Method for Detection of Trace Metal and Metalloid Contaminants in

Jun 28, 2010 - Method for Detection of Trace Metal and Metalloid Contaminants in Coal-Generated Fuel Gas Using Gas Chromatography/Ion Trap Mass ...
1 downloads 0 Views 63KB Size
Anal. Chem. 2010, 82, 6315–6317

Method for Detection of Trace Metal and Metalloid Contaminants in Coal-Generated Fuel Gas Using Gas Chromatography/Ion Trap Mass Spectrometry Erik C. Rupp, Evan J. Granite,* and Dennis C. Stanko National Energy Technology Laboratory, U.S. Department of Energy, P.O. Box 10940, Pittsburgh, Pennsylvania 15236-0940 There exists an increasing need to develop a reliable method to detect trace contaminants in fuel gas derived from coal gasification. While Hg is subject to current and future regulations, As, Se, and P emissions may eventually be regulated. Sorbents are the most promising technology for the removal of contaminants from coal-derived fuel gas, and it will be important to develop a rapid analytical detection method to ensure complete removal and determine the ideal time for sorbent replacement/regeneration in order to reduce costs. This technical note explores the use of a commercial gas chromatography/ion trap mass spectrometry system for the detection of four gaseous trace contaminants in a simulated fuel gas. Quantitative, repeatable detection with limits at ppbv to ppmv levels were obtained for arsine (AsH3), phosphine (PH3), and hydrogen selenide (H2Se), while qualitative detection was observed for mercury. Decreased accuracy and response caused by the primary components of fuel gas were observed. As worldwide energy demand increases, cleaner, more efficient technologies will become favored. Gasification of coal, such as in the Integrated Gasification Combined Cycle, will increase in prominence. During the gasification of coal, many of the trace elements present in coal are volatilized and remain in the gas phase through the gas turbine.1 Experimental and thermodynamic studies have indicated that Hg (as elemental Hg), Se (as hydrogen selenide, H2Se), As (as arsine, AsH3), and P (as phosphine, PH3) will be present as gaseous metal and metalloid species in fuel gas.2-6 While mercury has long been of environmental concern, there is increasing belief that the other trace elements will have a significant environmental impact if coal continues to be a primary source of energy. Arsine and hydrogen selenide are known * To whom correspondence should be addressed. E-mail: [email protected]. (1) Mojtehadi, W. Combust. Sci. Technol. 1989, 63, 209–227. (2) Helble, J. J.; Mojtehadi, W.; Lyyra¨nen, J.; Jokiniemi, J.; Kauppinen, E. Fuel 1996, 75, 931–939. (3) Norman, J.; Pourkashanian, M.; Williams, A. Fuel 1997, 76, 1201–1216. (4) Richaud, R.; Lachas, H.; Healey, A. E.; Reed, G. P.; Haines, J.; Jarvis, K. E.; Herod, A. A.; Dugwell, D. R.; Kandiyoti, R. Fuel 2000, 79, 1077–1087. (5) Dı´az-Somoano, M.; Martı´nez-Tarazona, M. R. Fuel 2003, 82, 137–145. (6) Lu, D. Y.; Granatstein, D. L.; Rose, D. J. Ind. Eng. Chem. Res. 2004, 43, 5400–5404. 10.1021/ac1012249  2010 American Chemical Society Published on Web 06/28/2010

