Saturation of the IMPROVE denuder for SO2 and HNO3
Author: Robert Eldred, CNL, University of California, Davis (raeldred@ucdavis.edu)
Date: May 7, 1998
Summary: The IMPROVE denuder has a collection efficiency for SO2 of >99% when freshly coated. There is some question about whether there is breakthrough of SO2 and HNO3 after being used for a year, at least at sites with high SO2 concentrations. This report will examine the effect of breakthrough of SO2 on the sulfate measurements and of HNO3 on the nitrate measurements. Our conclusion for SO2 is that the nylon filter does not collect SO2; no effect on the sulfate measurement on the nylon filter is expected or observed. Our conclusion for HNO3 is that is that no change in measured nitrate can be observed between old and new denuders; since the nylon filter does collect HNO3, the conclusion is that there is no breakthrough of HNO3. We are examining the option of revising the frequency of changing the denuders.
Denuder efficiency: The IMPROVE denuder consist of four concentric annular denuder sections coated with Na2CO3 plus 3% glycerin. The method of M. Possanzini,1 estimates the efficiency under the assumption that a particle is captured once it touches a wall. For the IMPROVE denuder the fraction of HNO3 or SO2 passing through without touching is < 10-9. Thus a typical gas molecule will touch the surface several times. Tests at Davis give a collection efficiency for SO2 of >99%. This drops to 90% with low RH after about 800 µg of SO2 has entered the denuder. In our current protocols, the denuder is replaced annually. Since about 3300 m3 of air passes through the denuder during a year, a 90% efficiency would be reached with an average SO2 concentration of 250 ng/m3. SO2 has been measured at many IMPROVE sites using a carbonate-impregnated filter. Table 1 shows that SO2 exceeds 250 ng/m3 at most sites.
Table 1. Average annual SO2 concentrations at IMPROVE sites. Nitric acid is not measured, but is probably much less, except at Sequoia. The units are ng/m3.
Appalachia
Shenandoah, WV 4453
Mammoth Cave, KY 3721
Great Smoky Mtns, TN 2613
Northeast
Acadia, ME 1550
Southeast
Okefenokee, GA 584
Chassahowitzka, FL 534
Everglades, FL 130
Voyageurs,
MN 388
Southwest
Saguaro, AZ 1523
Chiricahua, AZ 1226
Mesa Verde, CO 571
Canyonlands, UT 492
Big Bend, TX 491
Indian Gardens, AZ 427
Grand Canyon, AZ 370
California
Sequoia, CA 460
Death Valley, CA 246
Yosemite, CA 234
Pinnacles, CA 223
Redwood,
CA 27
Northwest
Mount Rainier, WA 183
Great Basin, NV 145
Rocky Mountains
Rocky Mountain, CO 186
Yellowstone, WY 102
Denali, AK 203
However, based on Stevens’ capacity measurements for the triple stainless CLAD glass denuder,2 the capacity of the IMPROVE denuder may be greater than 800 µg. The surface of the CLAD denuder is etched glass. The capacities of the CLAD denuders are 1700 µg at low RH and 5000 µg at 85% RH. The surface area of the IMPROVE denuder is 7-12 times that of the CLAD denuders. Even if the aluminum surface less carbonate than etched glass, a factor of 20 in µg/surface area seems too large.
The conclusion is that the IMPROVE denuder has a very high efficiency for SO2. There is some question about whether the efficiency for SO2 collection remains high after a year of operation at sites with high SO2 concentrations. The efficiency for collecting HNO3 after a year has not been measured for the IMPROVE or similar denuders. Because HNO3 is much more reactive than SO2, it may remain high even when that for SO2 decreases.
Sulfate: There are two factors here: (1) the amount of SO2 that passes through the denuder, and (2) the collection of SO2 by the nylon filter. Both must occur for there to be any effect.
Figure 1 compares sulfate measurements on nylon filters from two collocated samplers, one with a freshly coated denuder and one with no denuder. These samples were collected at Dolly Sods Wilderness, WV, during summer 1997. The mean concentrations of the two configurations differ by 0.3%. At a nearby site, Shenandoah National Park, the average SO2 molar concentrations are about one-half of the SO4 molar concentrations. The conclusion is that there is no significant collection of SO2 by the nylon filter.
Figure 1. Comparison between two measurements of sulfate on nylon, one with a freshly coated denuder and one with no denuder. Samples from Dolly Sods Wilderness, WV during summer 1997. The units are µg/m3. The correlation coefficient (r2) is 0.98 and the slope is 1.003 ± 0.025.
We can verify that no effect is observed by comparing the sulfate is measured on nylon with the sulfur measured on Teflon for corresponding samples. Figure 2 compares both sulfate measurements and SO2 at Great Smoky Mountains during fall 1997, when the denuder had been in use for a complete year. The average SO2 concentration was over 2 µg/m3. The left side compares the two sulfate measurements, while right side compares the difference with SO2. These plots show that there is no increase in the sulfate measurement on the nylon filter, even after the denuder is in use for a year.
Figure 2. Comparison between sulfate measured on nylon, sulfur (times 3) measured on Teflon, and SO2 measured with an impregnated filter. Samples from Great Smoky Mountains from fall 1997; the denuder had been in use for one year. The left side compares the two particulate measurements, while the right side compares the difference with SO2. All units are nmol/m3.
Nitrate: Figure 3 shows that HNO3 is collected on the nylon filter, increasing the nitrate measurement. This compares nitrate collected on a nylon filter following a freshly coated denuder with that collected with no denuder. The ratio of means is 2.1. Thus if the denuder had zero efficiency for HNO3, the measured nitrate concentrations would increase significantly.
