Introduction

The procedure for collection and analysis of air samples for hydrogen sulfide (H₂S) is described in OSHA Method No. ID-141 (12.1.). Briefly, H₂S is collected on Whatman No. 4 filter paper (Special Order, Whatman Labsales, Hillsboro, OR) which has been impregnated with silver nitrate (AgNO₃). The H₂S reacts with AgNO₃ to form silver sulfide, a greyish-black precipitate (12.2., 12.3.). The silver sulfide is dissolved in an alkaline cyanide solution and then analyzed by differential pulse polarography (DPP) for sulfide.

1. Experimental Procedure

This method has been evaluated using a sampling rate of 0.2 L/min and a total air volume of 2 L. The OSHA Ceiling Permissible Exposure Limit (PEL) was 20 ppm when this evaluation was performed; therefore, concentrations of approximately 10, 20, and 40 ppm H₂S were used for the evauluation. The evaluation followed the NIOSH protocol (12.4.) for method evaluation with some exceptions and consisted of the following experiments:

1. An analysis of a total of 18 spiked samples (6 samples at each of three test levels) prepared by spiking known amounts of sodium sulfide.
   
2. An analysis of a total of 18 samples (6 samples at each of the three test levels) collected from dynamically generated test atmospheres using a 2-L total air volume.
   
3. A determination of the collection efficiency and breakthrough of the AgNO₃-impregnated filters at about 40 ppm.
   
4. A test of the storage stability of 18 collected samples.
   
5. A determination of the detection limit of the method.
   
6. A comparison of sampling and analytical methods.
   
7. An assessment of the precision and accuracy of the method.
   
8. An assessment of the method using an independent laboratory.
   
9. Conclusions - including a discussion of recent PEL changes for H₂S.

2. Analysis

Samples were spiked and analyzed to determine analytical performance of the method and to assess desorption efficiency.

Procedure: Samples were prepared by adding known amounts of a standardized sodium sulfide (Na₂S) stock solution to AgNO₃-impregnated Whatman 4 filters. Approximately 25, 50, and 100 µg (as H₂S) were spiked. Details are discussed below.

2.1. A Na₂S stock solution was prepared by dissolving and diluting 35.855 g of Na₂S·9H₂0 to 500 ml with deoxygenated 0.1 M NaOH. The stock was standardized using the procedure listed in the method (12.1.). This stock solution was equivalent to 10,175 µg/mL H₂S. The solution was protected from light and stored in a refrigerator.

2.2. Silver nitrate-impregnated Whatman 4 filters were prepared as mentioned in reference 12.1.

2.3. The three sets of spiked samples were prepared by injecting known volumes of the standardized Na₂S stock solution on the AgNO₃-impregnated filters. The spiked filters were air dried in a dark environment for approximately 24 h prior to resorption. Each set contained six samples and a sample blank.

2.4. Sample solutions were prepared and analyzed for sulfide by DPP as described in reference 12.1. with one slight modification: Analytical standards were prepared by serial dilution of the standardized Na₂S stock solution (10,175 µg/mL H₂S) prepared in Section 2.1. Each sample was analyzed twice.

Results: The results of the spiked filter samples are given in Table 1. The desorption efficiency (DE) for each sample set analyzed was 0.923, 0.892 and 0.882, respectively. The DE at each level is significantly different from 1.0. The slightly low DE appeared to be due to difficulties in spiking the filters with the Na₂S stock solution and not to problems in the desorption of analyte. Potential problems when spiking these solutions were noted:

1. Sodium sulfide is an unstable compound (12.5.). Exposure of sodium sulfide crystals to air produces H₂S.
   
2. Aqueous solutions of Na₂S are slowly converted to sodium thiosulfate (Na₂S₂O₃) from standing in contact with air.

3. Sampling and Analysis

Procedure: To determine the ability of the method to sample and analyze for H₂S, known concentrations of generated samples were prepared. Test atmospheres of H₂S gas in this and all following sections were dynamically generated by diluting H₂S gas from a cylinder with purified humid air in a Teflon mixing tee. The cylinder concentration had been certified as 220 ± 5 ppm H₂S in nitrogen (Airco, Murray Hill, NJ). To verify the concentration of the cylinder, samples were taken and analyzed using the NIOSH sampling and analytical method no. S4 (12.6.) with some modifications. The procedure followed is listed below.

