ROOM TEMPERATURE HYDROGEN SENSOR BASED ON NANO-MICRO INTEGRATION FOR SPACE EXPLORATION

Transcription

1 ROOM TEMPERATURE HYDROGEN SENSOR BASED ON NANO-MICRO INTEGRATION FOR SPACE EXPLORATION PI: S. Seal (University of Central Florida) Co-PI: H.J. Cho (University of Central Florida) S. Shukla (University of Central Florida) L. Ludwig (Kennedy Space Center) P. Zhang (University of Central Florida) S. Deshpande (University of Central Florida) C. Drake (University of Central Florida) Abstract We are developing the sol-gel derived nanocrystalline indium oxide (In 2 O 3 )-doped tin oxide (SnO 2 ) sensor, in the form of thin film/nanowires/nanofibers, for room temperature hydrogen ( ) sensing application, for NASA, under the atmospheric conditions existing on the surface of the Earth. The nanocrystalline thin film sensor is incorporated into the microelectromechanical system (MEMS) device to achieve high sensitivity and selectivity with minimum detection and recovery time at room temperature. Effect of various test parameters such as the air pressure, the concentration as well as the MEMS design parameters such as the finger spacing and the number of fingers on the room temperature sensing characteristics of the present sensor has been demonstrated. The present nano-micro integrated sensor shows giant room temperature sensitivity (S= ) with high selectivity over CO. The current detection and recovery time at room temperature lie within the range of sec and sec respectively. New technological solutions for further reducing the response and the recovery time of the present nano-micro integrated sensor have been proposed. Sensor tests are underway to test the sensitivity of the present nano-micro integrated sensor under the atmospheric conditions existing on the surface of Mars and Moon. Introduction Hydrogen ( ) is the most abundant element in the universe and one of the most abundant on Earth. Due to the rapid consumption of the fossil fuels, much attention has been paid towards as an economical non-conventional energy source for the diversified industrial applications. For example, solid oxide fuel cell (SOFC) technology uses gaseous for the generation of power and heat. powered cars and buses are already in normal transit service in some of U.S. cities. Liquid has been used by NASA for launching the space-shuttles. As summarized in Figure 1, also finds applications in electronic, metallurgical, pharmaceutical, nuclear fuel, food and beverages, as well as glass and ceramic industries. Every day, millions of pounds of are used by hundreds of industries around the world. Due to the realization of the potential use of energy further interest has grown into the production of large quantity of different forms of, the enhancement of storage capacity, and the development of safe transportation system for. Depending on the quantities required, can be transported by road tanker or pipeline. North America alone has at least 700 km pipeline system. Pipelines for liquid have also been built by NASA for direct delivery of to the space vehicle at the launch pad. However, due to its very small size, is the most susceptible for the leakage through

2 the pipelines; typically about 1-3 % of in the existing systems is always lost, mostly through the joints in the pipes. If handled carelessly, is as dangerous for transport, storage and use as many other fuels. As a result, safety remains a top priority in all the aspects of energy and has been the prime motivation for the present work. In the recent years, nanotechnology has emerged as an attractive field for the development of novel materials having unusual properties, which have provided different pathways to solve many unresolved issues in various other fields. We strongly believe that the application of nanotechnology to sensors would help in advancing the science and the technology related to the development of sensor materials. Background Different experimental (metal-oxide-semiconductor (MOS)-based, catalytic resistor, acoustic wave and pyroelectric) and commercial sensors (catalytic combustion, electrochemical, semiconductor, and thermal conductivity) based on different principles are currently available but with major drawbacks as outlined in Figure 1. The very low sensitivity of these sensors at room temperature to low concentrations of has been invariable associated with the poor response and recovery time, which insists further investigation in these areas. In addition to this, the poor selectivity, which is a severe problem at room temperature, has been another major but pending issue, which needs to be delved in great detail. Particularly, attention must be paid in improving the selectivity by a novel approach without sacrificing the room temperature sensitivity and the response and recovery time of the sensor. Moreover, many of the experimental as well as commercially available sensors use the nanocrystalline materials, which are susceptible to changes in their physical properties (such as nanocrystallite size) if operated at highly temperature, which may reduce the potential life of the sensor. Such sensors, hence, must be operated at lower temperatures, where further research is still awaited. In Table 1, we summarize the list of companies, which manufacture and sell the sensor devices. Various characteristics of these sensor devices have also been shown for comparison. It appears to us that, these commercially available sensor devices are claimed to sense at room temperature within the concentration limits of 200 ppm-2%, which large enough for any practical application. Response time quoted for these sensor devices is less than 10 sec, but it is related to high concentration as high as 90% and not for ppm-level concentration. Although the sensor devices are claimed to operate at room temperature, the recovery time often is associated with higher temperature (70 o C). Moreover, the room temperature sensitivity values for these devices are often not quoted. In addition to this, there are only few manufacturers, which guarantee the cross-sensitivity to other poisonous gases. No claims are made regarding the suitability of these sensor-deices for sensing on other planetary conditions, which is an essential requirement for NASA. It is suggested that, the commercial sensors currently available in the market are designed only to meet the atmospheric conditions on the earth s surface. Modifying the sensor material properties to meet the NASA s over all requirements is imperative. Since our efforts are mainly focused on improving the semiconductor oxide based sensors, the current status of this particular class of sensor has been summarized in Table 2. Various forms of sensors such as thin films, random network of nanowires and

