Article pubs.acs.org/ac
Development of a Pretreatment System for the Analysis of Atmospheric Reduced Sulfur Compounds Youn-Suk Son,†,‡ Gangwoong Lee,§ Jo-Chun Kim,*,‡ and Jin-Seok Han∥ †
Exposure, Epidemiology, and Risk Program, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts 02115-6018, United States ‡ Department of Environmental Engineering, Konkuk University, Seoul 143-701, Republic of Korea § Department of Environmental Sciences, Hankuk University of Foreign Studies, Yongin 449-791, Republic of Korea ∥ Climate & Air Quality Research Department, National Institute of Environmental Research, Incheon 404-708, Republic of Korea ABSTRACT: A new pretreatment system was used to evaluate a technology to analyze reduced sulfur compounds (RSCs). To conduct this research, a self-developed custom dryer (Desolvator) and a thermal desorber system (TDS) were installed in the front of GC/PFPD. The syringe pump inside the TDS was devised in such a way that it can be desorbed in a relatively low desorption temperature and low vacuum (730 Pa). When comparing water removal efficiency of the Desolvator and frequently used Nafion dryer, the removal efficiency of the Desolvator stood between 94.6 and 96.1%, considerably higher and more stable than the Nafion dryer (81.3−94.5%). Moreover, analyses were made under various conditions in order to minimize the loss of samples when analyzing sulfur compounds using the TDS, and it was determined that adsorption temperatures less than −25 °C and a flow rate of 50 mL/min were appropriate for the efficient analysis of these sulfur compounds. Moreover, the desorption flow rate and the degree of a vacuum were found to be significant variables for the RSCs desorption. Besides, it was observed that a peculiar peak was formed by thermal decomposition when some sulfur compounds were rapidly desorbed at high desorption temperatures. flame photometric detector (PFPD) and the sulfur chemiluminescence detector (SCD) were recently introduced, which can detect major RSCs at the picogram levels.20−25 Most RSCs exist at tens or hundreds pptv in the ambient air. Therefore, the direct analysis of the ambient air sample using the system requires that analytes present in studied samples be concentrated prior to analysis.3,26,27 It is also known that water, when extracting or analyzing samples, lowers the adsorbent capacity, clogs the cold trap, destabilizes the baseline in chromatography, alters the retention time, and ultimately interferes with the normal detection of the samples.28 Research has been conducted to solve these problems. To this end, a thermal desorption and water removal system has been developed. The main technical part of a thermal desorption system was composed of an adsorption tube and a cooling unit to control the temperature of adsorption and desorption. Diverse adsorbents such as activated charcoal, silica gel, aluminum oxide, graphitized carbon black, molecular sieves, and porous polymers were tested for low concentration sulfur compound samples in the air.5,29 It was reported that silica gel, carbotrap, and the molecular sieve were optimal to concentrate sulfur compounds.26 However, it was also reported those adsorbents did
A
s the issues relating to global warming are gaining increased attention, the environmental significance of reduced sulfur compounds (RSCs) has grown in importance.1,2 A wide variety of RSC forms exist in the atmosphere, and they are found at levels of tens or hundreds of pptv.3 Major substances of such compounds include (CH3)2S (dimethyl sulfide, DMS), COS (carbonyl sulfide), CS2 (carbon disulfide), CH3SH (methyl mercaptan), H2S (hydrogen sulfide), and (CH3)2S2 (dimethyl disulfide, DMDS). In particular, compounds like DMS are formed from plankton in the sea, known to be produced in the ocean in large quantities, playing a pivotal role in the global atmospheric sulfur cycle.1,4 Diverse protocols are used to determine both total and individual RSCs in air matrices. Those protocols are divided into the gas chromatography (GC) and non-GC method.5 The detector types of non-GC methods are ion chromatography (IC),6,7 liquid chromatography−atomic fluorescence spectrometer (LC−AFS),8 and a sensor-based method.9−11 Then again, generally, the GC method, which flame photometric detector (FPD),12,13 mass spectrometer (MS),14,15 flame ionization detector (FID),16 and atomic emission detector (AED)17 are applied to as detectors, is the most common technique due to its excellent quantitative recovery and separation capability.5,18,19 However, the above-mentioned GC detectors are not sensitive enough for the accurate detection of RSCs at ambient levels.20 Accordingly, highly sensitive detection systems such as pulsed © XXXX American Chemical Society
Received: May 3, 2013 Accepted: September 30, 2013
A
dx.doi.org/10.1021/ac401345e | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
not have sufficiently high recovery rates for low molecular sulfur compounds like H2S. Moreover, a recovery efficiency of adsorption media tends to exponentially increase with a decrease in temperature.30 Therefore, a liquefied gas (nitrogen, argon, and etc.) or a Peltier cooler was used to cool down those adsorption media.5,20,23,24,26,31−33 Wardencki announced that the more efficient method was to cool the trap down to −196 °C by liquid nitrogen. However, this created a problem when excessive amounts of methane were present in a sample. And Pollmann et al. also reported that cryogenic preconcentration can cause sampling and chromatographic interference, as the coadsorption of atmospheric water vapor is inevitable at such sub-ambient temperatures.30,34 Therefore, the Peltier cooler which requires no cryogen could be more suitable for the analysis of ambient air samples. Moreover, in terms of maintenance and safety, the