CC BY
82HYDROGEN DETECTORS
V. M. Aroutiounian
Member of International Editorial Advisory Board
Yerevan State University 1 Malukyan St., Yerevan, Armenia Ph./Fax: 003741 555590; e-mail: [email protected]
Education: Semiconductors and dielectrics, Kiever Polytechnic Institute, Ukraine, 1964.
Dr. Phys-Math. Sci., Professor, Member of National Academy of Sciences and Engineering Academy of Armenia. Head of Dep. of Physics of Semiconductors and Microelectronics at Yerevan State University, Yerevan, Armenia.
Main field: Physics of Semiconductors, Conversion of Solar Energy into Hydrogen, Oxygen, Electricity and Heat, Sensors, Micro, Opto-and Nanoelec-tronics, Infrared and Microwave Devices, New Semiconductor Materials.
Governmental medal "Anania Shirakaci", Foreign Member of the Russian Academy of Natural Sciences, Corresponding Member of the International Engineering Academy etc.
Member of Intern. Solar Energy Society and European Soc. on Quantum Solar Energy Conversion. Editor-in-Chief of "J. Contemporary Physics" (Allerton Press, USA-NAS, Armenia), member of Sci. Council of International Scientific journal for Alternative Energetics and Ecology. He was involved also in the Editorial board of Int. Journal "Solar cells" (Switzerland).
Author of more than 300 papers in referred journals, several books, 42 certificates and patents of the USSR, USA, RF and RA.
At present, commercial hydrogen detectors are not suitable for widespread use, particularly in fuel cells and transportation, because the detectors are too bulky, expensive, and some are dangerous. From the standpoint of the safety with the global environment, it is necessary today to develop new hydrogen gas sensors working at or near room temperature without any power source.
There are many versions of hydrogen bulk sensors, which are reviewed in this paper. After short discussions of achievements in some modern sensors, main attention is given to discussions of results and problems for sensors made of different semiconductor metal oxides and their solid solutions. Such sensors should be small, cheap and easy to be implanted into microelectronic integrated circuits
Vladimir M. Aroutiounian
Introduction
Hydrogen is the most attractive and ultimate candidate for a future fuel and an energy carrier. Its generation can be realized by a variety of methods, including reforming natural gas, alcohols (methanol etc.), electrolysis and photoelectrolysis of water and biomass generation, as well as chemical decomposition of hydrogen containing compounds. Result of its burning is water, which is transferred again in hydrogen and oxygen. In hydrogen-based fuel cells, the electrical energy will be derived from the reaction of hydrogen and oxygen gases within the fuel cell to make water. Hydrogen does not require a fuel processor in fuel cells when hydrogen and is producible from renewable energy resources such as electricity from solar cells or wind. The biggest barriers to this hydrogen vision are the associated costs and establishment of a safe and effective infrastructure. A common need in this technology area is the ability to detect and monitor
gaseous hydrogen, but hydrogen gas sensors that can quickly and reliably detect hydrogen over a wide range of oxygen and moisture concentrations are not currently available. As hydrogen diffuses easily through most materials, containment of hydrogen is difficult
At present, commercial hydrogen detectors are not suitable for widespread use, particularly in transportation, because they are too bulky, expensive, and some are dangerous. The sensor working at high temperatures itself becomes a possible trigger of explosion, due to its enough input electrical energy of sensor operation. The concentration monitoring of hydrogen is very important for the application of fuel cells as well as for the case that hydrogen being an undesirable contaminant, for example, chemical industry. Not only the leak detection but also the concentration monitoring is important application of hydrogen sensor. From the standpoint of the safety with the global environment, it is necessary today to develop new hy-
Статья поступила в редакцию 02.02.2005 г. The artisle has entered in publishing office 02.02.2005
drogen gas sensors working at or near room temperature without any power source. In addition, they should be small, cheap and easy to be implanted into microelectronic integrated circuits.
There are many versions of hydrogen bulk sensors, for example, semiconductor oxide type or hot-wire type, where the resistance of the sensor materials changes dramatically when sensor surface reduced or the temperature of the sensor materials increased, respectively. Most of them are saturated and useless for high hydrogen concentration. There are thin film sensors like field effect transistors (FETs) with Pd catalyst or chemo-resistive Pd alloy resistor, where hydrogen at the Pd-gate interface results in a change of FET channel' current or the sensor resistance. There are other hydrogen sensors with different operating principles and type, based on different materials. Among them — optical fiber type, piezoelectric type, thermoelectric sensors, wire coated by Pd, Schottky and MIS diodes, solid electrolyte, polymers, amperometric and potentiometric sensors, carbon nanotube and fullerene sensors, graphite oxide, different metal oxides etc. (see, for example, review-papers and books [1-6]). Even short discussion of achievements in the field of solidstate hydrogen sensors is not possible in the framework of this review. Only some of types of hydrogen sensors are considered below. We divided the hydrogen sensors on the following groups:
1. Catalytic Bead Sensors. Catalytic bead sensors consist of two beads surrounding a wire operating at high temperature (450 °C). One bead is passivated, so that it will not react when it comes into contact with gas molecules. The other is coated with a catalyst to promote a reaction with the gas. The beads are generally placed on separate legs of a Wheatstone bridge circuit. When hydrogen is present, there is no measurable effect on the passivated bead, whereas there is a significant effect on the catalyzed bead. The increase in heat increases the bridge balance signal, which is the sensor signal.
Pellistor-a coil of Pt wire-can be encapsulated in a bead (1-2 mm diameter). It is working at 800-1000 °C. Catalytic bead technology is generally used in the 1 to 5 percent hydrogen range. The response times of the sensor varies from 10 to 30 seconds. The cost of the technology also varies, depending on both the manufacturer and application. The cost of over $500 per sensor and a control unit cost of over $1,200 are typical [7].
