aCR), the height of intergrain potential barrier is low compared to the electron bond energy in adsorbed molecule (VS << IA) and practically does not have an influence upon gas sensitivity.

Task 5.3: Operando investigations

Operando measurements have been applied on different materials based on NWs proposed by the project produced from different techniques like PVD) and CVD on Al2O3 Initial conductance measurements were performed for all type of samples. In the case of the CVD samples (SnO2, FeO, NiO-SnO2), sensors were showing an appropriate behaviour towards exposure of H2, CO and Ethanol in humid air. However, it has been shown that inhomogeneity of the layer indicated a more complex application of these sensors for operando investigations doing simultaneous DC and work function / DRIFT measurements.

Using more standard nanoparticles, the conductance measurements were performed in situ under atmosphere changing from pure air to a mixture of NH3 0.3 - 10 ppm in air. The measurements were performed in the temperature range of 25 to 200 degrees of Celsius.

Work function measurements were carried out using Kelvin Probe S and Kelvin Control 07 equipment (Besocke Delta Phi GmbH) in a seroing potential mode. A self-made setup was used for the experiments that included flow chamber, inside which a heating plate with supported Kelvin Probe (KP) samples was embedded. The work function measurements were performed simultaneously with DC-resistance measurements at a fixed temperature in the range of 25 to 200 degrees of Celsius and in situ changing gas composition.

On exposing to NH3 both parameters displayed transients to lower values, the responses increasing with ammonia concentration. Such behaviour was observed for both SnO2 and SnO2 / RuOx samples at all the tested temperatures (30 - 200 degrees of Celsius).

Having the opposite sign to CPD, work function difference turned to be positive on interacting with NH3. In common, reducing gases diminish sensitive layer resistance via the reduction of acceptor-like surface states (chemisorbed oxygen species) liberating the captured electrons and decreasing the Schottky surface barrier height (Vs). In this respect work function decrease could be anticipated because of Fermi level shift towards the conduction band. However, this turned to be not the case in the present study. Supposing conduction model (n-type polycrystalline semiconductor, depletion layers with the width depending on NH3 concentration) to remain unaffected by test gas, which is quite reasonable since the measurements were performed in air and ammonia concentration was rather small (0.3 - 10 ppm), the origin of work function increase should be an increment of electron affinity exceeding the surface potential height (Vs) decrease. Electron affinity being proportional to surface dipoles concentration provides additional information on gas interaction with materials surface that cannot be gained solely from electric measurements, since the processes under investigation could involve species forming the dipoles but not affecting the materials resistance. The surface dipoles that give rise to electron affinity increase are oriented by the positive pole to the solid and by the negative pole to the outside. Such dipoles could be formed by a surface metal atom or cation (positive pole) bound with a more electronegative atom or molecule from gas phase (e.g. water molecules or hydroxyl groups).

The NH3-induced work function differences and sensor responses of the SnO2 / RuOx sample at 200 degrees of Celsius (obtained from the data of the previous figure, see attached pdf) are summarised, as well as electron affinity differences are calculated as a function of test gas concentration. The increase with concentration suggests intensive surface dipoles formation during materials interaction with NH3 along with resistance response formation processes resulted from net surface charge redistribution. These dipoles could be due to adsorption of either initial NH3 molecules or intermediates and / or products of its interaction with the materials surface, e.g. NHy-derivatives, nitrogen oxides, hydroxyls, H2O. Compiling the values of both blank and RuOx-modified SnO2 for a fixed NH3 concentration (5 ppm) at all the tested temperatures, the electron affinity differences were plotted versus temperature. In general, Δχ increased with temperature for both the samples, but it was SnO2 / RuOx that displayed prominently higher electron affinity increment than SnO2 especially at raised temperature (100 - 200 degrees of Celsius). A comparison with sensor signal temperature graph suggested that: (i) SnO2 modification by RuOx increases the materials reactivity to ammonia gas; (ii) this gas-solid interaction involves, on the one hand, charge carriers of the solid and, on the other hand, surface bound dipolar species that do not influence the bulk resistance and, hence, results in modulation of either materials bulk electric properties or its local surface properties; (iii) the interaction with NH3 is intensified by raising the temperature.

These results have shown the very powerful of the operando techniques to correlate sensing properties with intrinsic parameters of the sensing materials.

WP6: Sensor system integration and functional characterisation
WP leader: EADS
WP participants: EADS, CNR, IREC, NIIET, MEPhI, RRC

Task 6.1: System integration

As detailed in task 6.3 while the work to address the sensor performances against the whole set of 18 M and 36 M parameters, it become clear that two aspects were difficult to handle: the response to ammonia, which was lower than the requested value, and the interfering effects from humidity, for which the 18 M requirements were fulfilled, but not the 36 M. So far, while completing the whole testing protocol, WP2 and WP3 activities have been targeted to respond to these needs.

Integrating these activities the consortium was able to respond also to those parameters. For example, as detailed in D9.7 the optimisation of the process to functionalise SnO2 NWs with RuOx nanoparticles permitted to get a response > 4 to 5 ppm of ammonia in humid conditions, RH equal to 36 % (in almost dry conditions, RH equal to 4 %, the response increased to about 10), thus fulfilling both the 18 M and 36 M requirements of the project. The development of materials for separation, in particular for separation of humidity, leaded to a system (separation unit + sensor) showing a very reduced cross sensitivity to humidity (please refer to D9.7 for details).

In the application field of drug and explosive detection the analyte abounds in particle from and is characterised by a very low vapour pressure. Prior the detection the vaporisation of the analyte particle is necessary. EADS has developed a sampling and desorption unit, which enables the evaporation of respective target particles direct into the appropriate gas sensor element.

Task 6.2: Solid / liquid sampling and vapour conversion

In the fields of drug and explosive detection the analytes abound in the form of solid particles. Therefore the analytes need to be collected and vaporised prior to detection. Within S3 EADS has completed its gas test lab so that all three kinds of samples (solid, liquid and gases) can be vaporised into a carrier gas flow and injected into a measurement chamber for gas / vapour detection. An additional open test setup (without chambers) allows determining intrinsic sensor response and recovery times. This latter kind of measurements demonstrated very clearly that the nanosensing materials respond within time scales of less than one second. Slower response and recovery time constants derive from adsorption / desorption effects at the cold surfaces of gas lines and/or measurement chambers (see D9.4 M18 and D9.5 M24).

List of websites: http://www.eurussias3.com

Related documents