cooler was recently considered more preferable than using a liquefied gas for the analysis of sulfur compounds.26 However, Kim reported that the basic detection properties of RSCs can be altered considerably because the operating conditions of GC-TDS (thermal desorber system) are more complicated than the simple GC setup.35 In result of various previous studies, the outcome of GC-TDS-based analysis also was affected by sample conditions such as concentration level and loading volumes.5,33,36,37 Consequently, further testing is required to improve the analytical reliability of GC-TDS for RSCs.34 On the other hand, to remove the existing water in the process of sample extraction and analysis, a Nafion dryer or CaCl2 is widely used.38−41 MacTaggert et al. reported that a Nafion dryer is suitable for sulfur compounds in the range of 1−20 ppbv,42 and it has advantages such as high capacity of water removal, a low and constant pressure drop, and easy maintenance. Hofmann et al. also tested recovery rates of sulfur compounds using Nafion dryer and reported that results were shown to be suitable for measurements of sulfur compounds in dry and wet air in the pptv range under laboratory and filed conditions.39 However, the Nafion dryer must be equipped with additional dry gas or system to make the count current. In this paper, to improve above-mentioned problems, we describe a novel technology which can concentrate and analyze tens of pptv level RSCs in the atmosphere by employing TDS (equipped with syringe pump to make low desorption temperature) and a self-developed custom dryer (which consists of a Peltier cooler and a narrow glass tube and is also called Desolvator in this paper). And then, the technology is evaluated for several chemically unstable odorous RSC substances. Furthermore, the water removal efficiencies of both Desolvator and Nafion dryer are compared with respect to water contents and sampling flow rates. Besides, this study aims at discovering the optimal analysis conditions and response behavior, depending on the temperature and flow conditions in the adsorption and desorption process inside the TDS.
Figure 1. Schematic diagram of a Desolvator showing sampling and cleaning stage.
ready, and then the sample gas is introduced into the TDS through the cool tube when the temperature of the cooler reaches −30 °C. The existing water vapor inside the air sample selectively condenses into a ring-shaped ice frost deposit formed inside the glass tube, while the sulfur compounds of concern such as hydrogen sulfide do not condense due to the considerably different boiling and melting points. Thus, the dry air samples pass out of the cooled tube and into the TDS. At this moment, the air flow rate through the glass tube is 50−100 mL/min. If it is too slow, the glass tube can easily be clogged by the frost, since there is not enough force to break the ice. On the other hand, if the flow rate is too fast, the size of the ring deposit inside the tube gets big because the ringshaped frost may be partially broken, when the desolvation efficiency of the system is reduced. Therefore, it is important to keep a suitable linear velocity. Once the inflow of the samples into TDS is completed, the Desolvator is ready for the second phase, the cleaning stage. In the cleaning stage, the purge gas, such as helium, flows into the 6-port valve to remove condensed frost in the sampling stage. At this time, the direction of the flow
■
EXPERIMENTAL METHODS Concept of Desolvator. To remove water vapor from the sample gas, the run process of a Desolvator, which consists of a Peltier cooler, is divided into two stages: sampling and cleaning. The operation modes of Desolvator are depicted in Figure 1. In the sampling stage, air in the atmosphere moves into a narrow glass tube (o.d.: 0.612 cm; i.d.: 0.4 cm; C: 2.61 cm; L: 11 cm) through a 6-port valve. In this process, a Peltier cooler located outside the tube cools down the glass tube. While the temperature of the Peltier cooler starts to decrease, the sample gas is vent to the air through the tube to get the Desolvator B
dx.doi.org/10.1021/ac401345e | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
vacuum inside the syringe (approximately 730 Pa) because the vacuum generated by the syringe pump can increase the rate of desorption and allow a lower thermal desorption temperature at the same time. After then, a certain amount of the collected samples inside the syringe pump fills the loop (size: 100 μL) before they flow into GC, and the sample in excess of the loop size is released into the atmosphere. By using a certain size loop, it is easier to separate the peak, and the stability of samples with exceedingly high concentrations is secured. In stage (d), samples that filled the loop flow inside GC. In this stage, the direction of the 6-port valve alters again so that the samples flow into GC. Lastly, stage (e) is the system cleaning mode, which cleans out all the paths that the samples pass through by using carrier gas. All the lines, valves, and loops inside the TDS used in all stages of this experiment are made of materials coated with sulfinert to prevent decomposition of sulfur compounds and adsorption. To minimize the uncertainty of analysis by reducing the thermal decomposition of chemicals concerned and the contamination of the systems followed by RSC desorption (except the adsorption tube component which samples pass through), the temperature of all the paths inside GC and TDS, such as switching valves and transfer lines, were maintained at 60−80 °C. Sulfur compounds were analyzed by using TDS, developed based on such principles. In order to secure the optimum recovery rate of analyzed substances, control variables such as desorption temperature, desorption flow, adsorption temperature, and adsorption