contaminants of fuel cells,7 and arsine and phosphine poison methanol synthesis catalysts,8 which are important secondary applications for coal-derived fuel gas. Outside of Hg, analytical detection of the target compounds has not been widely studied in the gas phase. Suh et al.9 proposed the use of inductively coupled plasma-dynamic reaction cell-mass spectrometry for the detection of arsine in the gas phase. This involves reacting AsH3 with molecular oxygen to form an AsOion for detection in a MS. Gras et al.10 used gas chromatography and a dielectric barrier discharge detector for ppb level detection of arsine and phosphine in light hydrocarbons. Kumar et al.11 created H2Se from Se found in environmental samples using hydride generation and measured H2Se using atomic absorption spectroscopy (AAS). None of this work was performed in a fuel gas atmosphere. There has been more research focused on analyzing for mercury in fuel gas than for the other trace elements. This research group has previously used an atomic fluorescence spectrophotometer (AFS) for online detection of mercury in nitrogen and noble gases,12 as well as a semicontinuous photodeposition technique to facilitate detection in flue gases.13 Expansion of this work indicated the need for an analytical system capable of handling a more complex gas stream. Uddin et al.14 used temperature programmed decomposition desorption (TPDD) to measure Hg0 removal from fuel gas using differential weight change and AAS to measure Hg0 desorption from activated carbon. Atomic absorption spectroscopy is a standard method for measuring Hg0 in a gas stream but requires a clean carrier gas, such as N2. In prior work from the same group, Wu et al.15 also reported the use of atomic adsorption spectroscopy, although rather than a clean carrier gas, a simulated fuel gas (7) Cayan, F. N.; Zhi, M.; Pakalapati, S. R.; Celik, I.; Wu, N.; Gemmen, R. J. Power Sources 2008, 185, 595–602. (8) Quinn, R.; Dahl, T. A.; Toseland, B. A. Appl. Catal., A: Gen. 2004, 272, 61–68. (9) Suh, J. K.; Kang, N.; Lee, J. B. Talanta 2009, 78, 321–325. (10) Gras, R.; Luong, J.; Hawryluk, M.; Monagle, M. J. Chromatogr., A 2010, 1217, 348–352. (11) Kumar, A. R.; Riyazuddin, P. Microchim. Acta 2006, 155, 387–396. (12) Granite, E. J.; Pennline, H. W.; Hargis, R. H. Ind. Eng. Chem. Res. 2000, 39, 1020–1029. (13) Granite, E. J.; Pennline, H. W. Semi-Continuous Detection of Mercury in Gases, U.S. Patent Application DOE S-104,279, Filed September 2007. (14) Uddin, M. A.; Ozaki, M.; Sasaoka, E.; Wu, S. Energy Fuels 2009, 23, 4710– 4716. (15) Wang, Y.; Duan, Y.; Yang, L.; Zhao, C.; Shen, X.; Zhang, M.; Zhuo, Y.; Chen, C. Fuel Process. Technol. 2009, 90, 643–651.

Analytical Chemistry, Vol. 82, No. 14, July 15, 2010

6315

Table 1. Operating Parameters for Varian GC/MS injector, °C

flow, mL min-1

Hg

350

10 psi

AsH3

180

1

PH3 H2Se AsH3, PH3, H2Se

180 180 180

1 1 1

contaminant

hold 75 °C (4.5 min), ramp 100 °C min-1 (0.25 min), hold 100 °C (0.55 min) hold 40 °C (0.5 min), ramp 20 °C min-1 (3.0 min), hold 100 °C (1.0 min) hold 40 °C (4.0 min) hold 150 °C (4.0 min) hold 40 °C (3.5 min), ramp 60 °C min-1 (1.0 min), ramp 100 °C min-1(0.5 min), hold 150 °C (1.0 min)

was used. Wang et al.16 used the Ontario Hydro Method to capture and speciate Hg, which is a multistep process involving capture with a series of impingers, followed by digestion and cold vapor atomic adsorption. Commercial analytical equipment is available for trace contaminants, with AFS being particularly useful for Hg and elements that form hydrides, such as As, Se, and P. Developing a rapid lab scale analytical technique to qualitatively and quantitatively determine the amount of contaminant present allows for evaluation of contaminant speciation postgasification and performance evaluation of sorbents, which are the most promising technology for trace contaminant removal. Flexibility and ruggedness are also important considerations if the technique were to be adapted to monitoring contaminants in an industrial setting. EXPERIMENTAL SECTION A commercial gas chromatography/ion trap mass spectrometry system has been used to detect AsH3, H2Se, PH3, and elemental mercury. The gas chromatograph is a Varian 450-GC equipped with a Varian CP-Porabond Q Fused Silica column (25 m × 0.25 mm) and a VICI gas sampling valve, used for AsH3, H2Se, and PH3. A Hamilton 250 µL gastight syringe was used for sampling Hg by manual injection. The mass spectrometer is a Varian 240-MS IT Mass Spectrometer equipped with internal electronic ionization. All tubing exposed to the sample gas was stainless steel. Ultrahigh purity (UHP) helium was used as the carrier gas, as obtained from Butler Gas Products, Inc. (McKees Rock, PA). Steady, known concentrations of arsine and phosphine were released using VICI Metronics G-Cal permeation devices. Arsine was released at 487 ng min-1 ± 5% at 25 °C. This device had a linear temperature dependence of 8.72 ng min-1 °C-1 between 10 and 60 °C. Phosphine was released at 55 ng min-1 ± 5% at 25 °C. This device had a linear dependence of 1 ng min-1 °C-1 between 10 and 60 °C. Hydrogen selenide was obtained as a gas mixture with N2 at 103 ± 2 ppm (Matheson Tri-Gas). A VICI Metronics Dynacal permeation device was used to obtain known concentrations of elemental Hg. The permeation device released 302 ± 2 ng min-1 at a temperature of 70 °C. The temperature was held constant using a Haake oil bath at 70 °C. The mass spectrometer trap temperature was 150 °C, the manifold was 50 °C, and the transfer line was 250 °C for all compounds. The GC/MS parameters for each contaminant can be seen in Table 1. The GC injector temperature was held constant (16) Wu, S.; Oya, N.; Ozaki, M.; Kawakami, J.; Uddin, M. A.; Sasaoka, E. Fuel 2007, 86, 2857–2863.