2. Electrochemical Sensors. Electrochemical sensors are composed of an anode and cathode "sandwiching" a chemically sensitive electrolyte. When hydrogen passes over the electrolyte, a reversible chemical reaction occurs. That generates a current proportional to the gas concentration. To ensure chemical reversibility, oxygen is required; therefore the sensor is not environmentally independent. Electrochemical sensors are typically used
to detect hydrogen in the range of 100 to 1000 ppm of hydrogen. Response times can be as low as several seconds, although typically these sensors are specified at 30 to 50 seconds for full-scale response. Cost per sensing point is similar to the catalytic bead technology, averaging approximate- s ly $1,200 per sensing point [7]. ^
3. Resistive Palladium Alloy Sensors. For rap- * id response in a chemically variable environment, f palladium metal and its alloys are used. The sur- ^ face of palladium acts catalytically to break the H- | H bond in diatomic hydrogen and allows the mon- ^ atomic hydrogen to diffuse into the material. The | level of dissolved hydrogen proportionally chang- s es the electrical resistivity of the metal (see be- § low). No other gases or environmental controls © are necessary for these measurements. As an exception, when dealing with a platinum surface, some care must be taken to remove sulphur-bearing contaminants, generally through the use of non-hydrogen specific filters.
4. Hydrogen Field Effect Transistors and Schottky diodes. By using palladium as the gate material for a standard FET or Schottky diode, small changes in the resistivity of the palladium produce large changes in the current-voltage characteristics of FETs or diode. This sensor works well in the range of 50 to 1000 ppm range of hydrogen.
5. Semiconductor Metal Oxide Sensors. Solid state sensors utilizing semiconductor metal oxides generally operate at temperatures above ambient (see below). The electrical resistance of the sensor material will also depend on temperature and on the chemical composition of the surrounding atmosphere. Most oxides change their resistance with the oxygen concentration in the environment changes, making these sensors unsuitable for fuel cell feedstock monitoring applications.
These sensors do offer fast response in the range of 0-1000 ppm hydrogen, therefore they are useful in leak-detection applications on the end of a sensor wand. Price per sensing point is in the range of $300-500 per sensing point [7].
Below we did not consider in details first two £ types of sensors. i
a
Field Effect Transistors |
and Palladium Sensors !
^
u
a
T
Although field effect transistors are widely | used successfully more than 25 years as ion-selec- £ tive ISFET devices, biosensors and chemical sensi- £
IX
tive devices for detection of different ions in liq- * uid media and gases, including hydrogen, such s FET devices cannot be used in fuel cells in many 0 real cases. The Robust Hydrogen Sensor is an example of a sensor that could be used for these applications. The fundamental advantage of this palladium-based technology for these applications is its independence from the local environment, as well as its small size and ability to be used at high
pressures. The optimization of a hydrogen sensor or sensor suite to respond quickly to sensing requirements for fuel cells is still not completely solved. DCH Technology offered the Robust hydrogen sensors, made of triple redundant Pd-al-loy TF-resistor, hydrogen-FET-six orders of mag-^ nitude in hydrogen detection and heater circuits * on a single chip. It is necessary 6-10 such sensors f on the vehicle and 10 sensors monitoring a garage
^ operation. Low cost $10-15 is expected per sensu
| ing point according to the DHL Information about
^ Fuel Cell applications [7].
1 Note that palladium is already known as highly
cy
& hydrogen sensitive material also more then 25 § years. In such application, diffusion of hydrogen © through the solid state layer is important for the performance of the sensor. But the blistering de-structs Pd at remarkable high hydrogen concentrations [8]. Palladium hydride is formed in many cases and most of the Pd-based thin film sensors have the potential for degradation of their performance over cyclic use at high hydrogen gas concentration, inherently. In order to overcome such problems, in cited above paper of K. Scharnagl et al. and paper [9], for example, a hybrid suspended gate FETs with Pt, Pd and Ir were reported. Experimentally well established facts according Pd do not allow using really very attractive idea about a decrease of the value of the work-function of electrons from Pd with an increase of the hydrogen concentration. This is reason that we limited ourselves by the list of some last publications in the field of Pd-sensors (see [10-14]). Pd-polymer nano-composite films were deposited by co-sputtering of Pd and polymer targets. The testing gas was carried out in 0.5-10.0 % hydrogen in nitrogen [15].
Thermoelectric Gas Sensors
Thermoelectric gas sensor for detection of high hydrogen concentration was developed in [16, 17] and allows to detect wide range and high-concentration hydrogen. The thermoelectric gas sensor 3. was fabricated by depositing platinum catalyst thin g film on the half surface of nickel oxide thick film.
g- The sensor detected hydrogen for the concentrai
tion from 0.025 to 10 % in air, with a good lin-1 earity of voltage signal versus gas concentration. 1 By coupling this sensor and the air mixing set-up, 8 high hydrogen concentration gas of range up to ! 100 % (1 atm hydrogen) balanced with nitrogen i was measured with a good linearity. g A hydrogen gas sensor using thermoelectric © (TE) Li-doped NiO thin film with platinum catalyst film was also fabricated. When this sensor was exposed to air mixed with the hydrogen gas, catalytic reaction heats up the platinum-coated surface, and then TE voltage builds up along the hot and cold region of the oxide film. The Li-doped NiO thin film was deposited by rf sputtering method
and followed by thermal annealing. As-deposited NiO thin film possessed amorphous state and Pt thin film on this NiO film has no catalytic activity. Contrary to the as-deposited sample, thermal annealed NiO thin film has crystallized and Pt thin film on this NiO film shows usual exothermic temperature increase arisen from catalytic reaction. The hydrogen sensing measurement shows that the Pt-NiO thin film deposited on the MgO (100) substrate could realize negligible noise level and make clear voltage signal even at room temperature.
Optical Hydrogen Sensors
Novel electrochromic devices made of a tungsten oxide WO3 thin film integrated with amorphous SiGe:H pin photodetector for hydrogen sensor were developed in [18]. With the addition of the palladium Pd film to ionize hydrogen gas, the WO3 thin film will react with hydrogen ion and transfer from transparency to blue color. This color change will degrade the absorption of light with a wavelength larger than the blue color. Therefore, the photocurrent generated by a-SiGe:H pin photo-detector will be lowered down during the detection of thus detecting the existing hydrogen gas. Especially, the WO3 pin photosensitive hydrogen sensor also shows highly selectivity with hydrogen gas.
Optical sensors on WO3, which are sensitive in visible-near infrared part of the spectrum, as well as integrated with Si-Ge photodetector were also reported in [19-20]. Films composed of noble metals such as Pd, Pt, Au and WO3 were prepared by simultaneous rf and dc sputtering deposition. At temperatures of 200-250 °C, no change in the visible-near infrared absorption of the WO3 film without noble metals could be detected by the presence of 1 vol. % hydrogen in air. With the increase in the hydrogen concentration, tungsten bronze (H3WO3) can be formed, which is enhanced by the dissociation of hydrogen on Pd. The combinations Pt-WO3 and Au-WO3 did not make the film sensitive to hydrogen.
