Grantee Research Project Results
Final Report: An Investigation of the Gas Sensing Properties of a Novel Manganese-Oxide-Supported Gold Catalyst
EPA Grant Number: R823130Title: An Investigation of the Gas Sensing Properties of a Novel Manganese-Oxide-Supported Gold Catalyst
Investigators: Gardner, Steven D.
Institution: Mississippi State University
EPA Project Officer: Hahn, Intaek
Project Period: October 1, 1995 through September 30, 1997 (Extended to September 30, 1998)
Project Amount: $98,580
RFA: Exploratory Research - Chemistry and Physics of Air (1995) RFA Text | Recipients Lists
Research Category: Air , Safer Chemicals
Objective:
The need to detect trace amounts of potentially hazardous gases has become increasingly important due to concerns over environmental pollution, occupational safety and combustion efficiency. In order to address these concerns, numerous commercial instruments have been made available that detect a variety of gases including CO, H2, O2, CO2, NOx, alcohols and hydrocarbons. Many of these instruments are based upon solid-state gas sensors composed of semiconducting metal oxides. In such cases, the detection schemes rely on surface phenomena that occur upon exposure of the sensor to the targeted gas. For example, the electrical conductivity of a semiconducting metal oxide such as SnO2 can be altered by adsorption of gases from the ambient. It follows that a calibrated response may be established between the SnO2 surface conductivity and the concentration of the target gas in the ambient. The net result is a rather simple and inexpensive gas sensing device.Despite the recent advances in semiconductor gas sensing instrumentation, the underlying technology itself must evolve significantly in order to address certain performance deficiencies. Perhaps the most critical problem associated with current semiconducting gas sensing devices is a lack of selectivity toward a specific gas. In addition, there is often the problem of gradual sensor deactivation. Considerable response times can also hinder the performance and overall applicability of gas sensors. In short, there is a critical need to enhance the selectivity, lifetime stability, and response time of semiconductor gas sensors while maintaining acceptable sensitivity toward the gas(es) to be monitored.
In previous research, Au/MnOx has exhibited unparalleled performance as a low-temperature CO oxidation catalyst. In particular, Au/MnOx has consistently proven to be superior to CO oxidation catalysts based upon SnO2, the latter of which is one of the most widely studied and successful semiconducting gas sensor materials for CO detection. Given the excellent catalytic properties of Au/MnOx and the fact that catalytic properties are ultimately related to gas sensing characteristics, it is highly probable that this material will also find applications in high-performance gas sensing instrumentation designed to detect reducing gases such as CO.
Much of the current development in gas sensing technology has been the result of empirical testing and evaluation. However, in order to address the performance issues cited above, basic research is needed to better understand the fundamental surface processes that are ultimately responsible for the gas sensing mechanism(s) in place. In the present research, experiments have been performed to investigate the gas sensing properties (based on electrical conductivity measurements) of a novel manganese-oxide-supported gold catalyst (Au/MnOx) and furthermore to explore the surface phenomena connected with the conductivity changes. The surface conductivity of Au/MnOx has been evaluated during stepwise, sequential exposures to CO, H2, NO, CH4 and dry air and the resulting surface composition has been determined using x-ray photoelectron spectroscopy (XPS) and ion scattering spectroscopy (ISS). Similar data have also been acquired for MnOx and Pt/SnO2 and the information is used to gain more insight into the gas sensing behavior observed.