flow were controlled in the experiment Target Compounds and Analysis Conditions. In this research, in order to make a calibration curve, standard gases of H2S (9.7 μmol/mol), CH3SH (9.8 μmol/mol), DMS (10.0 μmol/mol), and DMDS (10.0 μmol/mol) were used. The gases and gas cylinder that was used were manufactured by Rigas (Korea), and the uncertainty of each substance was about ±5%. The aforementioned four kinds of gases were diluted at a certain ratio in the Tedlar bag to conduct this research. The samples were analyzed using GC (DS 6200, Donam instruments Inc., Korea) /PFPD (model 5380, OI analytical) and TDS. The basic analysis conditions of GC/PFPD and TDS are represented as follows. Detector temperature of PFPD: 250 °C; flow rate (mL/min): air (1) = 10, air (2) = 10, H2 = 11.5; GC column: GC-GasPro (length: 30 m, i.d.: 0.32 mm, Agilent Technologies); carrier gas: He, 50 psi; cold trap: adsorption = −25−25 °C, desorption = 70−200 °C; outlet split: 49:1; flow path temperature: 60 °C; GC transfer line temperature: 80 °C. Calibration. Quantitative calibration was conducted when collecting samples at 50 mL/min for 20 min by using TDS, concentrating them at low temperature and injecting them into the GC/PFPD. The linear correlations between injected mass and peak area were high for all standard gases (r2 > 0.99), as shown in Figure 2. The values of method detection limit (MDL; three standard deviations of the response at the lowest calibration point) of H2S, CH3SH, DMS, and DMDS were 244, 141, 133, and 75 pptv, respectively, based on 1 L of air sample. The overall precision (defined as the variability in peak are response to repeated injections of samples of the same mass) was also below 10% (n = 7).
path inside the valve changes to the opposite direction from that of the sampling stage, and the temperature of the Peltier cooler increases above 150 °C, which cleans the electronic cooling tube in the opposite direction to the sampling. Experimental Method for Removal Efficiency of Water. A Desolvator self-constructed and a Nafion dryer (MD-050-48P-2, Perma Pure Inc.) were installed to compare the removal efficiencies of water vapor. In order to make constant relative humidity (RH: 50 and 100%), the air generated from the zero air system (model 701, API Inc.) was passed through a bubble meter, which is filled with water and provided to the pretreatment system such as a Desolvator and a Nafion dryer using MFC (FC-280S, Tylan) at constant flow rates (flow rates: 50 and 100 mL/min). To estimate the removal efficiency of water by the Desolvator, in the first stage, the sample gas that flowed into the system passed through an electronic cooling tube whose temperature was maintained at −30 °C and the existing water inside the sample was frozen on the tube surface. In the second stage, the remaining water that was not frozen in the electronic cooling tube in the first stage was eliminated in the second glass tube by using dry ice to cool it down to −60 °C. By using this method, the water from the extracted sample in the first and second stages was weighed on an extremely precise scale (Explorer, Ohaus) to the 0.1 mg level. The time spent for overall weight measurements was less than 30 s. The end of each glass tube was sealed by light plastic caps to prevent water evaporation, and the outside of the glass tube was dried very quickly prior to weighing. On the other hand, to estimate water removal efficiency by a Nafion dryer, the Desolvator was replaced by a Nafion dryer to remove water under the same conditions as in the aforementioned first stage. In the second stage, a glass tube was installed at the end of the Nafion dryer and was cooled to −60 °C by using dry ice to quantify the rest of the water that was not eliminated in the first stage. All the tests were carried out by passing through 1 L of air sample, and the temperature in a laboratory was maintained at 25 °C. Design of Thermal Desorber System. Preconcentration techniques are required to analyze traces of RSCs existing in the atmosphere by using GC/PFPD.5 In this study, a cold trap that utilizes an electronic cooling device (Pelier cooler) was used rather than using liquid nitrogen in order to concentrate RSCs existing in the atmosphere. Moreover, a self-developed TDS was used, which desorbs and stores the concentrated sample with the syringe pump and injected it through a loop to an analysis device. At this time, to minimize loss and contamination of the RSCs sample during the high temperature desorption process, the syringe pump was devised in such a way that it can be desorbed at a lower temperature with reduced pressure conditions. Sample movement stages of the self-developed TDS used in this research are divided from stage (a) to (e). Stage (a) is the adsorption mode. In this stage, a certain amount of air sample in the Tedlar bag (SKC Ltd.) is concentrated into the lowtemperature adsorption trap (silica gel 35−60 mesh, SigmaAldrich), using a vacuum pump and a 6-port valve. In Stage (b), the samples from stage (a) are desorbed and the carrier gas is injected into the 6-port valve (Vici Valco Instruments). At this time, the flow path of the 6-port valve changes, the carrier gas passes through the adsorption trap of a certain temperature, and the sample is desorbed. At the same time, the sample is collected inside the glass syringe pump, and then stage (c) is in progress. Here, the major role of the syringe pump is to create a
■
RESULTS AND DISCUSSION Removal Efficiency of Water. Figure 3 shows the removal efficiency in the sampling process based on variables of relative humidity and flow rates by using the Desolvator and Nafion C
dx.doi.org/10.1021/ac401345e | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 2. Calibration curves of four reduced sulfur compounds by GC/PFPD-TDS.