3.1. Three samples of H₂S gas from the cylinder were collected in midget impingers containing 10 ml of a cadmium hydroxide/STRactan absorbing solution at approximately 0.1 L/min for 5 min. A backup impinger containing this solution was connected in series to each sample to assure no loss of analyte from breakthrough.

3.2. The samples were analyzed for sulfide content by the methylene blue calorimetric procedure (12.6.). Standards were prepared from the Na₂S stock solution prepared in Section 2.

3.3. No H₂S was detected in the backup impinger of each sample (detection limit = 1.6 ppm H₂S).

3.4. The average H₂S concentration of the gas cylinder was 208 ± 14 ppm (95% confidence level). This experimental concentration did not significantly differ from the manufacturers' stated concentration of 220 ± 5 ppm. The stated value was used when calculating all theoretical (taken) results.

Test atmospheres were then generated using this gas and the generation system described in the Appendix. Six samples were simultaneously collected from the manifold for approximately 10 min. This was performed for each of the three test levels. The samples were collected on the AgNO₃-impregnated filters with model C-210 portable sampling pumps (Mine Safety Appliances, Pittsburgh, PA) at high (85 to 88%) RH and 25.5°C.

Results: The three sets of samples were analyzed twice by DPP using procedures described in the method (12.1.). The results of the sampling and analysis experiment are shown in Table 2. A correction for DE was not performed on the results. The theoretical (taken) concentration of the generated gas was determined from the measured flow rates of the diluting air and the certified H₂S gas from the cylinder at each test level. Two results were deleted as outliers since they failed to pass the outlier test at the 99% confidence level.

4. Collection Efficiency (CE)

Procedure: Six samples were simultaneously collected on AgNO₃-impregnated filters which were connected in series to backup midget impingers containing 10 ml of a cadmium hydroxide/STRactan absorbing solution. Samples were collected at high (85-88%) RH and 25.4°C. The samples were collected at a concentration of approximately 40 ppm for 10 min [Note: These were the same samples which were generated and analyzed in the sampling and analysis experiment (Section 3) at about 40 ppm]. The amounts of H₂S collected in the filters and backup impingers were then measured.

Results: The CE of each filter sample was calculated by dividing the amount of H₂S collected on the filter by the total amount of H₂S collected in the filter and backup impinger. The results are given in Table 3. The CE was 100%.

5. Breakthrough

Procedure: Two samples were simultaneously collected at approximately 40 ppm H₂S for 10, 15, and 20 minutes. The generation system test atmospheres were produced at a low (18%) and then a high (86%) RH (25°C) to determine any humidity effect on breakthrough. Each sample was collected on a AgNO₃-impregnated filter connected in series to a backup midget impinger containing 10 ml of cadmium hydroxide/STRactan absorbing solution. The midget impinger samples were analyzed for sulfide content using NIOSH method no. S4 (12.6.) for H₂S with some modifications. Breakthrough of H₂S into the impinger solutions at a concentration greater than 5% of the amount generated was considered significant.

Results: Breakthrough was determined for each sampling period by dividing the average amount of H₂S collected in the backup impingers by the known generated H₂S concentration. The results are presented in Table 4. Breakthrough of 6% occurred at both humidities and for a sampling period of 20 min. Breakthrough was below 5% for shorter sampling periods.

6. Storage Stability

A study was conducted to assess the stability of H₂S air samples collected on the AgNO₃-impregnated filters.

Procedure: Three sets (6 samples and a blank sample in each set) of H₂S samples were generated using AgNO₃-impregnated filters as the collection media. The generation system was set at high (85-88%) RH and 25.5°C. The samples were collected at approximately 20 ppm for 10 min. The samples were then stored in a dark environment at normal laboratory temperatures. A set of samples was analyzed after sample storage of 5, 14, and 30 days.

Results: The results of the storage stability study are shown in Table 5 [Note: The six samples listed for the 0 day storage test are the same samples taken at 20 ppm for the sampling and analysis (Section 3) experiment]. The results indicate samples are stable for a period of at least 30 days when stored under normal lab temperatures and in a dark environment.

7. Detection Limit

Procedure: Standard solutions in 0.1 N NaOH were prepared by serial dilutions of a stock standard solution of Na₂S. The procedure used is identical to the preparation of working standards in the method (12.1.), with the exception that the concentrations used for this experiment were from 10.2 to 203.5 ng/mL. Six standards at each concentration and six reagent blanks were prepared and analyzed.

