Resource Hub / Technical Literature Library / Sulfur Compound Stability in Air Sampling Canisters How Fill Gas Choice Affects Method Accuracy

Sulfur Compound Stability in Air Sampling Canisters: How Fill Gas Choice Affects Method Accuracy


When it comes to validating a method or protocol, we all rely on samples of known concentrations, whether we buy them or make them ourselves. We also know that for many types of samples, the matrix can bias the results (positively or negatively), particularly for complex matrices like soil or food. A sample of known concentration in a pure solvent often will not adequately approximate the results for a real sample, so some applications require a representative blank matrix that can be fortified and processed according to the method to get an accurate approximation of “real world” conditions. Sulfur compound stability in air sampling canisters is a good example of an application where choice of fill gas (blank matrix) can have a considerable impact on the accuracy of the results you ultimately report.

One might think that because air seems like a relatively clean sample matrix, compared to something like wastewater sludge or human tissue samples, that it does not need to be closely approximated when making samples for method validation or quality control (QC) checks. Indeed, it is common to find air sampling/analysis studies performed using high-purity nitrogen as the surrogate matrix for method development, validation, and routine performance checks. Air is approximately 78% nitrogen, so the use of nitrogen appears to be appropriate, and with it being readily available from the cryogenic equipment commonly used in air analysis laboratories, it seems like a good candidate for making analytical standards.

However, although the majority of our atmosphere is nitrogen, it is not 100% nitrogen. And, the oxygen present in true air samples can dramatically affect sulfur compound stability in air sampling canisters, especially in the presence of water vapor. Figure 1 shows that when using humidified air as the fill gas for a standard, significant loss occurs quickly for hydrogen sulfide in both a Restek and a competitor’s treated (i.e., coated) air canisters. A 50% relative humidity (RH) level was used in this demonstration, but even the driest regions of the most arid desert on Earth (the Atacama in Chile) still have average RH values of around 17%. Since water vapor exists in even the driest atmospheres, correlating results from dry nitrogen environments to actual air samples collected in the field may lead to error because it does not represent the impact of oxygen in the presence of water vapor. In a worst case scenario, if dry nitrogen is the fill gas for standards and QC samples, and you see passing QC recoveries along with undetected levels in your samples, you may report that no sulfur compounds were present in the samples when in fact they were there but had degraded prior to analysis.

Figure 1: Stability of hydrogen sulfide over time under 50% relative humidity conditions in air (error bars = 1 standard deviation; MS detector).

chart, line chart

Effects of Varying %RH in Air on Sulfur Compound Stability in Air Sampling Canisters

A relative humidity of 50% is not at all unusual to encounter in many parts of the world, but how is sulfur compound stability in air sampling canisters affected by different levels of relative humidity? From Figure 2, it is evident that severe degradation of hydrogen sulfide occurs in just a few hours at relative humidity levels above 10%. Considering most places being sampled for the presence of sulfur compounds will be well above 10% RH, this highlights the challenge of maintaining sample integrity when samples need to be shipped to a laboratory for analysis rather than being tested onsite.

Figure 2: Stability of hydrogen sulfide over time under different relative humidity conditions in air using Restek SilcoCan air sampling canister (error bars = 1 standard deviation; flame photometric detector).

chart, line chart

Effect of Humidity on Other Volatile Sulfur Compounds in Air

Hydrogen sulfide is just one of many different volatile sulfur compounds. Do others show the same dramatic decrease over time? Figure 3 illustrates the stability of various sulfur compounds at 21% RH, the first humidity level where substantial loss of sample is observed (performance gets worse at higher humidity levels). Many compounds exhibited a substantial loss of stability, except carbonyl sulfide. Dimethyl disulfide, on the other hand, actually increases in response over time, which is likely caused by its formation from the degradation of the other sulfur compounds in the mix. Polymerization of mercaptans is known to result in the formation of larger sulfur compounds [1].

Figure 3: Stability of various volatile sulfur compounds at 21% relative humidity in air. Note the performance of dimethyl disulfide in particular, which is discussed further in the text. Results for Restek SilcoCan air sampling canister only (error bars = 1 standard deviation; flame photometric detector).

chart, line chart

These results paint a pretty bleak picture of expected sulfur compound stability in air sampling canisters at typical ambient humidity levels when held for more than a few hours, let alone days. The best practice in this scenario is to use humidified air as the fill gas for standards and QC samples and to analyze samples immediately using onsite or mobile equipment. However, if we plot the stability data for the same compounds under 10% relative humidity conditions in air (Figure 4), we see an interesting opportunity to mitigate this stability issue.

Figure 4: Stability of various volatile sulfur compounds at 10% relative humidity in air. Note that even though there was an observed drop in recovery for dimethyl sulfide initially, its performance does stabilize near 80% of the initial measurement. Results for Restek’s SilcoCan air sampling canister only (error bars = 1 standard deviation; flame photometric detector).

chart, line chart

Effect of Sample Drying on Sulfur Compound Stability in Air Sampling Canisters

While Figures 1-4 suggest that sampling done for these compounds under most ambient conditions will lead to rapid degradation of sulfur compounds over a relatively short time even in treated air cans, regardless of the vendor, Figure 5 reveals an opportunity to mitigate this effect by actively drying samples at the point of collection, as has been noted elsewhere in the literature [2,3].