Figure 4. Recovery rates of four sulfur compounds using the Nafion dryer and Desolvator.
(Figure 4). The recovery rates of H2S (23 ppbv), CH3SH (16 ppbv), DMS (13ppbv), and DMDS (8 ppbv) for the Desolvator were approximately 102, 99, 97, and 102%, respectively, while those for the Nafion dryer were 93, 104, 90, and 96%, respectively. These results indicate that the Desolvator is more effective than the Nafion dryer for the analysis of four sulfur compounds. Analytic Efficiency of TDS. Adsorption Temperature. Experiments were performed by changing the temperature of the same silica gel adsorption tube to −25, −10, 0, and 25 °C in order to obtain the optimum conditions of sulfur compounds in accordance with the sample adsorption temperature. Figure 5
Figure 3. Removal efficiency of water by different flow rates and relative humidities of sampling.
dryer. The entire water content was extracted from the Desolvator (or Nafion dryer) and dry ice, and the removal efficiency of each device was calculated based on removed water content from dry ice out of the entire water content (removal efficiency = [entire water content − measured amount from dry ice]/entire water content × 100). As demonstrated in Figure 3, the water removal efficiency of the overall Desolvator stood at around 94.6 ± 0.6 − 96.1 ± 1.7%, which was higher than the removal efficiency of the Nafion dryer (about 81.3 ± 7.1 − 94.5 ± 0.8%) under the same conditions. Sundin et al. reported that the Nafion dryer removes over 90% of the water in the carrier gas at both room and elevated temperatures.43 As identified in Figure 3, the Desolvator was not much influenced by water contents or changes in flow in terms of water removal efficiency. However, the removal efficiency of the Nafion dryer increased from 81.3 to 87.8% when the relative humidity went up from 50 to 100%, which was similarly discovered in a previous research.28 Haberhauer-Troyer reported that the water removal efficiency by a Nafion dryer (MD-050-48P-2) increased from 81 to 90% by the increase of water content. Moreover, the water removal efficiency of the Nafion dryer rose from 87.8 ± 1.4 to 94.5 ± 0.8% as the flow rate increased from 50 to 100 mL/min. It was discovered that the water removal efficiency of the Nafion dryer was considerably influenced by water content and flow. It is assumed that the removal efficiency of the Nafion dryer is similar to that of the Desolvator when the sampling flow rate and water content increase. However, it is rare that relative humidity reaches 100% in the atmosphere. From this perspective, the Desolvator is more effective than the Nafion dryer for water removal since it shows consistent water removal efficiency on a wide range of relative humidity. We did recovery tests to compare the amount of sample lost by the Desolvator and Nafion dryer at the same conditions
Figure 5. Response peak area of sulfur compounds with respect to adsorption temperature in the TDS.
represents example peak areas of five standard substances, depending on the adsorption temperature. As Figure 5 shows, the peak area of DMDS, CH3SH, and DMS changed very little, even when the adsorption temperature changed between −25 and 25 °C. The peak area of H2S showed a similar result when the adsorption temperature ranged between −25 and −10 °C. Our statistical analysis showed that the difference between the two temperature groups was not statistically significant (p = 0.9343 at the significance level of 0.05). On the other hand, the peak area was slightly reduced when the adsorption temperature was 0 °C and the difference between −25 and 0 °C was statistically significant (p = 0.0185). Moreover, the peak area of H2S at 25 °C significantly decreased. RSD (relative standard deviation) on peak area of each malodorous substance in the adsorption trap was on the order of H2S (
