Comment

01 August 2004

Metal oxides at industry cusp

By Debra Deininger, Stephen Williams, and Clayton Kostelecky

Gas sensor technology possesses innate advantage for process control applications.

Nitrogen oxide (NO_x) emissions end up as acid rain. To help reduce acid rain, the Environmental Protection Agency (EPA) has a two-phased strategy to cut NO_x emissions from coal-fired power plants.

The Clean Air Act requires states to reduce ground-level ozone. Because NO_x and ozone travel long distances, the act also requires "upwind" states to implement programs that will help "downwind" states meet the ozone standards.

The EPA issued a rule that requires 22 states and the District of Columbia to revise their implementation plans to further reduce NO_x emissions by taking advantage of newer, cleaner control strategies and monitoring capabilities.

Extensive knowledge of materials chemistry and nanoengineering has led to the development of a mixed metal oxide semiconductor that is extremely sensitive to NO_x vapors, offering sub-parts-per-million (ppm) detection limits and enhanced selectivity. This NO_x sensor technology has demonstrated good performance and long-term durability.

Addition of dopants to base

The development of metal oxide semiconductor sensors traces back to 1952, when J. Bardeen and W.H. Brattain discovered that surrounding gases could modulate the electrical conductance of oxide semiconductors. T. Seiyama was among the first to point out that semiconducting oxides can work as gas sensors.

The past three decades have seen significant development of empirical and trial-and-error techniques to develop novel solid-state sensors. Recently, work has begun to identify the basic principles and phenomena that occur in a semiconducting sensing material.

In this present work, tungsten oxide (WO₃ )-based materials serve to create an NO_x sensor. WO₃  has a great affinity for NO_x. The addition of dopants to this base metal oxide material can enhance the properties of the sensor through modification of the conductivity (via the addition of excess electrons in the conduction band or through the addition of holes) and/or influence on microstructure (porosity, grain size, and grain boundaries).

Catalysts work in conjunction with conventional solid-state sensors to promote a response to a particular gas, thereby improving sensitivity and selectivity. For example, noble-metal catalysts often work to detect nonpolar organics, because they form strong homopolar or ligand field bonds with the adsorbate. It is very important that the catalyst material spread evenly over the semiconductor surface to ensure good performance. The present sensor contains a proprietary blend of dopants and catalysts to optimize the sensitivity and selectivity of the sensor's response to NO_x.

Via resistance welding

The solid-state NO_x sensors described herein exist in a commercial electronics package, referred to as a TO-39 header and cap. Because the sensors operate at relatively high temperatures of 200–350°C, the sensor element suspends in a way that minimizes heat transfer between the sensor element and the package. The sensors heat via a resistive material screen printed on the undersurface of the sensor elements. In this application, a commercially available ruthenium oxide paste forms the heater.

Solid-state NOx sensor

The electrical connections between the sensor element and the package are formed via resistance welding, leading to strong, reliable connections that are unaffected by temperature and the chemical environment. Each sensor has four electrical connections. Two measure the resistance of the sensing material, while two provide power to the resistive heater.

Solid-state metal oxide sensors are typically operated using a simple voltage divider. This requires two voltage supplies: heater voltage (V_H) and circuit voltage (V_C). The system applies V_H to the heater to maintain a constant, elevated temperature for optimum sensing performance. V_C allows a measurement of the output voltage (V_out) across a load resistor (R_L ).

One can figure sensor resistance (R_S ) using the following formula:

R_S = ((V_C –V_out) / V_out) R_L

In some cases a potentiostatic circuit, which applies a constant voltage to the sensor, is preferred. The applied sensing voltage may range from less than 100 millivolts to more than 6 volts, and the optimum input voltage may depend on the sensor chemistry.

Simple voltage divider circuit

Thus, inexpensive solid-state sensors with good performance characteristics, including extremely high sensitivity, are here. The sensors are rugged, inexpensive, and unaffected by relatively high temperatures. These sensors may be useful in applications where inexpensive, long-lived, and durable sensors are required, such as process control.