Summary/Accomplishments (Outputs/Outcomes):
Of the gases tested, the Au/MnOx catalyst exhibits the greatest overall sensitivity to CO. This is consistent with expectation based on the CO oxidation performance of this material as discussed above. Specimens containing 2 atomic percent gold (2 at% Au/MnOx) and 10 atomic percent gold (10 at% Au/MnOx) were examined in this case. Upon exposure to 0.5 Torr of CO at 130 oC, 2 at% Au/MnOx exhibits more than a three-fold increase in surface conductivity; whereas a two-fold increase is observed for 10 at% Au/MnOx. When the sample temperature is either decreased to 100 oC or increased to 250 oC, the CO sensitivity of both materials decreases. For MnOx samples that did not contain any gold, the surface conductivity at 130 oC is enhanced by a factor of two upon exposure to 0.5 Torr of CO. This demonstrates that conditions exist for which the gold and manganese oxide may interact synergistically toward CO detection.Perhaps the most intriguing result of the surface conductivity experiments is the behavior exhibited by Au/MnOx (and MnOx) upon exposure to dry air. Measurement of the Seebeck voltage indicates that both materials are n-type semiconductors, and therefore exposure to an oxidizing atmosphere would be expected to decrease the surface conductivity. Nevertheless, when these MnOx-based materials are exposed to dry air, there is an increase in the surface conductivity on the order of that observed for CO exposure. The same trends are observed when the dry air is replaced with oxygen. A possible explanation for this behavior is based in part on similar cases documented in the literature. Specifically, it has been observed that n-type semiconductors can behave as p-type semiconductors in the presence of sufficiently high partial pressures of oxygen. In such cases, it has been proposed that this condition might arise from a decreased concentration of electron charge carriers due to the formation of oxygen ions on the sensor surface. At a sufficiently high oxygen ion surface concentration, the decreased contribution of electrons in the conduction band may become more than offset by the number of holes created in the valence band. As a result, the sensor would begin to exhibit p-type behavior.
Experiments were performed to determine if the Au/MnOx and MnOx surface conductivities eventually show negative deviation upon exposure to lower partial pressures of oxygen. As the oxygen partial pressure is reduced, the effect is only to attenuate the conductivity increases, ultimately to a point where no response can be discerned. Therefore, it appears that the region of n-type response to oxygen occurs at an oxygen partial pressure too low to effectively measure a conductivity change for the conditions of this investigation.
Under the same conditions cited above (130 oC, 0.5 Torr CO), the 2 at% Au/MnOx conductivity response compares favorably to the four-fold increase in conductivity that was measured for a 2 at% Pt/SnO2 catalyst. Unlike Au/MnOx, however, the conductivity response of Pt/SnO2 is not reversible upon subsequent vacuum exposure. That is, upon CO removal, the conductivity of Pt/SnO2 does not return to the baseline value that is established before CO exposure. The reversible response observed for Au/MnOx may bode well for long-term stability as a CO sensor. Furthermore, the Pt/SnO2 response to dry air (or oxygen) is a decreased surface conductivity, in agreement with expectation for this n-type semiconductor and similar observations in the literature.
The surfaces of 2 at% Au/MnOx, MnOx and 2 at% Pt/SnO2 were all examined by XPS and ISS as a function of CO gas exposure conditions. In this case, the air-exposed specimens were characterized initially in order to provide a basis for comparison. Subsequently, each specimen was exposed to 0.5 Torr of CO at 130 oC for sufficient time to enable the conductivity to stabilize (typically less than 5 minutes.) The chamber was then evacuated and surface analysis experiments were performed immediately thereafter. The entire procedure was repeated with 0.5 Torr of dry air at 130 oC. Finally, the CO and dry air exposures were performed a second time yielding four successive gas exposures with intermittent surface analysis. The results from these experiments are summarized below.
The Mn 2p XPS spectra of air-exposed 2 at% Au/MnOx are consistent with the presence of a mixture of MnO, Mn3O4, Mn2O3 and/or MnO2, and this remains the case throughout the gas exposure sequence. During the gas exposures, the Mn/O atomic ratio increases from 0.7 to about 0.9 which may indicate progressive reduction of the manganese oxide support. More information about the chemical state of manganese would ordinarily be available from the extent of Mn 3s multiplet splitting, but unfortunately, the Mn 3s features overlap the Au 4f peaks.
The main features in the corresponding O 1s spectra are due to MnOx, hydroxyl groups and adsorbed water. The overall trend is a decrease in the relative fraction of oxygen present as hydroxyl groups as the gas exposures are carried out. Nevertheless, a significant amount of hydroxyls and/or adsorbed water remain on these surfaces. This is important because it has been proposed that these species are important to the low-temperature CO oxidation activity of Au/MnOx. Therefore, these results may implicate their importance to the CO gas sensing mechanism as well.