There is also some new development work in the area of optically sensing hydrogen using optical fiber technology. Several advantages occur using this approach: optical technologies are insensitive to noise; multiple measurements can be made using one optical box; and finally, optical technology allows using a different measurement technique in which the mapping of the hydrogen field can be done over long distances. Multiple sensing points with one processor is the key advantage of this technology. This advantage could prove to be cost-competitive over more conventional sensor technologies to monitor hydrogen levels in large bays with multiple hydrogen sources.
Room temperature Pd-coated WO3 fiber optical sensors were reported in [21, 22]. Hydrogen leak detection using an optical fiber sensor has been developed in [23] for the detection of hydro-
gen leakages on the cryotechnic engine of the European rocket ARIANE V. The principle of this sensor is based on the variation of reflectivity of a palladium micromirror deposited at the output end face of a multimode optical fiber. An experimental set-up using a monitored high power laser diode provides an optical heating of the palladium layer, which allows keeping up the performances in the wide range of temperatures. Fast and accurate detection of hydrogen concentrations inferior to the explosive limit has been performed.
Electrochemical Sensors
High-temperature SrCeO3- and CaZrO3-solid electrolyte oxides, which exhibit appreciable proton conduction under hydrogen-containing atmosphere at high temperatures, were investigated as hydrogen sensors in [24]. They are consisting of two electrochemical cells. One cell was employed for electrochemical pumping of hydrogen from hydrogen-containing atmosphere and the other was used for sensing of the ambient hydrogen with the pumped hydrogen as a standard gas. On applying a voltage above 2.5 V to the pumping cell, a sufficient EMF response against hydrogen partial pressure was observed over a wide range of pH2. Hydrogen sensors based on the change in electrical conductivity of high-temperature-type protonic conductor have been reported in other papers also. The PbF2-solid electrolyte H2, CO, NO and NO2 made and investigated in [25].
New information about typical electrochemical sensors can be found in [26, 27]. The potenti-ometric hydrogen gas sensor made of hydronium Nasicon, a hydrogen ion conducting solid electrolyte, is described in [26]. The device incorporates a silver-based reference electrode, which eliminates the need for a standard reference gas. The sensor is simple, fast, and capable of detecting hydrogen concentrations from at least 0.01 % to 100 %, a range of four orders of magnitude, in the absence of oxygen or other oxidizing agents. The room temperature performance of the sensor is possible.
A solid-state amperometric hydrogen sensor using Pt/C/Nafion composite electrodes was prepared by a hot-pressed method [27]. The results reveal that the sensitivity of the electrode increases with an increase of Pt loading. At a Pt loading of 3.0 mg/cm2, the sensitivities of the electrodes treated by direct wetting and indirect wetting methods are 0.263 and 0.716 mA/ppm, respectively, in the hydrogen concentration range of 12605250 ppm. This indicates that indirect wetting is a better method for moistening the hydrophobic electrode. Generally, both the response time and the recovery time decrease with increasing both the hydrogen concentration and the Pt loading. The electrode prepared in this study has the advantage of a higher catalytic surface area, which results in the 50 times higher sensivity than that
of electrodes prepared by other methods at the same Pt loading.
Note that conducting polymers are widely used for manufacturing of first electronic nose systems [28, 29], but seems their special use for hydrogen
detection is not known. New amperometric hydro- *
<
gen detector using Pd-polyaniline/oilyvinyl alco- j< hol-phosphoric acid/glassy carbon electrochemical I system is developed in [30]. Repeatable hydrogen g adsorption using nanostructured graphite at room -g
temperature is reported in [31]. |
d
\-
High Temperature Hydrogen Sensitive |
Schottky Diode Made of Silicon Carbide £
LP
C
Investigations of influence of hydrogen gas ° on properties of FETs made of SiC with Ru or Pt were carried out earlier, for example, in [32-36]. It was shown that SiC sensors can operate at rather high ambient temperatures. Pt/Ga2O3/SiC metal-oxide-silicon carbide devices operated as Schot-tky diodes were reported in [37]. The sensors have been tested towards different concentrations of hydrogen gas as a function of operating temperature. This study showed advantages of this structure compared to the pure thin film (90 nm) Ga2O3 conductometric sensor. The Ga2O3 thin films were prepared by the sol-gel process and deposited onto the transducers by spin-coating. For both types of sensors, the operating temperature was controlled by a micro heater located beneath the structure. At high temperature (above 500 °C), the response time of the sensors decreased.
Hydrogen sensing characteristics of a Pt-oxide-Al03Ga07As MOS Schottky diode are reported in [38].
Semiconductor Metal Oxide Sensors
All mentioned above versions of solid-state hydrogen sensors are still far from large-scale production. Their advantages should be still revealed. Today main gas sensors are fabricated from semiconductor metal oxide semiconductors. Their advantages and shortcomings, principles of the detection of gases and materials used for them are ^ well known and discussed elsewhere including our g previous review [1-4] etc. g-
After invention of the phenomenon of sensitiv- s ity of tin oxide SnO2 to different gases (N. Tagu- 1 chi, 1970), many efforts were made by T. Seiyama, | N. Yamazoe and other Japanese scientists in this | direction, which led to bringing such metal oxide | gas sensors into practice and organization of their 2 industrial production. Now scientists, engineers, g institutions and companies in many countries are © working in this field, which is dramatically extended and includes investigations of gas sensors also on other physical principles. Different gas sensing materials and multi-element arrays made of them (e-nose) proposed and are using today (including many semiconductor metal oxides and polymers).
Metal oxide low-cost semiconductor gas sensors are used in the continuous monitoring of emissions of small amounts of hazardous gases into the atmosphere and premises, for the detection of gases related to food quality (e.g. ethanol in fermentation processes, ethylene in fruit ripeness ^ determination/monitoring), safety and drags (nar-* cotics) control, to diagnose illness through breath | etc. In result of such achievements, obtained by -g many scientific groups in Japan, the USA, UK, I Italy, Germany and other countries, different com-^ panies were organized, which started produce and 1 use such semiconductor gas sensors in different
a
& ranges of human activity.
