Development of a Real-Time In Situ Organic/Elemental Carbon (OC/EC) Analyzer

EPA Grant Number: 68D99070
Title: Development of a Real-Time In Situ Organic/Elemental Carbon (OC/EC) Analyzer
Investigator: Robert A. Cary
Small Business: Sunset Laboratory, Inc.
EPA Project Officer: Dr. Jim Gallup/Kathryn A. Peele
Project Period: September 1999 - March 2000
Project Amount: $69,644
Research Category: Monitoring/Analytical

Project Summary:

Project Description:
The goal of the project was to develop a new, commercially viable, real-time in situ OC/EC instrument under this SBIR. Initial design and performance validation was be based on a previously described prototype. Phase I was designed to optimize the instrument sensitivity and reliability through a series of design improvements resulting in a field deployable working prototype. It incorporated the latest in electronic and optical developments along with an optimized oven and detector design for improved, reliable field operation. The resulting device when operated as a tandem instrument was projected to have a minimum time resolution of 30 minutes with a minimum quantifiable level of 0.5 µg/M3 each of organic and elemental carbon.

Summary of Findings:
The goal of the project was to develop the instrument around specialized state-of-the-art components leading a device which could be subsequently packaged for field deployment. Components had to be optimized for sufficient stability and sensitivity to provide adequate time resolution at ambient target analyte levels. The primary efforts were directed towards optimizing detector sensitivity and stability, development of custom electronics for both system control and sensitive data acquisition, development of the optical measurement system and development of associated sampling components to minimize artifact formation during sample collection and analysis. It was found that the detector performance could be made to be more than adequate with only modest hardware modifications and careful control of the system gases. Once optimization of these aspects of the system were achieved, detector performance was found to be well within the projected specification limits. Electronic control of the instrument focused on development of a custom data acquisition and control system which can control all of the external events required for operation of the system, collect data from the analog outputs and communicate with a host computer. This system uses a custom data system optimized to operate with an embedded controller. The data system has been determined to be both flexible and highly stable allowing accurate measurement of low level analog signals generated by the instrument. The embedded controller is a stable platform with a range of capabilities required for a field deployable instrument. It was known from the initiation of the project that the laser optical system was potentially the most difficult part of the system to develop within the desired performance goals. Initially, research focused on laser systems similar to those used in our laboratory system. Eventually, however, a custom system was obtained with sufficient power and stability to operate at the desired performance level. At the close of this program, the laser system in use operates with sufficient throughput and stability to detect readily changes in transmittance of less than 0.5%. One of the primary difficulties with quartz filter samplers for measuring organic (OC) and elemental (EC) carbon is the tendency to develop a positive artifact from the adsorption of organic gases on the quartz filter substrate resulting in artificially high OC levels. In order to avoid this with this system, a set of custom stainless steel organic denuders were developed. These devices eliminate the breakage problem associated with other devices currently on the market. Using a proprietary coating system with a much higher loading capacity, sample results suggest that these denuders completely remove the positive artifact problem associated with these systems. These high capacity devices used in conjunction with the real-time analyzer can be operated up to 48 hours under ambient urban conditions without breakthrough. Performance of the system initially had to focus on the comparability of the sample results to those obtained from composite samplers currently being used by monitoring groups. A sampling system was set up with a PM 2.5 cyclone inlet at 16.7 lpm. Flow from this inlet was split in two with one side directed to the real-time sampler and the other to a filter pack containing quartz filters. The time resolved data from the real-time instrument were summed for the time period and compared to the composite quartz filters collected during the same period. For the limited number of data sets collected, the OC recoveries for the real-time instrument ranged from 88% to 107% when compared to the composite filter while the EC ranged from 76% to 128%. Although the recovery range was larger for the EC, on average, the EC levels were very low (< 0.8 µg/M3) and all of the results statistically overlap. Also, with such low levels, a small error in split time on either measurement could easily result in this much variation. In fact, data from co-located composite samplers rarely result in any better agreement. Sometimes, unfortunately, even two punches from the same filter will not agree any better when levels are this low. Therefore, with the limited results available, it appears that the real-time instrument results are readily comparable with existing accepted methods of measurement for OC and EC. For the purposes of demonstrating the feasibility of this project, we opted to build only a single sided instrument. This limited the minimum time resolution to approximately one hour. As it presently stands, this means collecting a sample for 35 minutes with the subsequent analysis cycle of 25 minutes. No sample is collected during the analysis cycle. The off-sample analysis time can be reduced by changing the analysis heating profile and by reducing the system purge time. With this limitation noted, for most general monitoring purposes, a time resolution of 2 hours is satisfactory. Using changes in the analysis temperature profile and purge time, sampling time losses of < 20% can be realized. When operated asynchronously in a tandem mode, the instrument will produce 30 minute time resolution with no loss of sample.

Conclusions:
We feel that the results presented for the instrument performance in terms of detection limits, time resolution and comparability to currently accepted sampler systems demonstrate the successful completion of this Phase I development project. Because we were unable to participate in the Phase II program with its' narrow time window, we are proceeding with the commercialization of the instrument on our own. We are working with two different vendors for the development of a cabinet for containment of the entire system as a single unit. The goal is to have a production prototype by early August 2000.

Publications:
n/a

Commercialization Keywords:
Carbon aerosol, OC/EC analyzer, field deployable, real-time, PM 2.5

1