Results:

7.1. Qualitative Detection Limit

The analytical procedure requires the use of a blank subtraction software routine for each sample or standard. This routine sets any constant background signal to zero. Therefore, parametric or non-parametric tests which require measurable analytical signals from reagent blank samples could not be used to determine the qualitative detection limit (12.7.). The qualitative limit was estimated from results in Table 6 to be 0.020 µg/mL or 1.0 µg of H₂S in a 50 ml sample extraction volume. This corresponds to 0.4 ppm H₂S (2-L air volume). The coefficient of variation (CV) of replicate determinations of standards at this level was 0.128. Concentrations of H₂S below 0.020 µg/mL were not consistently resolved when analyzed.

7.2. Quantitative Detection Limit

The quantitative detection limit is 0.050 µg/mL or 2.5 µg of H₂S in a 50 ml sample extraction volume. This corresponds to 0.9 ppm H₂S (2-L air volume). As shown in Table 6, the CV of replicate determinations of standards at this concentration is less than 0.10.

8. Method Comparison

A side-by-side method comparison was performed as an independent measurement of the generated concentrations. The NIOSH sampling and analytical method no. S4 (12.6.) was used with some modifications. The samples for the NIOSH method were taken simultaneously with impregnated filter samples during the sampling and analysis experiment (Section 3).

Procedure: Six samples (NIOSH method) of the test atmosphere were simultaneously collected from the sampling manifold for approximately 10 min at a flow rate of 0.2 L/min at each of the three test levels. Each set of six samples was concurrently collected with the filter samples at that test concentration. The procedures used for the NIOSH method are listed below.

8.1. Samples were collected in midget impingers containing 10 ml of a cadmium hydroxide/STRactan absorbing solution with MSA Model C-210 portable sampling pumps.

8.2. The samples were analyzed for their sulfide content by the methylene blue colorimetric procedure with standards prepared from the standardized Na₂S stock solution.

Results: The average H₂S concentration (95% confidence level) determined in each set of impinger samples is:

H2S Taken Concentration		H2S Found Concentration
10.06 ppm		10.07 ± 0.17 ppm
22.18 ppm		21.94 ± 0.29 ppm
39.22 ppm		37.19 ± 0.92 ppm

The first two experimental concentrations found did not significantly differ from the taken concentrations. The difference between the taken and found concentrations for the third test (39.22 vs. 37.19 ppm) was possibly due to analytical difficulties in recovering the cadmium sulfide precipitate from the impingers at this test level.

9. Precision and Accuracy

The precision and accuracy data, based on the NIOSH statistical protocol (12.6.), are presented in Tables 1 and 2. The pooled coefficients of variation for spiked (CV₁) and generated (CV₂) samples and the overall pooled CV_T are as follows:

CV₁ = 0.031        CV₂ = 0.036        CV_T = 0.038

The bias was -3.1% and overall error was ±10.7%.

10. Independent Assessment of Method

Procedure: A series of sample filters impregnated with AgNO₃ were prepared at the OSHA Analytical Laboratory and submitted to an independent laboratory for sample collection at low and high humidity using their H₂S generation system. Their system dynamically generated H₂S gas by diluting a gas stream from a cylinder of H₂S with a stream of air from a compressed air cylinder. The dilution took place in a Teflon manifold. Flow rates were controlled using rotameters. The high humidity experiment was conducted by bubbling the diluting air in water before mixing. The known (taken) concentrations were determined by measuring aliquots of the generated atmospheres with a model 5700 gas chromatography (Hewlett-Packard, Avondale, PA) equipped with a flame photometric detector. The detector response was calibrated using a H₂S permeation tube.

Samples were taken at three different concentrations and two different humidities by employees of the independent laboratory. A flow rate of 0.2 L/min was used with model 222-3 sampling pumps (SKC Inc., Eighty Four, PA). The samples were then submitted to the OSHA Analytical Laboratory for analysis. The samples taken at low humidity were collected and analyzed within 30 days of preparation; high humidity samples were collected about 45 days after preparation. The samples taken at high humidity were analyzed about 70 days after the filters were impregnated.

Results: Results are shown in Table 7. The results of the two different humidity tests indicate good agreement with the theoretical values with the exception of samples collected at about 40 ppm. All recoveries for the 40 ppm samples were about 50% lower than expected. The independent laboratory indicated monitoring problems occurred when performing the first test (low humidity) at this concentrate ion. The problem was attributed to a faulty H₂S permeation tube when calibrating the gas chromatography. For the second high concentration test (high humidity), recoveries were also about 50% low; however, the independent laboratory indicated the impregnated filters used during the 40 ppm test were not handled according to specifications. Blank air was collected through these filters a month prior to their use.