§ As it was mentioned above, gas detection is en© abled by a change in the electric resistance of semiconductors arising from surface phenomena. The response of the surface is dependent on many factors of metal oxide using in the sensor as the work body, its elemental composition including a doping, electronic and defect structure as well as its microstructure. Changes in the surface characteristics of the active layer can induce changes in the sensor performance. Note that the use of SnO2, Zr2O3 and other oxides as an active material for gas sensing is widely related to their combined use with catalytic metals particles dispersed on the surface of semiconductor oxides. The interaction between gas molecules and the surface of the sensing film is affected by many factors such as the operation temperature, gas being analyzed, sensor porosity, geometry and packaging. It is necessary to have long-term and successful stable and reproducing characteristics of sensors, their selectivity to detecting gases. Note that many questions were solved today. Serious achievements are obtained also in the field of polymer and acoustic gas sensors, but we will concentrate our attention below on metal oxide sensors based on the phenomena of adsorption of detected gas by the surface of a semiconductor, which leads to a remarkable change in electrical resistance (conductivity) of the sensor and allows measuring the concentration of interacting gas. « Analysis of the literature shows that today <t the main success is reached with gas sensors made g. of tin oxide and In2O3. It is the well-established I fact that SnO2 has an excellent potential for its 1 applications due to high capacity to adsorb gas-| eous molecules and promote corresponding reac-I tions (usually its oxidation) while showing chang-§ es in its surface conductivity. On the other hand, £ it is also well known that this material possesses g some disadvantages, such as poor selectivity and © high working (operating) temperatures. Many investigations were carried out in order to understand the role of the introduction of different dopants in SnO2. If the ceramic technology is used, such a doping is conducting by the use of additions of corresponding metal oxides. It was found that gas sensitivity and selectivity of tin oxide
sensors depend on type of different dopants. For example, it is established that additions of V, Nb and Ti in SnO2 increased the sensitivity of the sensor to C3H6, introduction of ThO2 and thermal annealing at 670 °C increased the sensitivity of SnO2 to carbon monoxide, addition of silver and annealing at 370 °C — to hydrogen, introduction of bismuth oxide — to carbon monoxide and methane etc. There are many investigations carried out on the Taguchi chemiresistors made of these metal oxide without dopants or specially doped (for example, tin oxide with dopants Pd, Pt, Mg, Th, Hg, In, Ni, Sb, Al, and Au) as well as SnO2 — In2O3 solid solutions. Gas sensitive Schottky diodes and metal-oxide semiconductor devices using these two main oxides and catalytic metals were investigated rather intensive.
Unfortunately, the maximum gas sensitivity of semiconductor SnO2 and In2O3 sensors is detected at temperatures which are well-above the room temperature. The literature data shows that the operating temperature of the SnO2 sensors lies usually in the range 250-400 °C, below and above of which the gas sensitivity is dramatically decreases. Note that the addition of Pt decreased the operation of ceramic tin oxide sensors up to 310 °C only with simultaneous absence of the selectivity to different gases. High operating temperature leads to an increase in consumed electric power and makes difficult and expensive large-scale use and manufacture of such sensors as demands special heating circuit and package to each sensor. Some of attempts to eliminate of some sources of interference (for example, by filtering the analyzed atmosphere, in the case of some sensors) have been successful, but the challenge of producing high performance SnO2 remains an open issue.
We mentioned above that the use of tin oxide SnO2 as working material for different gas sensors is very attractive, but there are many difficulties, which make the use of sensors made of SnO2 limited, although they are industrially produced now by several companies. Attempts to avoid these problems have included different approaches:
First one is the necessity to continue the search of other metal oxides, which will be more sensitive to different gases and have lower operation temperature.
Second, investigators try to reach a decrease in operation temperature and response time, enhance of sensitivity to different gases by combined use of metal oxide with catalytic noble metals like Pt and Pd as particles dispersed on the surface or dopants or Schottky barriers to metal oxide. As was mentioned above, many studies have evaluated the effect of such dopants on the morphology and properties of tin oxide. Note that cations of noble metals favor the adsorption of gaseous molecules, increase sensor sensitivity and the amount of reagent converted by the catalyst. This activity is most likely caused by the catalytic
properties of the noble metal itself and by the formation of adsorbing centers. However, according to the literature data, the interaction between the tin oxide and noble metals is difficult to understand and sometimes leads to non-stable results and poisoning of catalysts.
Third, the introduction of rare earth cations (usually trivalent cations) influences the tin oxide electronic structure, favoring, for instance, the adsorption of oxygen species and resulting in the manufacture of sensor materials with suitable morphological characteristics (high surface area, small particles, etc.). It is therefore expected that the concomitant use of noble metals and rare earth cations can produce some kind of synergetic effect, resulting in high performance tin oxide-based catalysts.
Doping of SnOm with 1.7% Si decreases the temperature of pre-heating of hydrogen sensor up to 150 °C [39].
N. Yamazoe et al. [5] studied the SnO2La2O3Pd, SnO2La2O3Pt, SnO2La2O3 and SnO2 systems for ethanol sensor applications. Their results indicated that the system containing Pt is the most sensitive, followed, in decreasing order of sensitivity, by the system containing Pd, the system doped only with La2O3 and, finally, the pure system. The SnO2La2O3Au system was also analyzed as the CO sensors. These authors proposed that the system is appropriate and selective for this application, suggesting that the effect of La is to reduce the number of acid sites (Sn4+) and that of Au is to produce specific sites for the CO adsorption.
Forth, it is established that the SnO2 thin film gas sensors exhibit enhanced sensitivity below 10 nm crystallite size; however, such films tend to show an increased response time with decreasing the nanocrystallite (grain) size below this size range. Existing gas sensing mechanisms, proposed for explanation the gas sensitivity enhancement below 10 nm crystallite size, can not satisfactorily explain the increased response time below this size range. The recovery time has also been observed to increase dramatically at lower operating temperatures (<200 °C). Any way, it was found very promising to develop new technologies like solgel, spin coating or thin film pyrolysis and sputtering in order to manufacture and investigate nanocrystalline tin oxide gas sensors. Recently, outstanding results were demonstrated on such sensors made by sol-gel technology [40-41]. Correspondingly, new problem is arising for such technology connected with the role of the calcinations temperature and film thickness. Both factors are important for the realization of high sensitivity and stability of nanocrystalline sensors.