As mentioned above, the binding energies of the Au 4f peaks are near those of the Mn 3s features and this hinders the spectral interpretation. In the case of 2 at% Au/MnOx, the intensity of the Au 4f peaks is not strong enough (relative to the Mn 3s signal) to provide a clear indication of the gold's chemical state. The Au 4f7/2 binding energies are approximately 84.2 eV which is slightly above that of metallic gold (84.0 eV) and well below the binding energy of Au2O3 (86 eV). Based on similar investigations from the literature, it is speculated that the gold is probably present as small metallic clusters and the small positive binding energy shift may be due to enhanced chemical interaction with the support.
The ISS data are consistent with those of XPS even though ISS is more surface sensitive (essentially outermost atomic layer only). On air-exposed Au/MnOx surfaces, ISS detects carbon, sodium, and chlorine, in addition to gold, manganese and oxygen. After the initial CO exposure, the relative concentrations of carbon, sodium, oxygen and chlorine are diminished. After subsequent exposure to dry air, there is an increase in the oxygen presence which would be consistent with replenishment of oxygen lost during prior CO reduction of the surface. The oxygen concentration follows a similar trend (decrease and increase, respectively) during the second CO/dry air cycle, although to a lesser extent. However, after the first CO exposure and beyond, no carbon is detected by ISS. This is an important observation and several factors may be involved. It is possible that CO adsorbs on the Au/MnOx surface such that the carbon atom is physically shielded from the ISS probe ions by oxygen or hydrogen. (ISS is unable to detect hydrogen.) Shielding of the carbon atom would also be likely if a unidentate or bidentate carbonate species is present. Considering the fact that the surface conductivity of Au/MnOx returns to the baseline value after CO evacuation (i.e., the conductivity change is reversible as mentioned above), the ISS spectra are probably indicative of CO desorption from the surface as a result of vacuum exposure.
The fact that XPS and ISS of Au/MnOx reveals the presence of sodium and chlorine raises interesting questions about the potential influence of these elements on the surface conductivity. It is difficult to prepare Au/MnOx specimens that are chlorine-free using chloroauric acid as a presursor in the co-precipitation method of the present investigation. However, washing the precipitates with hot water significantly reduces the presence of surface chlorine. Chlorine is known to be detrimental to the performance of many CO oxidation catalysts, and this may well be true with CO gas sensing as well. The influence of sodium on surface conductivity could be examined by using a different precipitating agent other than sodium carbonate, such as potassium carbonate or lithium carbonate, although this has not been considered in the present study.
Under the conditions of the investigation, the Pt/SnO2 specimen exhibits the largest conductivity change (increase) upon exposure to CO. As noted above, unlike MnOx and Au/MnOx, however, the response to CO is not reversible. This may be due to residual CO chemisorbed on the surface, but XPS and ISS are not able to yield conclusive evidence. Subsequent exposure of Pt/SnO2 to dry air decreases the measured conductivity. Small amounts of reduced Sn are detected on the Pt/SnO2 specimen (possibly alloyed with Pt), the presence of which decreases upon progressing through two complete cycles of CO and dry air exposure. However, since the conductivity change of Pt/SnO2 is even greater during the second CO exposure, the presence of reduced Sn may adversely affect the sensor response. Hydroxyl groups are also detected on Pt/SnO2 and their surface concentration decreases and increases, respectively, upon CO and dry air exposure. This may be an indication of their importance to the Pt/SnO2 gas sensing mechanism. In all cases the relative fraction of oxygen on Pt/SnO2 present as hydroxyl groups/adsorbed water is less than that measured on Au/MnOx and this is in agreement with the previous statement and the superior CO oxidation activity of Au/MnOx. The fact that the CO conductivity change is not as great for Au/MnOx is consistent with different CO gas sensing mechanisms with respect to utilization of surface hydroxyl functions. This assertion is further supported by XPS which indicates a significant fraction of Pt is present as Pt(OH)2.