3 WHY ARE HYDROGEN SENSORS NEEDED? Electronic Industries Carrier Gas for Active Trace Elements as Arsine and Phospine for Manufacturing Semi- Conducting Layers in Integrated Circuits. Nuclear Fuel Industries Protective Atmosphere for Fabrication of Fuel Rods Metallurgical Industries Reducing Atmosphere Steel Making Copper Brazing Aerospace Industries Launching Space Shuttles Power Life Supports and Computers Produce Drinkable Water as Byproduct Hydrogen Applications Chemical and Petroleum Industries Production of Ammonia and Methanol Sulfur Removal Catalytic Converter Pharmaceutical Industries Production of sorbitols used in cosmetics, adhesives, surfactants, vitamins a and C. Food and Beverages Industries Hydrogenating Liquid Oils Converting them into Semisolid Materials Glass and Ceramic Industries Manufacturing Float Glass to Prevent Oxidation of Large Tin Bath. Hydrogen Production Crude Oil, Coal, Natural Gas Nuclear Solar, Hydro, Wind Wave, Geothermal Wood, Organic Waste, Biomass Hydrogen Storage Metal Hydrides Liquid Hydrides Carbon Nanotubes Compressed Hydrogen Hydrogen Transportation Specially Designed Tanker Trucks Pipelines STATUS OF HYDROGEN SENSORS Exiting Hydrogen Sensors Experimental MOS Structures Catalytic Resistors Acoustic Wave Pyroelectric Commercial Catalytic Combustion Electrochemical Semiconductor Thermal Conductivity Precincts O Sensitive Film Damage at High Hydrogen Concentrations Hydrogen Induced Drift Insensitive to Low Concentrations Poor Response at Room Temperature Limited Hydrogen Detection Range Poor Selectivity Cannot Operate in Helium and in Vacuum Large Size Figure 1. Chart summarizing the need for an immediate development of hydrogen sensors based on innovative approaches for overcoming the limitations of the current sensor technology and meeting the requirements of the hydrogen based industries.

4 Table 1. Summary of characteristics of commercially available sensors. Company Sensor Range Temperature Range ( o C) Humidity % Power Response/ Recovery (sec) Fuel Cell- Sensor ppm ppm 80 <95 1 W 12 V DC <10 sec 90% Concentration Scan LLC Pd/Ni Thin Film 0.5%-10% 70 (Gas Temp.)/1 atm 0-40 (Operating) <95 At 40 o C 2 W V DC In sec Depending on Concentration Neodym Technologies RKI Instruments MOS <2% At 40 o C MOS with Molecular Sieve+Transmittor 2000 ppm- 2% 1 W 12 V DC 4 (Recovery 10) Alarm Point 2000 ppm -9 to V DC 20 sec for 90% Concentration Applied Nanotech. Inc. Pd Nanoparticles 0.5-2% In Volume microwatts <10 sec (Recovery <10 sec at 70 o C) Arrgh! Manufacturing Inc. - 1% -10 to V DC - Industrial Sci. Corp <99 3 V DC - Enmet Co ppm Nanotubes, as well as aligned or single nanowires of semiconductor oxides have been developed for the sensing application. It seems that, the maximum sensitivity at room temperature has been reported for the thin film form of the sensor with the response and the recovery time in few minutes. On the other hand, a single nanowire type sensor exhibits the response time in few seconds but with extremely low sensitivity. The random network of nanowires, having large porous structure, may be a good choice for selecting the appropriate form of the sensor, which may compromise the sensitivity to some extent for improving the response time. Recently, the single-walled carbon nanotubes (SWCNTs) and multi-walled carbon nanotubes (MWCNTs) have been synthesized and utilized for the (and other gases as well) sensing application. Some of the typical gas sensing results reported in the literature for these new gas sensing materials are tabulated in Table 3 and can be compared with the gas sensing properties of semiconductor oxides gas sensors, Table 2. Three different forms of carbon nanotubes such as single, parallely aligned, and