Fifth, the degree of the porosity of working body and its surface is very important for realization of both gas, and ion-selective sensors made of tin oxide. The gas sensitivity is observed to increase with increasing of amount of film po-
rosity. It is demonstrated [41] that the gas sensitivity of the nanocrystalline tin oxide semiconductor thin film sensor increases from 100 to 900 % with increase in the amount of film porosity from 32 to 65 %.
Sixth, today any exhaustive and universally * mechanism and synthesis approach have not de- ^ veloped nor for tin oxide, nor other metal oxides. * Even for the most studied SnO2 the nature of sen- | sitivity, forms of chemisorbed species, ways to reach ^ the selectivity to different reducing and oxidizing | gases etc. remain debatable. ^
Seventh, today sensors on the SnO2 and other f metal oxide are mainly working at operating tem- S peratures 250-650 °C, which are well above room § temperature and pose major problems. It is impor- 0 tant also to eliminate high temperatures in the process of the manufacture of sensors (for example, during the calcination or pyrolysis procedures).
In the multicomponent gas mixtures of real applications, the metal oxides do not selectively react to the detected but they react to a wide range of gases with similar chemical behavior. Temperature step methods may help in certain applications, but their use is limited. To overcome these problems, intense work in the field of gas filters has been performed with high-temperatures stable gas filters located directly at the surface of the sensing thin film. The action of a gas filter may be two-fold:
- to hinder the interfering gas from reaching detection substance but to allow the gas to be detected to reach the sensor surface;
- to transform the gas mixture to be detected into mixture which favors the detectability of the target gas.
Since conventionally used charcoal filters located aside from the sensor element which collect unwanted gas species like solvents tend to become saturated, a catalytic action of the filter is needed. This filter type does not collect unwanted gases but it should transform them into a species not hindering the detection of the target gas using a catalytic conversion. By its definition, the catalyst itself is not consumed which means the filter will not saturated. <
Function and applications of metal oxide gas i sensors are discussed also in [42]. Cryogenic hy- p-drogen sensors are developed in [43]. ^
>s
s
Hydrogen Metal Oxide Sensors ^
s
i
We focused below on hydrogen sensors made £ of metal oxide semiconductors. Sensors are pro- | duced by several companies. The best among such * sensors are the Figaro sensors. Below characteris- s tics of the special sensor for hydrogen gas TGS 821 0 are given.
TGS hydrogen sensors cannot work at high humidity conditions, need in remarkable pre-heat-ing of working body (Pt-SnO2) and consumed electrical power, have low change ratio of sensor resistance and value of it, rather expensive standard
Electrical Characteristics
Item Symbol Condition Specification
Sensor Resistance Rs Hydrogen at 100 ppm/air 1k ~ 10k
Change Ratio of Sensor Resistance a log[Rs (H2 100 ppm)/Rs (H2 1000 ppm)] 0.6 ~ 1.20
log (1000 ppm/100 ppm)
Heater Resistance Rh Room temperature 38.0 ± 3.0
Heater Power Consumption Ph Vh = 5.0 V 660 ± 55 mW
Standard Circuit Conditions
Sensitivity Characteristics
Item Symbol Rated Values Remarks
Heater Voltage Vh 5.0 ± 0.2 V AC or DC
Circuit Voltage Vc Max. 24 V AC or DC *PS 15 mW
Load Resistance Rl Variable *PS 15 mW
circuit, price of which is 5-6 time expensive in comparison with own price of sensor, rather low selectivity to hydrogen in comparison to other gases. Today the price of one hydrogen sensor with corresponding circuit is more than $300-500.
Below we give a Table with the list of reference on some of important efforts published recently with mention of pre-heating temperature of working body of the sensor.
Hydrogen Sensors for Fuel Cells
Note again that all fuel cell technology ultimately utilizes hydrogen in a gaseous form. In
Rs/R0
100
0.01
1000 10000 Concentration, ppm
addition, fuel cells generally use platinum catalysts to break down the diatomic hydrogen gas into monatomic hydrogen ions. Different electrolytes are used in fuel cell technology to carry the monatomic hydrogen ions to O2 to form H2O. It is well known problems with the poisoning of platinum catalysts by other chemical species, therefore
Table
Metal oxide film (thickness) Temperature Reference (First Author, Source) Comments
1 2 3 4
SnO2 6 nm 350 °C N. S. Baik et al. Sensors and Actuators B 63(2000)74, 65(2000)97; Go Sakai et al. Ibid 77(2001)116
SnO2 8-10 nm 300 °C F. Lu et al. Ibid 66(2000)225
SnO2 15 nm 260-400 °C Ibid Q. Pan et al. 66(2000)237; Y. Wang et al. 85(2002)270
SnO2 3-15nm 400 °C F. Li et al. Ibid 81(2002)165
SnO2 High Temp Ibid Y. Shimizu et al. 46(1998)163; 52(1998)38; S. Nicoletti et al. 60(1999)90; M. Gaidi et al. 62(2000)43 Loading Pt, Pd
SnO2/Pt-Pd-SnO2 200 °C Ibid A. Kawahara et al. 65(2000)17; T. Hyodo et al. 64(2000)175; J.Woellenstein et al. 70(2000)196
SnO2 Cu/Pt 270-320 °C Ibid R. K. Sharma et al.72(2001)160 SiO2 Filter
SnO2 250-150°C Ibid V. A. Chauhary et al. 55(1999)154 Ru/Pd, Ru/Pd/Ag
SnO2/Si 250 °C Ibid J. M. Hamond and C. Liu 81(2001)25 10.6-100 % H2
Table
1 2 3 4
SnO2-Co3O4 200 °C Ibid U.-Sung Choi et al. 98 (2004)166 n- and p-type
SnO2-CdO Ibid Z. Tianshu et al. 60(1999)208
ТЮ2-&2О3 350-450°C Ibid Y. Li et al. 83(2002)160 p-type
TiO2:Nb; ТЮ2; TiO2-Pd 200-250°C Ibid M. Ferroni et al. 68(2000)140 J. Hayakawa et al. 62(2000)55 Y. Shimizu et al. 83(2002)195
ZnO nm-size 300 °C J. Xu et al. ibid 66(2000)277
SnO2-ZnO-CuO High Temp W. Moon et al. ibid 87(2002)464
Fe2O3-6 wt. % Ag2O 320 °C Wang. Ibid 84(2002)95; S. Tao et al. 61(1999)33
1П2О3 350 °C W. Chung et al. Ibid 65(2000)312; 46(98)139 spin-coating
In2O3-CuO 210-230 °C T. Belysheva et al. ISJAEE No.2(2004)60
Ga2O3 High Temp M. Fleischer Sensors and Actuatofs B 69(2000)205; Ibid R. Pohle et al. 68(2000)151; T. Weh et al. 146 and 78(2001)202 200-nm filter SiO2
NiO-TiOx 250 °C C. Imawan et al. Ibid68(2000)184 MagnetronSputter
MoO3-SiO2-Si 300 °C C. Imawan et al. Ibid 78(2001)119 MagnetronSputter
MoO3-V2O5 150 °C C. Imawan et al. Ibid 77(2001)346 MagnetronSputter
MoO3-TiO2 MoO3-WO3 350 °C Ibid K. Galatsis et al. 83(2002)276; E. Comini et al. 84(2002)26
W/Pt-GaN MOS and Schottky Diode 350-600 °C Ibid B. S. Kang et al. 104(2005)232
the hydrogen gas stream has to be kept as pure as possible to avoid degradation of the catalyst and concomitant performance degradation of the fuel cell. All fuel cells require two types of sensors; the first monitors the quality of the hydrogen feed gas, and the second more important sensor system, for leak detection. The demands on the sensors are different [7, 44].