The sensors do require significant power input (several hundred milliwatts) and thus are not suited for portable applications requiring battery operation.

The sensors are also not suitable for applications where one must measure levels of NO and NO₂  independently. For applications requiring measurement of total NO_x, the sensor response can adapt to individual circumstances via variations in sensing and heating potential to optimize the sensor response for an intended application.

In this way, both high sensitivity and high range detection are possible with a single, low-cost sensor.

Commercial-grade performance

The use of nanostructured material as the active sensing element significantly improves sensor performance.

Here is a comparison of the response of a WO₃-based sensor prepared from commercial coarse-grained powders and one prepared from a nanostructured material (grain size less than 100 nanometers).

The sensors were in an air background, and 8 ppm NO₂ entered the sensing area from 200 to 400 seconds. The introduction of NO₂ causes an increase in the sensor resistance.

To facilitate comparison between two different sensor materials, the sensor response plots as RG/RA, the resistance at any point divided by the average resistance in air. This clearly shows that the sensor produced from the nanostructured sensor has a significantly larger and faster response to the challenge gas of 8 ppm NO₂.

WO3-based sensor response to 8 ppm NOx.

Effect of temperature

For a given sensing material composition, the operating temperature defines the single most effective control of sensor performance. One can readily adjust the response of the sensors via control of the temperature at the sensor surface.

The effect of temperature on the WO₃-based NO_x sensor is clear in this figure. The sensors are in a background of dry air, with 5 ppm NO₂ introduced for two 300-second exposures at times of 300 and 900 seconds. The introduction of challenge gas causes an increase in sensor resistance.

In the case of the WO₃-based NO_x sensor presented here, lower temperatures are preferred for higher sensitivity, particularly at the parts per billion (ppb) level, while higher temperatures can reduce both the sensor resistance in air and the response to gas, allowing a larger dynamic range to happen.

For this sensor, upper detection limits depend on saturation of response and instrumental limitations of measuring very high resistance values. Particularly at the lower temperatures, the measurement of high resistance values is an area of great concern.

Effect of temperature on sensor response to 5 ppm.

Strong voltage response

For applications where lower detection limits are not critical, the simple voltage divider circuit worked very well.

The response of the sensor to a range of NO_x concentrations using the voltage divider circuit is in this figure. In this case, the sensor signal displays as a voltage.

Across this range of test concentrations, the sensors show a strong, reproducible response, with an excellent signal-to-noise ratio.

Sensor response to NO2 using voltage divider circuit.

Typical response data

Here is the response of a typical sensor to 1 ppm of NO and NO₂.

The sensors are in a background of dry air, with challenge gas introduced for 300-second exposures at 300 and 900 seconds. At the operating temperature of ~250°C, the sensor's response to both of these gases is very large, with a shift in resistance of more than an order of magnitude on switching from air to challenge gas.

The sensors typically show a somewhat stronger response to NO₂ than to NO. The sensors have a complete recovery after gas exposure, although both response and recovery time are somewhat slow.

NOx sensor response to 1 ppm challenge.

Higher concentration and faster

Here is the response of another typical sensor to a concentration of 10 ppm NO and NO₂. The operating temperature is approximately 250°C. The sensors are in air with challenge gas applied at 300 and at 900 seconds.

At these higher concentrations, the response of the sensor is faster, while the recovery is significantly slower. Complete recovery takes about ten minutes.

Logging linear relation

This figure shows that the sensor response is very reproducible. Analysts exposed 10 ppm NO₂ to the sensor for 300 seconds over a background of dry air at 300 and 900 seconds, several hours apart, in two trials.

The sensor responses are almost identical in terms of baseline, sensitivity, and response time.

An evaluation of the stability of the sensors over much longer time periods (months to years) is currently ongoing, and preliminary results are very encouraging.

Reproductibility of sensor response 10 ppm NO2. Sensors operated at 300°C.

Process detection limits

The sensors show a strongly increasing resistance with increasing gas concentration. This response is linear when viewed on a log (resistance) versus log (concentration) plot. A typical plot is shown here.