5 random network have been investigated for the gas sensing. Comparison reveals that, relative to the semiconductor oxides gas sensors, the carbon nanotubes exhibit very low gas sensitivity (<2) and high detection time ( sec). Moreover, the recovery time associated with the carbon nanotubes based gas sensor has been few hours. Table 2. Typical gas sensing results reported recently for the various forms of semiconductor oxides and our results Sensor Material Sensor Form Synthesis Method Operating Temperature ( o C) Gas (Amount) Sensitivity (R air /R gas ) Response Time (Sec) SnO 2 Single Nanobelt Vapor Phase Evaporation 200 C 2 H 5 OH (250 2 Few Seconds NO 2 ( SnO 2 /Pd Single Nanowire Thermal Evaporation 200 (?) Pd Single Nanowires Electrochemical Deposition 25 (5 %) (msec) In 2 O 3 Random Network of Nanowires Carbothermal Reduction 370 C 2 H 5 OH ( ZnO Random Network of Nanowires Thermal Evaporation 300 C 2 H 5 OH ( TiO 2 Nanotubes Array Anodization 290 ( In 2 O 3 - SnO 2 SEAL Thin Film Sol-Gel 20 (

6 Table 3. Typical gas sensing results reported recently for the various forms of carbon nanotubes. Sensor Material Sensor Form Synthesis Method Operating Temperature ( o C) Amount of Gas Sensitivity (R air /R gas ) Response Time (Sec) SWCNT- Pd Single Tube Patterned CVD Growth RT ( SWCNT Parallely Aligned PECVD 165 NO 2 (100 ppb) SWCNT- PABS Random Network Arc Discharge 32 NH 3 ( SWCNT -Pd Random Network Arc Discharge ( %) MWCNT Random Network Modified PECVD 25 NH 3 ( CNT Random Network PECVD 165 NO 2 (100 ppb) 0.56 Few Minutes Experimental (a) Sensing Tests The MEMS devices, which utilize an oxidized Si-wafer (Si-SiO 2 ) as the platforms, were patterned with four interdigitated (Au) electrodes, using thermal evaporation, photolithography, and wet chemical etching techniques. The MEMS devices were designed with the different number of fingers (8 and 20) and different finger spacing (10 μm and 20 μm). The tin-isopropoxide solution in iso-propanol and toluene, corresponding to the concentration of 0.23 M of tin-isopropoxide, was used with the addition of calculated amount of indium(iii)-isopropoxide to obtain the thin films of SnO mol% In 2 O 3 via a sol-gel dip-coating process. The dried gel films were sputtered with a thin Pt-layer for 10 sec using a sputter-coater. The coated-mems devices were dried at 150 o C for

7 15-30 min in air. The dip-coating, the sputtering, and the drying processes were repeated to obtain a desired film thickness. Finally, the Pt-sputtered dried gel films were fired at 400 o C in air for 1 h and utilized for the characterization and sensing tests. The coated and calcined MEMS devices were wire-bonded to an integrated circuit chip and installed in the 32 pin socket assembly, which was in turn placed centered over the sensor test-board designed using the LPKF CircuitCAM 4.0 software and cut using the LPKF Boardmaster 4.0 software on a single-sided copper clad prototype boards. All sensing tests were conducted in the dynamic test condition at room temperature (22 o C with the relative humidity of 35-50%). In this type of sensor-testing, the air-pressure within the test-chamber was reduced and maintained at a desired level using the turbo-pumps. A mixture of appropriate amounts of nitrogen (N 2 ) and was admitted into the test-chamber through the respective mass-flow-controllers. The N 2 (15000 was used as a carrier-gas. The amount of in ppm was calculated using the ratio of number of moles of admitted into the test-chamber per minute to the total number of moles of gas molecules (that is, the summation of number of moles of N 2,, and air) within the test-chamber. Thus, in the dynamic test condition, a desired amount of was continuously blown into the test-chamber per minute and simultaneously pumped out of the test-chamber throughout the test-duration. Thus, the dynamic test condition simulates the condition, which may be encountered in an actual service application, for example, leakage through a pipe line. (b) Nanowires/Nanofibers Sensor Material Development Two different techniques, electrospinning and thermal evaporation, have been utilized to synthesize the nanocrystalline SnO 2 -based nanowires/nanofibers. In the electrospinning Figure 2, tin (II) chloride (SnCl 2 ) precursor was dissolved completely in a highly concentrated polymeric solution, which was then taken in a syringe placed on a syringe pump, that fed the syringe tip with the polymeric solution at a constant speed. A Cu-plate covered with Al-foil was placed in front of the syringe at a distance of 10 cm and a high voltage (15 kv) was applied in between the Cu-plate and the syringe tip. The Syringe Pump Syringe Metallic Syringe Tip Jet Bending Instability of Jet Polymer Solution Taylor Cone Cu-Plate Covered with Al-Foil kv - High Voltage Power Supply Figure 2. Schematic of electrospinning process for nanowires/ nanofibers formation.

