Project Title: Development and Application of a Rapid and Robust Sensor to Determine Nitrogen Species in the Coastal Atmosphere
Principal Investigator(s): Joel E. Baker, Ronald L. Siefert, Amy K. Zander

Accomplishments
Scheduled Tasks:
i) Prepare and submit a manuscript describing the instrument.

Progress on Tasks
A manuscript has been prepared describing this instrument and will be submitted to Analytical Chemistry. The manuscript describes the instrument developed for this project and describes both laboratory testing and field inter-comparisons. This manuscript is also a chapter in Randy Larsen’s Ph.D. thesis. Dr. Larsen recently defended his thesis this past summer and is currently preparing the manuscript for submission to Analytical Chemistry. Attached is a copy of this chapter describing the instrument

Difficulties Encountered
As stated in the previous progress report, the proposed micro-porous hollow fiber membrane module (MHFMM) system did not have a known constant removal rate of gas-phase ammonia and gas-phase nitric acid within a continuous atmospheric sampling environment. The data that has been gathered through mass transfer experiments with the custom MHFMM did not result in a definitive range of mass transfer coefficients. However we have been able to circumvent this problem by modifying the design of the instrument. The modification includes incorporating two mist chambers with two separate inlets. One of the inlets uses a coated denuder to scrub out the gas-phase species while the other inlet uses a uncoated denuder that let's gas-phase species pass through. Therefore one channel measures total gas-phase and particulate phase species, while the other channel measures only particulate phase species. By difference the concentration of gas-phase species can be calculated. This instrument has been tested in several field projects associated with this project and also with a project investigating ammonia/um chemistry and emissions in the Chesapeake Bay airshed.

Anticipated Success in Meeting Project Objectives in Scheduled Project Period
We have met the overall objective of our proposal although we have modified the design due to difficulties that we encountered. We have built a system that can simultaneously measure gas-phase and particulate-phase atmospheric ammonia/um concentrations. The system can measure these two channels in near real-time with a temporal resolution of less than 10 minutes. Field experience with the instrument has shown that an ammonia/um preliminary concentration can be calculated within about minutes. The data can then be further refined to provide more precise and accurate values by incorporating instrument drift into the calculation of ammonia/um concentrations.

Preliminary Results
The following is an abbreviated portion of the attached thesis chapter describing intercomparison studies involving the continuous mist chamber / fluoromoter instrument and standard air sampling methods for ammonia/um. See the attached thesis chapter for more details.

The continuous mist chamber/fluorometer instrument was deployed along side other ammonia/ammonium samplers in July and December 2001. The summer deployment occurred at the Fort Meade, MD in conjunction with an ambient particulate matter sampling program (Chen et al., in review). The instrument was deployed from July 15 20, 2001 on a scaffold approximately five feet off the ground in a field on the Fort Meade army post. Comparing the results against independent measurements of ammonia and ammonium demonstrate the validity of this instrument as a tool to measure atmospheric ammonia and ammonium. Another group of investigators collected ammonia and ammonium using a sequential gas sampler (SGS) at the same time and location (Chen et al., in review). Plotting the (SGS) values against the 24 hour averages from the mist chamber results in a good correlation (R2 = 0.85). In general the mist chamber/fluorometer tends to slightly underestimate the concentrations relative to the SGS. This may be due to the averaging process that includes spurious signals that occurred due to bubbles in the fluorometer which were not sufficiently filtered. The correlation would be improved to an R2 = 0.95 with the removal of the one large outlier. This outlier occurred on July 20th when ammonium concentrations were on the decline, but the calculated daily average did not include this decrease because the instrument was shut down early.

Total ammonia (ammonia gas plus ammonium aerosol) ranged from 0.5 to 7 µg N m^-3. Clear diurnal patterns emerged with concentrations varying over 2 µg N m^-3 within 24 hours. These details would not be possible with traditional filter packs or denuders. In general, total ammonia concentrations were on the rise from the 15th to the 18th. At which point a storm passed through the area, resulting in a drop in total ammonia concentrations from the 18th to 20th. Particulate phase ammonium was the predominant form during this period. Missing data occurred on the morning of July 19th due to difficulties related to storm.

Another deployment occurred in December 2001 at the Beltsville, MD, USDA Research Center, in a field adjacent to a dairy farm. Comparison with total ammonia collected with a honeycomb denuder/filter pack assembly was poor, but there are several potential explanations. One reason was differences in inlet configurations, the mist chamber/fluorometer had 10 feet of 0.5 inch heated polyethylene tubing to draw air from outside into a trailer were the equipment was located. The denuder/filter pack was mounted outside within three feet of the mist chamber inlet, but was not heated. This may have caused differences in the aerosol cut sizes. Additionally, heavy fog and drizzle occurred during the early days of the study. Biased high levels of ammonia would be measured if ammonia contaminated water condensed onto the denuder/filter pack surfaces. This would explain the under estimates of the mist chamber during some of the days. Overall, attention needs to be made when designing inlets.

Extremely high ammonia concentrations occurred when winds were from milking station. This provided an opportunity to evaluate the mist chamber/fluorometer under high ammonia conditions. When winds were blowing from the milking station, total ammonia (ammonium plus ammonia) rose into the tens of µg N m^-3, compared to typical ambient concentrations of 1 to 2 µg N m^-3. Both this deployment and the Fort Meade deployment clearly indicate that atmospheric ammonium concentrations can double within 12 hours which further supports the need for an instrument, such as this, to provide sub-hourly measurements.

Overall the instrument presents a valuable way of improving our understanding of atmospheric processes by providing significantly greater detail into the fluctuations of atmospheric ammonia concentrations. The instrument is both rugged and deployable having been operated outside for over a week with minimal operator intervention, withstood continuous vibrations and rocking motion of a boat, and can operate off of a portable generator. This instrument has the ability to provide ten minute resolution measurements of ammonia and ammonium which is much faster than traditional integrated sampling methods (12 to 24 hours), about equal to wetted parallel plate denuders and slower than DOAS or laser (< 1 second) based methods. However the latter three methods do not analyze particulate phase ammonium which was the primary form of ambient atmospheric ammonia/ammonium. The LOD was 0.05 mg N m^-3, while the precision base on three times the signal to noise was 0.5 mg N m^-3. The accuracy was 0.5 mg N m^-3 based on a comparison of independent field measurements during the Fort Meade deployment.

Tasks and activities for next reporting period

Tasks for the next reporting period
Submit the manuscript describing the instrument to Analytical Chemistry.

Concerns or difficulties
None.