Hydrogen leak-detection sensors must detect over the general level of ambient hydrogen levels, and in a variety of environments. For example, a sensor must be able to discriminate between ambient low-level sources of hydrogen and those which will be generated by a hydrogen leak. Successful commercial hydrogen-leak detection for the fuel cell marketplace will depend on the following factors: no false alarms; integrated autonomous fuel cell shutoff/venting procedures; multiplexed sensor arrays; and reliable sensing, calibration and self-testing. Leak detection can be accomplished by two types of sensors: one to detect the leak and a second alarm sensor with a set-point set at 50 percent of the Lower Explosive Limit (LEL) of 4 percent hydrogen in air. The need for leak detection is especially acute where there is a large volume of hydrogen consumed (for example, in residential/stationary applications and in automotive applications, where the fuel cell sits in an enclosed volume).
Explosive-limit sensors are continuously on, do not need to sense below a certain threshold value and are generally slow to respond. For such alarm sensors, the detection range is 0.5 to 4.0 percent hydrogen in air, with a response time of sev-
eral seconds to tens of seconds. Once a large-scale leak is detected, the buoyant nature of hydrogen means that it is easily dissipated, making it ultimately a safer gas to use than heavy flammable gases which tend to settle.
Spatial identification of a leak requires a sensor element is similar to other leak-detection methods. The sensor element itself must be faster acting to prevent operator fatigue, be able to sense down 1000 ppm range, and may need to have a larger upper detection limit, depending on application. In addition, there is a need to have a leak sensor which may be independent of the environment.
Fuel cells will likely develop faster in applications where electrical power is not available at reasonable cost, and so the initial energy source will be highly variable. Hydrogen sensors are then required to monitor the hydrogen content in the gas, as well as the pressure. Hydrogen concentrations can range from 40 to 99.9 percent, depending on fuel cell design, and should be to follow the fuel cells' power generation. In contrast to leak detection, explosive-limit monitoring, hydrogen sensors are also required to monitor hydrogen going into a fuel cell (hydrogen sensors for feed stock monitoring). Technical challenges originate from the multitude of potential generation sources for the gas. These challenges include hydrogen from reformers, tanks, electrolyzers, hydrogen storage media, biomass and other storage technologies. One of the key challenges is the ability to load track. Sensors are now tasked to respond quickly to serve as process monitors, as well as monitor the relative purity of the hydrogen gas stream. In
contrast to leak detection, these sensors must be capable of monitoring hydrogen at high-purity levels and pressures.
All sensors on the market suffer from slower response times (8 to 30 seconds) than the likely s duty cycles needed for most applications. Maturity ^ and specialization of these sensors will require sig-* nificant yet achievable cost reductions in the senf sors, as well as a better definition from the emerg-^ ing fuel cell market for the expected duty cycle I loads, hydrogen fuel stream pressures, flow rates ^ and composition. It will be excellent if heater cir-f cuit, which keeps the sensor elements at a nomi-S nally constant temperature, made on a single chip § in addition to the sensing elements. Sensors can
0 be multiplexed together using a sensing architecture developed for these applications in form an e-nose [45], and so the cost per sensing point reduces proportionally to the number of sensors required. For example, the number of sensors required for an automotive application is usually 6 to 10 sensors on the vehicle, and possibly up to 10 sensors monitoring a garage operation.
Experience of the Yerevan State University Team in the Field
We are working in the field of semiconductor ceramic chemical sensors since 1985 and have 56 publications in English in this field in peer-reviewed scientific journals and Materials of Conferences as well as 11 USSR, USA, Russian Federation and Armenian Patents. Some gas and smoke sensors were investigated, developed in YSU and made of Bi2O3, SnO2 and Fe2O3 thin films, BiFeO3, Bi4Fe2O9 and Bi2V4O11 ceramics as well as LaAlO3-CaTiO3 and NaBiTi2O6 thin films [4]. Note that operating temperature of all sensors made of SnO2 and Fe2O3 thin films as well as BiFeO3 and Bi4Fe2O9 ceramics was 300-400 °C. In case of sensors made of Bi2O3 and NaBiTi2O6 thin films as well as LaAlO3-CaTiO3 and Bi2V4O11, operating temperature was 20-70 °C. Sensors working at low temperatures were sensitive to smoke, ethanol and sulfur vapors as well as hui midity. Sensor made of Fe2O3 <K> was sensitive to t humidity at room temperature. | We have also long-term and successful experience in work with TiO2, Fe2O3, ZnO, CuO, Bi2O3 s and other simple and complicate metal oxides. For
1 example, we have investigated in details properties | of TiO2 <Ta, Mo, W, Nb, Re, Cr, V, Fe, Mn, Ni, | Co, Al etc.>, Fe2O3<Sn, Ge, Zr, Hf, Nb, Ti, Ta x etc.>, ZnO <Al, Y, In etc.> and other oxides as § well as solid solutions TiO2-MnO2, TiO2-Cu2O, CdO-® ZnO, Fe2O3-Nb2O5. We published about 70 papers
(including review-papers), part of which with Dr. J. Turner and other colleagues in NREL (see, for example, [46-48]. It is evident that these ceramic materials can be investigated in detail in order to understand its advantages and abilities for its use for the manufacture of thin film gas sensors.
