The evaluation of OSHA Method No. ID-172, Carbon Dioxide in Workplace Atmospheres

1. The validation consists of the following experimental protocol:
2. Analysis of 3 sets of 6 spiked samples having concentration ranges of approximately 0.5, 1, and 2 times the PEL.
3. Analysis of 3 sets of 6 dynamically generated samples having concentration ranges of approximately 0.5, 1, and 2 times the PEL.
4. Determination of the qualitative and quantitative detection limit for analysis of CO₂ by gas chromatography.
5. Determination of any variation in results when sampling at high and low humidity levels.
6. Comparison of other methods used for CO₂ workplace determinations with the gas chromatographic method.
7. Determination of the storage stability of CO₂ samples collected in gas sampling bags.
8. Assessment of the performance of the gas chromatographic method and conclusions.

2.1. Gas sampling bags were flushed with CO₂-free compressed air [ambient CO₂ and other potential contaminants were removed from the compressed air by using an air scrubber/filtration system (Figure 1) consisting mainly of a Ascarite II/Drierite bed]. A vacuum was then applied to completely evacuate the bags. 

2.2. A known amount of CO₂-free air was metered into each sampling bag. Compressed air flow rates were measured immediately before and after each experiment using a soap bubble flowmeter. Air flow was regulated by using a regulator-rotameter system as shown in Figure 1. Blank samples of the compressed air were periodically collected and analyzed along with the samples and standards. 

2.3. A known amount of CO₂ was metered into each sampling bag containing diluent air. Carbon dioxide flow rates for the spiked samples were determined immediately before and after each experiment using a soap bubble flowmeter. A gas cylinder containing 1.93% CO₂ in air (Air Products, Long Beach, CA, certified analytical standard) was used for spiking the samples. 

2.4. Reference and analytical standards were analyzed along with the spiked samples. The reference standard was purchased commerically (Scott Specialty Gases, Houston, TX, 0.9968% CO₂ in nitrogen, methane, carbon monoxide, oxygen, and hydrogen, certified analytical standard) and analytical standards were generated by dilution of Bone Dry grade CO₂ (Union Carbide, 99.8% min. purity). Carbon dioxide flow for analytical standards was regulated using a regulator-mass flow controller system as shown in Figure 2. 

2.5. Spiked samples, blanks, reference, and generated standards were analyzed by gas chromatography (9.1). Samples were analyzed within 2 days of preparation. Analytical instrument parameters are displayed in Appendix 1. 

Results: Spiked sample results were calculated using a linear regression concentration-response curve. Integrated peak areas were used as signal measurement. Spiked sample recoveries are listed in Table 1. All Analysis data passed both the outlier and Bartlett's tests. The data (Table 1) indicates good precision and accuracy. The coefficient of variation for analysis (CV₁) was 0.034 and the average analytical or spiked recovery was 95.2%. 

3.1. Gas sampling bags were flushed with CO₂-free compressed air. Ambient CO₂ and other potential contaminants were removed from the compressed air by using an air filtration system similar to the one shown in Figure 1. A vacuum was then applied to completely evacuate the bags. 

3.2. A dynamic gas generation system was assembled as shown in Figure 2. A humidity, temperature, and flow control system (Miller-Nelson Research Model 301) was used to control and monitor air flow. A dry test meter (Singer Co., Model # DTM 115) was to measure air flow immediately before, during, and after the experiments. The flow control system was calibrated in-house temperature, humidity, and flow prior to use. Calibration of the dry test meter was done using a spirometer as a primary standard. 

3.3. Carbon dioxide (Bone Dry grade, 99.8% min. purity) gas was introduced into the flow system via a mixing chamber as shown in Figure 2. Carbon dioxide flow rates were taken immediately before and after each experiment using a soap bubble flowmeter. Flow rates were controlled using mass flow controllers (Tylan Model FC260 mass flow controller). 

3.4. Generated samples, blanks, reference, and spiked standards were analyzed by gas chromatography. Reference standards were determined by direct injection of the gas from the canister into the gas sampling valve of the gas chromatograph. Analytical standards were prepared in gas sampling bags and then injected. Samples were analyzed within 2 days of preparation. Instrument parameters used during the analysis are displayed in Appendix 1. 

Results: Generated sample results were calculated using a linear regression concentration response curve. Integrated peak areas were used as signal measurement. Generated sample recoveries are listed in Table 2. As shown in Table 2, the gas chromatographic determinations of CO₂ using gas sampling bags are within NIOSH accuracy and precision guidelines (9.3). The CV_T was 0.026 and the overall recovery was 99.5%. The Sampling and Analysis data shows excellent precision and accuracy. All data passed both the outlier and Bartlett's test. 

4.1. Same as Section 2.1. 

4.2. Blank samples were generated using the flow, humidity, and temperature control system mentioned in Section 3.2. 

4.3. Low concentration CO₂ standards were prepared by mixing CO₂ (1.93% in air), via a mixing chamber, with the treated air. Concentrations of 205.1, 398.9, and 662 ppm were used as standards. 

4.4. Samples were then analyzed by gas chromatography. Analytical conditions used are given in Appendix 1. 

Results: Qualitative and quantitative detection limits are listed in Tables 3 and 4, respectively. The qualitative limit is 200 ppm. The quantitative limit is 500 ppm. A 1-mL sampling loop was used for all analyses. Lower detection limits for CO₂ are possible with larger sampling loops, but should not be necessary for workplace determinations. This assumption is based on the fact that ambient air will always contain a certain amount of CO₂. In well ventilated areas, the level of CO₂ is normally in the range of 300 to 700 ppm. 