6316

MS scan range, m/z

GC column oven protocol

Analytical Chemistry, Vol. 82, No. 14, July 15, 2010

150-250 50-250 30-70 60-85 30-250

Table 2. Linear Response Range and Detection Limits

AsH3 PH3 H2Se

R2

linear range, ppm

concentration detection limit, ppb

relative standard deviation

0.9951 0.9898 0.9933

1.25-5 0.33-3.95 4.1-41

26 37 207

5.1% 1.4% 3.9%

at 180 °C, except for Hg, where it was held at 350 °C. Carrier gas flow through the column was 1 mL min-1, except for Hg, where the column pressure was held constant at 10 psig. The MS scan ranges were centered on the mass-to-charge (m/z) ratio of the parent ion. Oven parameters varied depending on the contaminant, with the stated goal of elution and detection from the column between 3 and 5 min. Combinations of the parameters in Table 1 were used when AsH3, PH3, and H2Se were present in the same gas stream. RESULTS AND DISCUSSION Column Selection. The Varian PoraBOND Q Fused Silica column was selected due to the column’s broad application range, high stability, and resistance to degradation due to water. Water is a significant component of fuel gas, and any proposed detection method needs to take this into account. This column has previously been used by Gras et al.,10 who also offer a short review of other column options. Response and Detection Limits. Arsine, phosphine, and hydrogen selenide were detected quantitatively and qualitatively using the GC/MS system. The linear response range and detection limits can be seen in Table 2. Each of the three compounds was calibrated in a similar manner. Three injections were made at four known concentrations as well as three blank injections. The linear ranges for the compounds can be seen in Table 2. The selected concentrations were dependent on the source of the contaminant and the experimental system. In each case, the contaminant flow was diluted with varying flow rates of He. Since the permeation tubes (AsH3 and PH3) release a constant mass, the calibration concentrations were altered by adjusting the He flow rate across the permeation tube. Hydrogen selenide concentrations were adjusted by altering the He flow rate to dilute the H2Se/N2 stream. Helium was the only other species present in the detection of arsine and phosphine, while N2 was present with H2Se because of the bottled mixture. In all three cases, the R2 value for a linear regression was at or above 0.99, indicating an acceptable linear response. The precision of the detection was checked by making 10 injections at a concentration on the lower end of the linear

Table 3. Detection Limits in Fuel Gas

AsH3 PH3 H2Se

concentration detection limit, ppb

relative standard deviation

804 123 5335

11.2% 6.6% 11.3%

calibration range in a helium environment. Concentrations for the precision checks were 2.1 ppm (AsH3), 4.1 ppm (H2Se), and 1.0 ppm (PH3). The relative standard deviation for each compound can be found in Table 2. The concentration detection limit (CDL) is defined as CDL )

t·s m

(1)