We have also achievements in investigation of the phenomenon of influence of the porosity on different properties of silicon, which will be used during our activity in the framework of A-322 ISTC project. This knowledge will be applied to understanding of chemical and physical processes realized in chemical sensors. Among solving another problems, which at first glance did not connected with gas sensing, we focused our attention on further investigations and development of new type smoke detector made of bismuth oxide [49]. Our smoke detectors may become competitive with photoelectric and ionization detectors of smoke, which currently are widely used in fire-alarm systems. Note also that the development of the market of gas-sensing metal oxide semiconductor smoke detectors in Europe is stimulated by favorable legislation of European Union. It is connected with the fact that only high-sensitively smoke detectors can give the alarm at appearance of a smoldering seat of fire that may be effective for early detection of fire. So, in past we have developed and proposed a new version of high sensitive adsorption smoke detector made of thin-film sensor, functioning in a wide temperature range and at the surrounding relative humidity up to almost 100 %, with increased noise immunity and stability. The proposed detector highly competitive with known other adsorption (Taguchi) type smoke detectors in the sensitivity to smoke concentration and some other parameters, but does not need practically in the preheating of the work body of sensor. The lack of any humidity effect and temperature variation on the operation of the smoke detector allows reducing the sensors' dimensions and therefore minimizing consumed power. All this should lead to the smoke detector miniaturization as a whole, already to date having much less dimensions in comparison with large-scale produced detectors.
We carried out independent testing of the ad-sorptive type smoke detector in the USA according to corresponding new US standards in July 2004. Testing was carried out in the framework of financial support of the A-322 project. Our detectors conducted a testing according bilateral treaty with ISTC in Specialized Institute in the USA. Results are good; in some cases the parameters of our detectors exceeded last ones having other principle of the operation.
Today we have all necessary and sufficient technological equipment and corresponding setups and PC-controlled equipment for manufacture and corresponding measurements of the structure, electric and electronic properties of gas sensitive materials and prototypes of such metal oxide gas sensors. We study sensitivity of many of ceramic materials to hydrogen. We are ready to realize gas-sensitive metal oxide thin films with the thickness of the film less than 200 nanometers and nano grains with sizes equals to several nanometers as well as investigate noise phenomena in sensors [50-53].
Conclusion
The hydrogen economy exists and will grow quickly, as society evolves away from a petroleum-based energy infrastructure. These are exciting times for sensor and fuel cell companies that are developing this new energy technology.
Problem of manufacture of hydrogen sensors for fuel cells are very complicate, many research groups in Universities and Companies in the world are working today in this field.
Several factors will drive the designs for hydrogen sensors for fuel cells. As the portable fuel cell market increases, cost will drive the design to less expensive, robust and reliable sensors.
We can made the following conclusions:
■ Promising hydrogen-sensing technologies exist currently for the emerging fuel cell market.
■ Cost per sensing point is currently high, although as the volumes of fuel cells increase, the sensing costs will reduce proportionally. Any way, expensive palladium-or exotic polymer based sensor technologies offer a sensing solution over very limited range of technical requirements that are needed for fuel cells. Such sensors have non-stable characteristics.
■ It is necessary to the increase the selectivity, sensitivity and stability of hydrogen sensors and their consumed power, which can be realize by the decrease of the temperature of pre-heating of working body of the sensor.
■ Size of the sensitive area and components of the circuit of the sensor should be decrease for further integration of the sensor. By multiplexing the sensor units to a single control panel, additional economies can be realized.
■ There is the reason to carry our further search of new promising materials, mixed oxides and technologies, decreasing of the film thickness, physical and chemical investigations of role of the size of nanocrystallites (grains), interface and barrier phenomena, the thickness and porosity of thin films, operation temperature, contents of impurities, surface modification etc.
References
1. Chemical Sensors (Ed. by T. Seiyama). Ko-dasha Ltd. and Elsevier. Tokyo, 1981.
2. Solid State Gas Sensors (Ed. by P. T. Mose-ley and B. C. Tofield) A. Hilger Bristol, 1987.
3. W. Gopel et al. Chemical and Biochemical Sensors // Sensors. A Comprehensive Survey. Vol. 2.VCH Weinheim, N.Y., Basel, Cambridge 1991.