Figure 3. Comparison of nitrate following a freshly coated denuder and nitrate with no denuder at Dolly Sods Wilderness. The difference is the nitric acid that is not captured in the inlet, stack, and cyclone. The units are µg/m3.
We can examine the measured nitrate concentration both before and after replacing the denuder with a freshly coated denuder. If any significant fraction of the HNO3 were to penetrate the denuder with the old denuder, then we would expect the measured nitrate before the change would be higher than that after the change. Figures 4 through 19 plot the nitrate measurements on both sides of the annual sampler maintenance for a wide variety of sites. The first sample with a new denuder is on the center line.
Figures 4 to 7 are at eastern sites with high SO2 concentrations, while 8 to 10 are at eastern sites with moderate SO2. Figures 11 and 12 are at western sites with very high nitrate and moderate SO2; presumably the HNO3 concentrations are large. Figures 13 to 19 are at western sites with comparatively low SO2 concentrations.
There is no evidence of any decrease following the changing of denuders. The conclusion is that the collection efficiency for HNO3 is always high, even at sites with high SO2. Thus the protocol of changing the denuders annually does not have any adverse affect on the nitrate measurements.
Figure 4. Comparison of measured nitrate before and after replacing the coated denuder at Dolly Sods Wilderness. The first sample with a new denuder is on the center line. The average annual SO2 concentration is probably similar to that at Shenandoah, around 5 µg/m3.
Figure 5. Comparison of measured nitrate before and after replacing the coated denuder at Shenandoah National Park. The first sample with a new denuder is on the center line. The average annual SO2 concentration is around 5 µg/m3.
Figure 6. Comparison of measured nitrate before and after replacing the coated denuder at Great Smoky Mountains National Park. The first sample with a new denuder is on the center line. The average annual SO2 concentration is around 3 µg/m3.
Figure 7. Comparison of measured nitrate before and after replacing the coated denuder at Washington DC. The first sample with a new denuder is on the center line.
Figure 8. Comparison of measured nitrate before and after replacing the coated denuder at Acadia National Park, ME. The first sample with a new denuder is on the center line. The average annual SO2 concentration is around 2 µg/m3.
Figure 9. Comparison of measured nitrate before and after replacing the coated denuder at Chassahowitzka NWR, FL. The first sample with a new denuder is on the center line. The average annual SO2 concentration is around 0.5 µg/m3, which is much lower than at the Appalachian sites.
Figure 10. Comparison of measured nitrate before and after replacing the coated denuder at Boundary Waters Canoe Area Wilderness, MN. The first sample with a new denuder is on the center line.
Figure 11. Comparison of measured nitrate before and after replacing the coated denuder at San Gorgonio Wilderness, CA, which has the highest nitrate concentration in the IMPROVE network. The first sample with a new denuder is on the center line.
Figure 12. Comparison of measured nitrate before and after replacing the coated denuder at Sequoia National Park, which has the second highest nitrate concentration in the IMPROVE network. The first sample with a new denuder is on the center line. The average annual SO2 concentration is around 0.5 µg/m3.
Figure 13. Comparison of measured nitrate before and after replacing the coated denuder at Yosemite National Park. The first sample with a new denuder is on the center line.
Figure 14. Comparison of measured nitrate before and after replacing the coated denuder at Three Sisters Wilderness, OR. The first sample with a new denuder is on the center line.
Figure 15. Comparison of measured nitrate before and after replacing the coated denuder at Glacier National Park. The first sample with a new denuder is on the center line.
Figure 16. Comparison of measured nitrate before and after replacing the coated denuder at Great Basin National Park, NV. The first sample with a new denuder is on the center line.
Figure 17. Comparison of measured nitrate before and after replacing the coated denuder at Canyonlands National Park. The first sample with a new denuder is on the center line.
Figure 18. Comparison of measured nitrate before and after replacing the coated denuder at Grand Canyon National Park. The first sample with a new denuder is on the center line. The average annual SO2 concentration is around 0.4 µg/m3.
Figure 19. Comparison of measured nitrate before and after replacing the coated denuder at Weminuche Wilderness, CO. The first sample with a new denuder is on the center line.
Conclusions: The efficiency of the current denuder for SO2 is >99% for a freshly coated denuder. The efficiency for SO2 collection may drop off at sites with high SO2 with the current protocol of annual change. However, since the nylon filter does not collect SO2, no change in measured sulfate is observed. Any HNO3 passing through the denuder would be collected, increasing the measured nitrate. Since no change is observed between used and fresh denuders, the conclusion is that the efficiency for HNO3 remains high even with the current protocols.
Future plans:
1. We will directly measure the capacity of the denuder for SO2 and HNO3.
2. We will examine whether the capacity can be increased by revising the protocols for coating the denuders and by etching the aluminum surface.
3. We will examine the lifetime of the denuder by operating two collocated samplers, one with the denuder replaced frequently and one with no replacement and measure nitrate in both.
4. We will consider the option of increasing the frequency of replacement of the denuders, based on these measurements.
References
1. Possanzini, M., A. Febo, and A. Biberti, “New design of a high-performance denuder for the sampling of atmospheric pollutants,” Atmospheric Environment, 1983, 17, 2605-2610.
2. Stevens, R.K., L.J. Purdue, H.M. Barnes, R.P Ward, J.O. Baugh, J.P. Bell, H. Sauren, J.E. Sickles, L.L. Dodson, “Annular denuders and visibility studies,” in Visibility and Fine Particles, edited by C.V. Mathai, Air & Waste Management Assn., Pittsburgh, 1990, pp. 122-130.