6.1. Detector tubes (in-house study) 
A side-by-side (in-house) determination of CO₂ was performed using different types of short-term CO₂ detector tubes and simultaneous gas sampling bag-gas chromatography analysis. Detector tubes were chosen since they were listed as the OSHA sampling method for CO₂ (9.8). As mentioned in the Introduction of the CO₂ method (9.1), an alternative titration method (9.7) was considered unsuitable to use for comparison at the generated concentration levels. Gas chromatographic and detector tube samples were taken at different humidity levels. A synopsis of the side-by-side testing is shown in Table 6. The overall recovery and CV for the gas sampling method displays an improvement over the detector tube technique. Further information regarding the short-term detector tube evaluation can be found in reference 9.6. 

6.2. "Numbering error. This section contains no data." 

6.3. A preliminary evaluation of long-term detector tubes was also performed (9.9). The Draeger model no. 6728611 long-term detector tube and the Mine Safety Appliance Vaporgard Dosimeter were examined. Preliminary testing revealed the Draeger tube unsatisfactory for CO₂ compliance determinations. Only 10 MSA dosimeters were tested and all were from the first lot of production. Dosimeter results were satisfactory; however, further testing and assessment of lot-to-lot variability are necessary. 

6.4. Detector tubes (field study) 
A side by side field sampling evaluation was also performed. The sampling was done at a food freezing plant which used liquid CO₂ as the refrigerant. Gas bag samples and detector tubes were taken in various areas around the plant. Detector tubes were taken at random times and in close vicinity to the sampling bags. A log normal distribution was applied to the data to determine TWAs (for a further discussion of grab samples used to determine TWAs, see references 9.10 and 9.11). Detector tube readings were also taken directly from the personal gas bag samples during the gas chromatographic analysis. Results of the field testing are also listed in Table 6. 

6.5. Miran 1A 
The Miran 1A infrared gas analyzer was also assessed for possible use in CO₂ determinations. As a direct reading instrument, the Miran 1A appeared too sensitive to assess large CO₂ levels sometimes found in industrial settings. An off-scale reading was given when CO₂ concentrations were above 5,000 ppm. However, this response characteristic of the MIRAN appears to make it useful for indoor air quality investigations because CO₂ levels less than 5,000 ppm are normally used to determine ventilation system performance. Carbon dioxide is also used in air quality assessments as a tracer gas to monitor ventilation efficiency. An attempt was also made to use the 5.4-L sampling cell of the gas analyzer as a closed-system analyzer. Samples were collected in gas sampling bags and aliquots were taken from the bags using gas-tight syringes. These aliquots were then injected into the closed system. It was necessary to take 50 to 100-mL aliquots to achieve an adequate signal for CO₂ measurement at the generated levels. The aliquots were considered very large and made accurate and precise analysis difficult. 

Total Samples	Day	Ave % Recovery
10	1-5	95.6
8	9-18	89.7
8	26-31	77.2
7	>31	68.5

Figure 3 graphically depicts the recovery dropping below 90% after 14 days of storage. Sample refrigeration appears to slightly retard CO₂ loss; however, gas bag samples should be sent to the laboratory and analyzed as soon as possible.

9.1. Occupational Safety and Health Administration Technical Center: Carbon Dioxide in Workplace Atmospheres (OSHA-SLTC Method No. ID-172). Salt Lake City, UT. Revised 1990. 

9.2. Mandel, J.: Accuracy and Precision, Evaluation and Interpretation of Analytical Results, The Treatment of Outliers. In Treatise On Analytical Chemistry, 2nd ed., Vol. 1, edited by I.M. Kolthoff and P.J. Elving. New York, NY: John Wiley and Sons, 1978. pp. 282-285. 

9.3. National Institute for Occupational Safety and Health: Documentation of the NIOSH Validation Tests by D. Taylor, R. Kupel and J. Bryant (DHEW/NIOSH Pub. No. 77-185). Cincinnati, OH: National Institute for Occupational Safety and Health, 1977. pp. 1-12. 

9.4. National Bureau of Standards: Experimental Statistics by M.G. Natrella (NBS Handbook 91). Washington, DC: U.S. Department of Commerce, National Bureau of Standards, 1966. Chapter 16, pp. 8-14. 

9.5. Long, G.L. and J.D. Winefordner: Limit of Detection -- A Closer Look at the IUPAC Definition. Anal. Chem. 55: 712A-724A (1983). 

9.6. Occupational Safety and Health Administration Analytical Laboratory: Carbon Dioxide Detector Tubes(PE-2). Salt Lake City, UT. 1987. 

9.7. Norton, J. F., ed.: Standard Methods for the Examination of Water and Sewage. 9th ed. New York, NY: American Public Health Association, 1946. pp. 33-40. 

9.8. U.S. Department of Labor-Occupational Safety and Health Administration: Chemical Information File. Online Database -- OSHA Information System. Washington, DC: Directorate of Technical Support, U.S. Dept. of Labor, OSHA, 1985. 

9.9. Occupational Safety and Health Administration Analytical Laboratory: Carbon Dioxide Detector Tubes -- Long Term (PE-3). Salt Lake City, UT. 1987. 

9.10. National Institute for Occupational Safety and Health: Occupational Exposure Sampling Strategy Manual by N.A. Leidel, K.A. Busch and J.R. Lynch (DHEW/NIOSH Pub. No. 77-173). Washington, DC: Government Printing Office, 1977. 

9.11. Leichnitz, K.: Detector Tube Measuring Techniques. Ecomed, Federal Republic of Germany, 1983. pp. 251-264.