4. Aroutiounian V. M. // Sov. Microelectronics. 1991. Vol. 20. P. 337.
5. Yamazoe N. // Sensors and Actuators. 1991. Vol. B5. P. 7.
6. Gusev A. L. et al. // ISJAEE. 2002. No. 6. P. 12.
7. Peter Jardine A. // DarnellCom PowerPulse, 2000.
8. Scharnagl K. et al. // Sensors and Actuators. 2000. Vol. B80. P. 163.
9. Katti V. R. et al. // Sensors and Actuators. 2002. Vol. B84. P. 219.
10. Tabuzzi A. et al. // Ibid. Vol. B83. P. 175.
11. Chen H.-I. et al. // Ibid. Vol. B85. P. 10.
12. Nakano S. et al. // Sensors and Actuators. 2004. Vol. B104. P. 75.
13. Tabib-Azar M. et al. // Sensors and Actuators. 1999. Vol. B56. P. 158.
14. Kalygina V. M. et al. // Semiconductors. 2002. Vol. 36. P. 1215.
15. Fedtke P. et al. // Sensors and Actuators. 2004. Vol. B100. P. 151.
16. Matsumiya M. et al. // Thin Solid Films. 2002. Vol. 419. P. 213.
17. Shin W. et al. // Sensors and Actuators. 2004. Vol. B97. P. 344.
18. Li K. et al. // Sensors and Actuators.
2000. Vol. B69. P. 96.
19. Ando M. et al. // Sensors and Actuators.
2001. Vol. B76. P. 13.
20. Davazoglou D., Drtsas // Ibid. 2001. Vol. B77. P. 359.
21. Tabib-Azar M. et al. // Sensors and Actuators. 1999. Vol. B56. P. 158.
22. Sekimoto S. et al. // Sensors and Actuators. 2000. Vol. B66. P. 142.
23. Bevenot X. et al. // Sensors and Actuators. 2000. Vol. B67. P. 57.
24. Katahiraa K. et al. // Sensors and Actuators. 2001. Vol. B73. P. 130.
25. Alejnikov N. N. et al. // ISJAEE. 2004. No. 1. P. 54.
26. Maffei N., Kuriakose A. K. // Sensors and Actuators. 1999. Vol. B56. P. 243; 2004. Vol. B98. P. 73.
27. Yu-Chuan Liu et al. // J. of The Electrochemical Society. 2002. Vol. 149. P. H173.
28. Slater J. M., Paynter J., Watt E. J. // Analyst. 1993. Vol. 118. P. 379.
29. Freund M. S., Lewis N. S. // Proc. Natl. Acad. Sci. USA. 1995. Vol. 92. P. 2652.
30. Abalyaeva V. V., Efimov O. N., Gusev A. L. // ISJAEE. 2002. No. 6. P. 23.
31. Kajiura H. et al. // Appl. Phys. Lett. 200. Vol. 82. P. 1929.
32. Spetz A. L. et al. // Sensors and Actuators. 2000. Vol. 70. P. 67.
33. Boehme O. et al. // Adv. Mater . 2001. Vol. 13. P. 597.
34. Kim C. K. et al. // Sensors and Actuators. 2000. Vol. B66. P. 116.
35. Roy S. et al. // J. of The Electrochemical Society. 2003. Vol. 150. P. H135.
36. Pascal T. et al. // Sensors and Actuators. 2004. Vol. B98. P. 63.
37. Trinchi A. // Sensors and Actuators. 2004. Vol. B100. P. 94.
38. Cheng C.-C. et al. // Sensors and Actuators. 2004. Vol. B99. P. 425.
39. Gusev A. L. et al. // ISJAEE. 2002. No. 6. P. 12.
40. Cabot A. et al. // Sensors and Actuators. 2000. Vol. B70. P. 87.
41. Shukla S. et al. // Sensors and Actuators. 2004. Vol. B97. P. 256.
42. Kohl D. // J. Phys. D.: Appl. Phys. 2001. Vol. 34. P. R125.
43. Gusev A. L. // ISJAEE. 2003. Spec. Issue. P. 110.
44. Simon I., Arndt M. // Sensors and Actuators. 2000. Vol. A97-98. P. 2104.
45. Gardner J. W., Bartlett P.N. Electronic Noses: principles and applications. Oxford Science Publications, 1999.
46. Aroutiounian V. M., Arakelyan V. M., Shah-nazaryan G. E. // Solar Energy. 2005. Vol. 78.
47. Aroutiounian V. M. // Sov. Phys. Uspe-khi. 1989. Vol. 32. P. 521.
48. Aroutiounian V. M., Arakelyan V. M., Shahnazaryan G. E., Turner J. A. et al. // Int. J. of Hydrogen Energy. 2002. Vol. 27. P. 33.
49. Adamyan A. Z., Adamyan Z. N., Aroutiou-nian V. M. // Sensors and Actuators. 2003. Vol. B93. P. 416.
50. Hakhoyan A. P., Aroutiounian V. M., Adamian Z. N. et al. // Proc. Int. Symp. on Environmental and Industrial Sensing. USA 4576-39, 2001. P. 227.
51. Adamyan A. Z., Adamian Z. N., Aroutiounian V. M. // Sensors. 2003. Vol. 3. P. 438.
52. Aroutiounian V. M., Hakhoyan A. P., Ada-myan A. Z., Adamian Z. N., Barseghyan R. S. // Proc. 17th European Conf. On Solid-State Transducers. Guimaräes, Portugal, 2003.
53. Asriyan H. V., Gasparyan F. V., Aroutiounian V. M. et al. // Sensors and Actuators. 2004. Vol. A113. P. 338.
, j-'Tj- T
НОВОСТИ НАУКЙ И ТЕХНИКИ
водородный мотоцикл на топливных элементах
Британская компания «Intelligent Energy» разработала мотоцикл, работающий на водороде. Представленное устройство получило название «Emissions Neutral Vehicle» (ENV).
«Сердцем» двухколесного транспортного средства является отсоединяемая база с топливными элементами и резервуаром для хранения водорода, изготовленным из композиционного материала на основе углеродного волокна. Топливные элементы вырабатывают энергию для небольшого электромотора мощностью 6 кВт, который и приводит мотоцикл в движение. Время разгона до 50 км/ч составляет примерно 7,5 с, максимальная скорость — 80 км/ч. Запаса водорода теоретически должно хватить либо на 160 км пробега, либо на четыре часа непрерывной работы. Кстати, благодаря тому, что база с топливными элементами отсоединяется, генератор может применяться для питания какой-либо другой техники, например, моторной лодки.
Главным достоинством мотоцикла «Emissions Neutral Vehicle» является полное отсутствие вредных выбросов в атмосферу. Единственный побочный продукт (водяные пары) настолько чист, что, по заявлениям разработчиков, может напрямую использоваться для получения питьевой воды. Кроме того, водородный мотоцикл практически не производит шума. Кстати, это достоинство, по мнению некоторых, может обернуться недостатком и послужить причиной дорожно-транспортных происшествий. Впрочем, отмечают в «Intelligent Energy», мотоцикл можно без проблем оборудовать динамиком, воспроизводящим искусственный рев двигателя.
В случае начала массового производства стоимость водородного мотоцикла будет составлять порядка $8300.
Источник: http://rus.delfi.ee/
водородные автобусы «ford» для аэропортов
8-10 марта в Лас-Вегасе прошла выставка оборудования для аэропортов . «Ford» приурочил к ней показ экологически чистых автобусов E-450. Автомобили оснащены адаптированными к работе на водороде двигателями V10 внутреннего сгорания и позиционируются как транспортные средства с нулевой эмиссией. Машины рассчитаны на перевозку 12 пассажиров, дальность пробега на одной заправке составляет 250 км.
В рамках сотрудничества с «Ballard» компания отрабатывает классические технологии с применением топливных элементов. По такой схеме, в частности, создано третье поколение водородных автомобилей «Focus», представленных в США осенью минувшего года.
Источник: http://zr.ru/