where t is the value of Student’s t at 98% confidence interval and 10 samples (t ) 2.821), s is the standard deviation of the 10 injections, and m is the slope of linear calibration curve. The concentration detection limit of arsine and phosphine is in the low ppb range, while hydrogen selenide is about 10 times higher. Each of the compounds has been observed at lower concentrations (as low as 5 ppbv for AsH3), but the linearity and precision of the response are unknown. The detection limits are 1 to 2 orders of magnitude below the possible range of these contaminants in fuel gas, indicating an acceptable level of detection. Simulated Fuel Gas. The presence of simulated fuel gas affected the response of the GC/MS, but linearity was maintained. The simulated fuel gas is 50% N2, 27% CO, 20% H2, 3% CO2, and 0.2% H2S. The concentration detection limit, calculated in the same manner as above, can be seen for AsH3, PH3, and H2Se in Table 3. There is an order of magnitude decrease in the concentration detection limit for the three compounds. This is attributable to an increase in the relative standard deviation of the response and a decrease in the absolute response (in the CDL equation, this is accounted for by the slope, m). The increase in standard deviation can be attributed to a more complex gas matrix, where instead of just He (and some N2 for H2Se) and the target compound, there are a number of compounds present due to the fuel gas. The decreased absolute response can be explained by the complex gas matrix, as well as by possible interactions with the S present in the simulated fuel gas. Sulfur is known to interact with the target compounds15,17-19 and could possibly limit the mass of the target contaminant reaching the MS due to upstream deposition. Despite this decrease in accuracy, each of the target compounds was detectable down to approximately 50 ppbv. The concentration detection limits below were presented in order to offer a comparison between clean He detection and fuel gas detection. The decreased level of detection is still appropriate for monitoring trace contaminants in fuel gas. Mercury. Mercury was the initial focus of this work, as it is more regulated and of a broader concern than the other contami(17) Presto, A.; Granite, E. J. Environ. Sci. Technol. 2007, 41, 6579–6584. (18) Uddin, M. A.; Yamada, T.; Ochiai, R.; Sasaoka, E. Energy Fuels 2008, 22, 2284–2289. (19) Presto, A. A.; Granite, E. J.; Karash, A. Ind. Eng. Chem. Res. 2007, 46, 8273–8276.

nants. Quantitative detection of mercury proved to be the most problematic of the four targeted contaminants. Mercury was not observed when the gas sampling valve was used, likely due to deposition in the valve. Sampling was changed to manual injections using a gastight syringe in order to avoid the sampling valve and Hg was detected, although the detection was qualitative only. An attempted calibration of Hg resulted in a R2 value of 0.092. The response was roughly the same independent of the concentration of mercury that was injected. The Hg calibration concentrations ranged from 0.2 to 1.8 ppmv, so condensation of Hg due to a temperature change from 70 °C to room temperature (where the Hg vapor pressure corresponds to roughly 1.8 ppmv) is not responsible for the inability to calibrate Hg. This may be due to difficulties in the electronic ionization of Hg (in its elemental form), contamination of the sampling syringe, or losses within the column. Multiple Contaminant Detection. Although the majority of the work was done with single contaminants in a He gas, research has been performed with all three contaminants present in He and all three contaminants in the simulated fuel gas. Mercury was not included in this work. It was possible to create a chromatographic temperature profile that had the three contaminants elute at different times. The response was linear and in line with the response achieved with individual contaminants. This was repeated in a simulated fuel gas with the same results. CONCLUSION Arsine, phosphine, and hydrogen selenide were detected with a repeatable, linear response using a Varian GC/IT MS. The detection was performed in multiple gaseous environments and with multiple contaminants present, in a time frame (3-5 min) suitable for online measurement. Mercury was detectable but only qualitatively. Fuel gas components lowered the accuracy and response of the MS detector to the point that detection levels are not acceptable for trace contaminant monitoring, as concentrations within fuel gas are expected to be below the determined CDL. However, this procedure is a valuable laboratory tool in controlled conditions and indicates a need for an improved analytical technique. These results indicate that this system could be used for real time detection of gaseous metal and metalloid contaminants in fuel gas at a quantifiable level with further improvement. There is a possibility that these results could be extended to mercury with further experimental refinements. ACKNOWLEDGMENT E.C.R. thanks the National Energy Technology Laboratory for financial support through a postdoctoral fellowship administered by the Oak Ridge Institute for Science and Education (ORISE). Funding support from the DOE Gasification Program is greatly appreciated. The authors also thank Rick Bailey and Rob Tapper from the Varian Corporation for helpful advice. References in this paper to any specific commercial product, process, or service is to facilitate understanding and does not necessarily imply its endorsement by the U.S. Department of Energy. Received for review May 10, 2010. Accepted June 17, 2010. AC1012249 Analytical Chemistry, Vol. 82, No. 14, July 15, 2010

6317