Carbon Monoxide Measurement
Potentially Useful Equipment: Gas Filter Correlation CO Analyzer – Thermo
Environmental Model 48, or Dasibi Model 3008. Mass flow controllers, molecular sieve desiccant, Hopcalite catalyst, compressed air "unknown," chart recorder or data logger.
1. Determine the noise level, whether the noise is Gaussian, the effect of varying the signal averaging time on signal-to-noise level, detection limit, and response time.
2. Measure the mixing ratio of CO in a compressed air cylinder.
3. Measure the amount of CO in the ambient air over the course of a day, and discuss the diurnal variation in terms of emissions and meteorology.
1. Read the instruction manual(s) and "Modifications of a Commercial Gas Filter Correlation CO Detector..., J. Atmos. Ocean Technol., 5(3), 424, 1988, and complete the
qualifier. You will determine the procedure, but these steps should get you started.
2. Turn on the Thermo Environmental CO instrument with the pump off, and connect the outputs to a chart recorder or suitable data acquisition system. Let the instrument warm up until the signal is stable, about half an hour. On channel 1, set the range to the most sensitive level (1.0 ppm full scale), set the time response to the shortest time (10 s), and adjust the zero such that the signal is about 1.0 V. Do not change the span at any time.
Put pen one of the chart recorder on a 2 V full scale setting.
3. Set up channel 2 the same way, except put the time response on 60 s. Record the signal with the pump off for 50-100 ten-second observations. If there was any discernable drift in the 60-s signal (that is greater than the variability), flush the cell with zero air and allow the instrument to warm up another 5 min.
4. Calculate the standard deviation of the 10-s readings, create a histogram, and check for fit to a normal or Gaussian distribution. Check and report the internal temperature and pressure of the instrument. Compute a detection limit.
5. Turn on the pump and let the system equilibrate about 5 min. Check for leaks. Check the calibration with a cylinder of known [CO]; be careful not to waste expensive calibration gas. Check and record the temperature and pressure.
6. Measure the CO in a compressed air cylinder; record the mixing ratio and uncertainty. Use this cylinder to determine the response time.
7. Measure ambient CO for a full day, with an eye on emissions and meteorology that might explain the variability. How important is instrument noise?
Hints: The detection limit is a direct function of the molecular number density of CO in the multi-pass cell, and thus is related to internal pressure and temperature. Please use a filter for particles when sampling ambient air. It is easiest (and most accurate) to measure the contents of a cylinder without dilution, that is neat, but do not measure our NIST standard that way.
For the Report: Calculate the detection limit for a specific signal-to-noise ratio, confidence interval (σ level), and integration time. Can you improve the detection limit with greater averaging times? Do you think there is any limit to the averaging time that can be used? From the literature, determine the "clean air background" CO level; did you observe such concentrations in College Park? With what uncertainty can these instruments measure the background mixing ratios?