One the major goals of the investigation was to correlate sensor response to surface composition. As mentioned above, sensors fabricated with 2 at% Au/MnOx did not contain enough gold to yield acceptable signals during XPS and ISS surface analysis. This rendered an assessment of the gold's chemical state difficult at best. For this reason, the remainder of the investigation focused on evaluating the response characteristics of 10 at% Au/MnOx. Furthermore, since Au/MnOx exhibited less than a 10% conductivity change upon methane exposure and nitric oxide exposure (100 oC to 250 oC, 0.5 Torr), additional data were acquired only for hydrogen exposures. These results are summarized below.
There is approximately a two-fold increase in the surface conductivity of 10 at% Au/MnOx upon hydrogen exposure (0.5 Torr) and the response does not vary significantly over a temperature range of 100 to 250 oC. In addition, the conductivity changes are essentially reversible as explained above for CO. Subsequent dosing with dry air results in conductivity increases that are roughly 70% of those measured during prior hydrogen exposure. Therefore, the prior exposure to hydrogen does not change the sensor's p-type response to oxygen as described above. The trends in the corresponding XPS and ISS spectra are similar to those identified above for the CO/dry air exposures. Multiple chemical states of manganese are still evident, but the XPS Mn/O atomic ratio maintains a value near 0.8. This would be consistent with less overall surface reduction relative to CO exposure for these conditions. Although the Au XPS signals are more clearly defined in this case, the Au 4f7/2 binding energies appear near 84.2 eV as above. Therefore, the enhanced Au spectra are unable to provide additional clarification of the gold's chemical state.
Conclusions:
The surface conductivity of Au/MnOx exhibits sufficient sensitivity to carbon monoxide and hydrogen to be considered further as a potential sensor for these gases. The increase in surface conductivity can be as high as three-fold for CO and H2 partial pressures of 0.5 Torr within the temperature range of 100 to 250 oC. On the contrary, the conductivity changes are less than 10% for nitric oxide and methane at these same conditions. The conductivity responses are essentially reversible in all cases which may bode well for long-term operational stability. Over the range of variables considered, a given Au/MnOx specimen exhibits comparable sensitivity to CO and H2, and therefore the material's response is practically nonselective in this case. Typical CO and H2 detection limits are near 0.1 Torr of partial pressure which equates to about 150 ppm and 10 ppm, respectively, in atmospheric air.Relative to MnOx alone, the Au/MnOx specimens exhibit an enhanced conductivity response to CO, H2 and dry air suggesting a synergistic interaction between Au and the MnOx support. In all cases, exposure dry air (or oxygen) results in a conductivity increase which is atypical for these and other n-type semiconductors. The conductivity changes are essentially reversible in each case. The surface analysis data suggest that the chemical state of Mn on MnOx and Au/MnOx is similar, indicating a mixture primarily of Mn3O4 and Mn2O3, although MnO is also detected. On the contrary, significant differences are noted with respect to the presence of adsorbed water and hydroxyl groups, the concentration of which is greater on Au/MnOx. Since the presence of hydroxyl groups has often been shown to promote CO oxidation performance on similar materials, this correlates well with the superior CO oxidation activity of Au/MnOx (relative to MnOx) and it is consistent with the enhanced conductivity response of Au/MnOx as well. The XPS data provide a basis for speculating that gold is present in the metallic state on Au/MnOx, and there may be a chemical interaction with the support.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
Other project views: | All 3 publications | 1 publications in selected types | All 1 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Srinivasan B, Gardner SD. Investigation of the gas sensing properties of Au/MnOx: Response to CO exposure and comparison to Pt/SnO2. Surface and Interface Analysis 1998;26(13):1035-1049. |
R823130 (Final) |
not available |
Supplemental Keywords:
air, ambient air, atmosphere, exposure, carbon monoxide, CO, hydrogen, environmental chemistry, measurement methods, Mississippi, MS, Region 4., Scientific Discipline, Air, Geographic Area, Physics, Chemistry, State, Engineering, Chemistry, & Physics, EPA Region, metal catalysts, auger electron spectroscopy, region 4, semiconducting metal oxides, chemical composition, Mississippi (MS), analytical chemistry, gas sensing system, intermittent surface analysis, manganese-oxide supported gold catalyst, ion scattering spectroscopy, ultra-high vacuum temperaturesProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.