The response to low levels of both NO₂ and NO in this figure range from concentrations of 1–20 ppm. It is evident from this plot that there is a very large sensor response to these relatively low gas concentrations.

An analysis of the response function in conjunction with measured noise levels shows that the sensors are capable of theoretical detection limits of 30 ppb NO₂ and 60 ppb NO, at a signal-to-noise ratio of three, under laboratory conditions.

In practice, the achievable detection limits will obviously be considerably higher due to variations in flow, environmental temperature, and humidity affecting the stability of the sensor.

Actual process detection limits are likely to be approximately 0.5 ppm and depend significantly upon instrumentation and frequency of calibration.

Sensitivity of NOx sensor operated at 250°C.

Selectivity has limits

Shown here is the response of the sensors to some typical challenge gases.

To facilitate comparison, gases that cause an increase in sensor resistance (NO₂, NO) have their response reported as R_G/R_A, while gases that cause a decrease in resistance (H₂, NH₃) have their response reported as –R_A/R_G.

The sensors show a very small response to hydrogen (H₂) and ammonia (NH₃), which is the opposite direction of the response to nitrogen oxides. The sensors show no response to carbon dioxide (CO₂), sulfur dioxide (SO₂), and methane (CH₄). Note that exposure to low levels of sulfur gases does not inhibit the sensors' response to NO_x.

The knock on NOx

NO_x causes a wide variety of health and environmental impacts because of various compounds and derivatives in the family of nitrogen oxides, including nitrogen dioxide, nitric acid, nitrous oxide, nitrates, and nitric oxide.

Ground-level ozone (smog): Smog is formed when NO_x and volatile organic compounds react in the presence of heat and sunlight. Children, people with lung diseases such as asthma, and people who work or exercise outside are susceptible to adverse effects such as damage to lung tissue and reduction in lung function. Ozone travels by wind currents and causes health problems far from original sources. Millions of Americans live in areas that do not meet the health standards for ozone. Other effects of ozone include damaged vegetation and reduced crop yields.

Acid rain: NO_x and sulfur dioxide react with other substances in the air to form acids that fall to earth as rain, fog, snow, or dry particles. Wind can carry them for hundreds of miles. Acid rain deteriorates cars, buildings, and historical monuments, and makes lakes and streams acidic and unsuitable for many fish.

Particles: NO_x reacts with ammonia, moisture, and other compounds to form nitric acid and related particles. Human health concerns include effects on breathing and the respiratory system, damage to lung tissue, and premature death. Small particles penetrate deeply into sensitive parts of the lungs and can cause or worsen respiratory diseases such as emphysema and bronchitis and aggravate existing heart disease.

Water quality deterioration: Increased nitrogen loading in water bodies, particularly coastal estuaries, upsets the chemical balance of nutrients used by aquatic plants and animals. Additional nitrogen accelerates "eutrophication," which leads to oxygen depletion and reduces fish and shellfish populations. NO_x emissions in the air are one of the largest sources of nitrogen pollution in the Chesapeake Bay.

Global warming: One member of the NO_x family, nitrous oxide, is a greenhouse gas. It accumulates in the atmosphere with other greenhouse gases, causing a gradual rise in the Earth's temperature. This will lead to increased risks to human health, a rise in the sea level, and other adverse changes to plant and animal habitat.

Toxic chemicals: In the air, NO_x reacts readily with common organic chemicals and even ozone, to form a wide variety of toxic products, some of which may cause biological mutations. Examples of these chemicals include the nitrate radical, nitroarenes, and nitrosamines.

Visibility impairment: Nitrate particles and nitrogen dioxide can block the transmission of light, reducing visibility in urban areas and on a regional scale in national parks.

Behind the byline

Debra Deininger, Stephen Williams, and Clayton Kostelecky are scientists and engineers with Synkera Technologies (www.synkera.com). The authors wish to recognize the support of the National Science Foundation, NASA, and the Department of Energy–Office of Information Technology in the development of this sensor technology.