The length of time between preparation and analysis for high humidity samples (at concentrations at or below 30 ppm H₂S) indicates a storage stability of at least 45 days.

11. Conclusions

This sampling and analytical method has been shown to be precise and accurate for determining Ceiling exposures of 10 to 40 ppm when using 0.2 L/min flow rates for 10 to 15 min. To determine compliance with the Final Rule STEL of 15 ppm H₂S, the same sampling and analytical conditions can be used.

Storage stability did not pose a significant problem under the conditions tested. Breakthrough was evident after 20 min of sampling at 40 ppm which places a limitation on sampling. Due to the potential for breakthrough, it is recommended to sample for TWA exposures at a lower flow rate of 0.1 L/min. Eight 1-h samples are recommended for TWA assessments of H₂S. Samples taken near the TWA PEL of 10 ppm (0.1 L/min for 1 h) will have the same total amount collected as the samples collected at 20 ppm (0.2 L/min for 15 min). Therefore, at the lower flow rate the method should not exhibit any significant sample collection problems when determining TWA exposures.

12. References

12.1. Occupational Safety and Health Administration Technical Center: Hydrogen Sulfide in Workplace Atmospheres by T. Wilczek (USDOL/OSHA-SLTC Method No. ID-141). Salt Lake City, UT. Revised 1989.

12.2. Natusch, D.F.S., H.B. Klonis, H.D. Axelrod, R.D. Teck, and J.P. Lodge, Jr.: Sensitive Method for Measurement of Atmospheric Hydrogen Sulfide. Anal. Chem. 44: 2067-2070 (1972).

12.3. Natusch, D. F. S., J.R. Sewell, and R.L. Tanner: Determination of Hydrogen Sulfide in Air -- An Assessment of Impregnated Paper Tape Methods. Anal. Chem. 46: 410-415 (1974).

12.4. 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.

12.5. Windholz, H., ed.: The Merck Index. 9th ed. Rahway, NJ: Merck & Co., Inc., 1976.

12.6. National Institute for Occupational Safety and Health: NIOSH Manual of Analytical Methods. 2nd ed., Vol. 2 (DHEW/NIOSH Pub. No. 77-157-B). Cincinnati, OH: National Institute for Occupational Safety and Health, 1977. pp. S4-1-S4-100

12.7. Occupational Safety and Health Administration Analytical Laboratory: OSHA Analytical Methods Manual (USDOL/OSHA-SLCAL Method No. ID-102 (Backup Report)). Cincinnati, OH: American Conference of Governmental Industrial Hygienists (Pub. No. ISBN: 0-936712-66-X), 1985.

Generation System for Producing Dynamic Test Atmospheres of Hydrogen Sulfide

A generation system was designed such that the H₂S gas and diluent air were connected to a mixing tee which was then connected to a sampling manifold. The diluent air was conditioned to the RH and temperature required for the particular experiment. Details of the system are listed below.

1. Description of equipment used to monitor and control gases.
   1. The flow rate, temperature, and RH of the diluent air were controlled with a model HCS-201 Mass Flow, Temperature, and Humidity Control System (Miller-Nelson Research Inc., Monterey, CA). The humidity and temperature of the air were measured with a model 400D % Relative Humidity/Temperature Monitor (General Eastern, Watertown, MA).

   2. The flow rate of the H₂S gas was controlled with mass flow controllers (Tylan Corp., Carson, CA).

   3. The flow rate of the H₂S gas and the air were measured just prior to and just after sampling. The flow rate of the H₂S gas was measured with a model 823-1 Electronic Bubble Flowmeter (Mast Development Co., Davenport, IA). The flow rate of the humid air was measured with a dry test gas meter which had been previously calibrated against a primary standard.

2. Description of sampling manifold and connective tubing:
   1. A 12-port sampling manifold was constructed from Teflon tubing and fittings. There were six sampling ports located on opposite sides of the sampling manifold.

   2. The sampling manifold was directly connected to a Teflon mixing tee by means of Teflon tubing and Teflon L fittings.

   3. The sampling devices were attached to the manifold with Teflon tubing and to the sampling pumps with Tygon tubing.

   4. All other connective tubing and fittings were made of stainless steel or Teflon to avoid any contamination problems.