8 polymer fibers were drawn from the syringe tip, which were subsequently deposited on the Al foil (or the Si-SiO 2 substrate). Since, the polymer fibers contain the Sn-precursor, low temperature calcination temperature burnt off the polymer leaving behind the inorganic SnO 2 -based nanowires/nanofibers. In the thermal evaporation technique, Figure 3, a high temperature furnace was utilized. The Sn-precursor powder and the Si-SiO 2 substrate (with Pt catalyst) were taken into the Al 2 O 3 crucible, which was placed in the center of the furnace. Argon (Ar) gas was blown continuously into the furnace. The furnace was ramped to 900 o C, held at that temperature for 5 h, then cooled naturally to room temperature. Furnace Ar Out Ar In Substrate Precursor Powder Figure 3. Schematic diagram describing the thermal evaporation process for nanowires/nanofibers formation. Results and Discussion (a) Room Temperature Sensing Characteristics of Nano-Micro Integrated Sensor As shown in Figures 4 and 5, the present nano-micro integrated sensor shows very high sensitivity at room temperature as high as Moreover, the sensor being insensitive to CO at room temperature, it exhibits very high selectivity over CO. As shown in Figure 6, the response time of the present sensor lies within the range of sec and the recovery time lies within the range of sec. The variation in the response kinetics of the present nano-micro integrated sensor is presented in Figure 7. Improved response kinetics with increasing concentration within the range of ppm is noted. Within this range room temperature sensitivity lies within the range of As demonstrated in Figure 8, the room temperature sensitivity of the present sensor is almost insensitive to the air pressure level within the range of Torr. The sensor is also sensitive to at room temperature in the helium (He) atmosphere, Figure 9. The room temperature response kinetics of the present nano-micro integrated sensor is superior for smaller finger spacing, Figure 10.

9 1.0E Sensor Resistance (Ω) 1.0E E E+06 (b) 1.0E Figure 4. Room temperature (900 sensing with high sensitivity. Sensor Resistance (Ω) 1.0E E E E CO 1.0E E Figure 5. Room temperature selectivity over CO. 1.00E Sensor Resistance (Ω) 1.00E E H E Figure 6. Low room temperature response ( sec) and recovery time ( sec). (Note: Data points are separated by 50 sec time interval.)

10 Sensitivity (R air /R gas ) Figure 7. Room temperature sensing for different concentration levels. Sensitivity 1.0E E E E Sensor Resistance (Ω) E E E E E E E He Atmosphere Figure 8. Room temperature sensing for different air pressure levels ( Torr). UV-OFF UV-ON Figure 9. Room temperature B B sensing in He atmosphere.

11 Sensitivity (R air /R gas ) μm 20 μm Figure 10. Comparison of room temperature response kinetics for two different finger spacing. (b) Nanowire/Nanofiber Sensor Material Development The nanocrystalline SnO 2 based fibers deposited on the MEMS device via electrospinning technique are shown Figure 11; while, that derived using the thermal evaporation process are shown in Figure 12. Determining the room temperature sensing characteristics of the present nanofibrous sensor is under investigation. (a) (b) 500 μm 1 μm Figure 11. SEM images of SnO 2 nanowires synthesized via electrospinning. In (a), microelectromechanical system (MEMS) device with interdigitated gold (Au) electrodes is seen.

12 2 μm Figure 12. SnO 2 nanowires synthesized via thermal evaporation technique. Future Work - Utilize nanoelectromechanical system (NEMS) design (instead of MEMS) for reducing the room temperature response and recovery time below 60 sec. - Utilize the nanofibrous SnO 2 -based sensor for the room temperature sensing. - Utilize the thin film as well as nanofibrous sensor to sense under the atmospheric conditions existing on the surface of Moon and Mars. - Develop a prototype sensor device for NASA operating at room temperature. March 2006