To evaluate this approach, we filled an air canister with 50 ppb of sulfur mix and 50% relative humidity air. We analyzed that sample initially, and then immediately following analysis, we passed the sample from the first air can, through a length of treated ¼" stainless-steel tubing packed with calcium chloride as a drying agent, and into a second air canister where it was analyzed again. After the second analysis, the relative humidity in the second air can containing the dried (i.e., dehumidified) sample was measured. The amount of drying agent was varied by packing stainless-steel tubing of three different lengths. In all three setups studied, the relative humidity dropped from the initial 50% to 0%, as verified with a hygrometer. The results are presented in Table I.

Table I: Relative responses of sulfur compounds after passing through a drying apparatus.

Length of 1/4" Stainless-Steel Tubing Packed with Calcium Chloride drying agent (cm) % Response Compared to Initial Direct Analysis at Time Zero and without Drying
Hydrogen Sulfide Carbonyl Sulfide Methyl Mercaptan Ethyl Mercaptan Dimethyl Sulfide Dimethyl Disulfide
2.5 124 100 137 121 99 79
5 112 98 118 94 100 73
10 13 109 65 46 103 54

This experiment illustrated that a simple device filled with calcium chloride can effectively and dramatically reduce the relative humidity inside the air can, which translates into a much more stable sample compared to air cans that collect air at most ambient humidity levels. The values greater than 100% are believed to be the result of a loss of response in the initial measurement, because even though the sample was tested immediately after it was created, the high humidity air might still create opportunity for sample degradation to occur in the transfer and analysis hardware. This loss in the initial reference sample created an elevated relative response in the stable, dehumidified sample.

Also, the results suggest that too much desiccant in the 5 and 10 cm devices is counterproductive, resulting in analyte loss. A separate water vapor capacity experiment was conducted in which the 2.5 cm length of tubing was packed with 0.3244 g of calcium chloride. This amount of calcium chloride was able to dry 89.5% RH air down to 0% RH for 9.6 L of air (a 0.17 g water capacity), demonstrating that it would be more than sufficient to dry a 6 L air can’s volume of humid air. Therefore, the larger amounts of calcium chloride that negatively impacted recoveries are not necessary. A commercially available hydrocarbon trap containing a calcium chloride drying bed was also tested and its results were poor for the same reasons (results not shown).

In sum, these experiments indicate that drying samples at the point of collection with an appropriate amount of calcium chloride may preserve sample integrity long enough to permit a quick shipment to an offsite laboratory for analysis. However, it is important to emphasize again that method accuracy will still depend in part on how well the blank matrix used for calibration standards and QC samples matches the actual sample matrix. So, if the actual samples will be dried at the point of collection, then calibration standards and QC samples should be treated the same way.

In light of these results, one may wonder if drying the air sample during collection is enough to create a sample that is stable for at least a few days and maybe getting a cheaper, uncoated electropolished stainless-steel air can will do the job. As shown in Figure 5, major loss occurs in an untreated canister, even when reducing humidity in the air canister to 10%. Thus, a coated (deactivated) air sampling can, such as Restek’s SilcoCan air canister, is necessary to avoid losing active compounds like some of the sulfurs analyzed here.

Figure 5: Stability of various volatile sulfur compounds at 10% relative humidity in air. Results for Restek’s electropolished TO-Can air sampling canister only (error bars = 1 standard deviation).

chart, line chart


Even though nitrogen is a major component of the air being sampled, it is not sufficient as a fill gas, even if humidified, when used to represent sulfur compound stability in air sampling canisters for actual samples. The best way to validate the performance of a method or process is to approximate the test conditions as closely as possible. For air sampling, using humid air is a more accurate representation of real-world conditions, making it a better choice than nitrogen fill gas for method validation and quality control. However, because sulfur compound stability in air sampling canisters is often poor under humid conditions, steps should be taken to dry actual air samples or analyze them immediately. By removing humidity at the time of collection, sulfur compound stability can be increased, which can preserve sample integrity during transport, helping to ensure accurate results.


  1. H.W. Chin, R.C. Lindsay, Ascorbate and transition-metal mediation of methanethiol oxidation to dimethyl disulfide and dimethyl trisulfide, Food Chemistry 49 (4) (1994) 387-392.
  2. S. Trabue, K. Scoggin, F. Mitloehner, H. Li, R. Burns, H. Xin, Field sampling method for quantifying volatile sulfur compounds from animal feeding operations, Atmospheric Environment 42 (2008) 3332-3341.
  3. C. Merlen, M. Verriele, S. Crunaire, V. Ricard, P. Kaluzny, N. Locoge, A preconcentration method based on a new sorbent for the measurement of selected reduced sulfur compounds at ppb level in ambient air, Microchemical Journal 143 (2018) 47-54.