Gas to Particle Conversion-Gas Exchange Technique for Direct

Sep 23, 2014 - ABSTRACT: A novel gas to particle conversion-gas exchange technique for the direct analysis of metal carbonyl gas by inductively couple...
0 downloads 0 Views 403KB Size
Download PDF
Technical Note pubs.acs.org/ac

Gas to Particle Conversion-Gas Exchange Technique for Direct Analysis of Metal Carbonyl Gas by Inductively Coupled Plasma Mass Spectrometry Kohei Nishiguchi,† Keisuke Utani,† Detlef Gunther,‡ and Masaki Ohata*,§ †

J-Science Lab Co. Ltd., 3-1 Hiuchigata, Kamitoba, Minami-ku, Kyoto 601-8144, Japan ETH Zurich, Department of Chemistry and Applied Biosciences, Laboratory of Inorganic Chemistry, Wolfgang-Pauli-Strasse 10, 8093 Zurich, Switzerland § National Metrology Institute of Japan (NMIJ), National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1, Umezono, Tsukuba 305-8563, Japan ‡

ABSTRACT: A novel gas to particle conversion-gas exchange technique for the direct analysis of metal carbonyl gas by inductively coupled plasma mass spectrometry (ICPMS) was proposed and demonstrated in the present study. The technique is based on a transfer of gas into particle, which can be directly analyzed by ICPMS. Particles from metal carbonyl gases such as Cr(CO)6, Mo(CO)6, and W(CO)6 are formed by reaction with ozone (O3) and ammonium (NH3) gases within a newly developed gas to particle conversion device (GPD). The reaction mechanism of the gas to particle conversion is based on either oxidation of metal carbonyl gas by O3 or agglomeration of metal oxide with ammonium nitrate (NH4NO3) which is generated by the reaction of O3 and NH3. To separate the reaction gases (remaining O3 and NH3) from the formed particles, a previously reported gas exchange device (GED) was used and the in argon stabilized analyte particles were directly introduced and measured by ICPMS. This new technique provided limits of detection (LOD) of 0.15 pL L−1 (0.32 ng m−3), 0.02 pL L−1 (0.07 ng m−3), and 0.01 pL L−1 (0.07 ng m−3) for Cr(CO)6, Mo(CO)6, and W(CO)6, respectively, which were 4−5 orders of magnitude lower than those conventional applied for detecting these gases, e.g., gas chromatography with electron captured detector (GC-ECD) as well as Fourier transform-infrared spectroscopy (FT-IR). The achieved LODs were also similar or slightly better than those for ICPMS coupled to GC. Since the gas to particle conversion technique can achieve the direct measurement of metal carbonyl gases as well as the removal of reaction and ambient gases from metal carbonyl gases, the technique is considered to be well suited to monitor gas quality in semiconductor industry, engine exhaust gases, and or waste incineration products.

I

can be directly introduced into ICPMS, whereas trace amounts of gases such as metallic gas are exchanged by Ar and therefore lost within the GED. It is well-known that the analysis of metal carbonyl gases such as Ni(CO)6 and Fe(CO)6 in CO and other gases is very difficult because the gas sampling method such as impinger method using acids can not trap them as sufficient as necessary for quantification.12−17 In order to overcome the limitation, we developed an effective sampling method which can trap the metal carbonyl gas in hydrochloric acid (HCl) in a vacuum cylinder.12,13,17 The ozone (O3) gas was required to oxidize the metal carbonyl gas to a metal oxide which was then absorbed in HCl. When the required oxidation process was carried out, we found white fume occurred immediately by the reaction between metal carbonyl gas and O3, which was anticipated because of the conversion from metal carbonyl gas to particles of metal oxide. On the basis of this observation and the knowledge about the

nductively coupled plasma mass spectrometry (ICPMS) is widely applied for trace element analysis due to its high sensitivity, multielement capability, and a wide linear dynamic range.1,2 Most of the samples are introduced as liquid or solid aerosols covered in argon (Ar) or mixtures of Ar and helium. Unfortunately, a high flow rate (several hundred milliliters per minute) of ambient air would extinguish the ICP immediately, which limits the direct analysis of airborne particles. To overcome this limitation a gas exchange device (GED) has been proposed and its capabilities have been successfully demonstrated on a variety of applications.3−11 For example, direct real-time multielement analysis of airbone particulate matter has been extensively reported.3−6,8 The GED has also been coupled to laser ablation (LA-ICPMS) for direct atmospheric sampling of laser generated aerosols for the determination of trace elements as well as the isotope ratio analysis, in particular for larger samples, which cannot be placed in a commonly required ablation chamber.7,9,10 Because the GED consists of a porous silica membrane tube, it can not only be used for the exchange of ambient air but also for other gases. For example, direct multielement analysis of tobacco smoke has been demonstrated.11 Though particles in ambient air or other gases © 2014 American Chemical Society

Received: June 12, 2014 Accepted: September 23, 2014 Published: September 23, 2014 10025

dx.doi.org/10.1021/ac502168h | Anal. Chem. 2014, 86, 10025−10029

Analytical Chemistry

Technical Note

Figure 1. Schematic diagram of the reaction chamber for a gas to particle conversion device (GPD) which is based on ozonation/NH3 supported particle formation.

possibility of direct analysis of 0.1 μm particles using GED,3−11 we studied several gas reactions for gas to particle conversion in more detail and found the reaction using O3 and NH3 gases was the most efficient to generate particles from the metal carbonyl gas. The present study demonstrates a novel gas to particle conversion-gas exchange technique for the direct analysis of metallic gases such as Cr(CO)6, Mo(CO)6, and W(CO)6 by ICPMS, and the working principle, the reaction mechanism, and figures of merits are discussed.

■

Table 1. Operating Conditions of Gas to Particle Device (GPD), Gas Exchange Device (GED), and ICPMS GPD (J-Science Labo) sample introduction flow rate: NH3 gas flow rate:

EXPERIMENTAL SECTION

1% O3 gas flow rate:

100, 150, 200, 250 mL min−1 (MFC-1 + MFC-2 in Figure 2) 25 and 50 mL min−1 from 2% or 4% NH3 solution 25 and 50 mL min−1

GED (J-Science Labo) sweep gas flow rate of Ar:

2500 mL min−1

ICPMS (Agilent 7500cs) ICP: rf forward power: plasma gas flow rate (Ar): auxiliary gas flow rate (Ar): carrier gas flow rate (Ar): measurement mode: measured mass and elements: dwell time for each measured mass: collision gas:

Instruments. Figure 1 shows the schematic diagram of the reaction chamber of the gas to particle conversion device (GPD) developed in the present study (J-Science Labo. Co. Ltd., Kyoto, Japan).18 Because the GPD was prototype, the size as well as the length of the reaction chamber was not optimized with respect to the gas to particle conversion. Therefore, we studied flow rates of sample, O3, and NH3 gases to optimize the operating condition as well as characterize the reaction mechanisms. The sample gas containing the metallic gas (Cr(CO)6, Mo(CO)6, and W(CO)6) is introduced into the reaction chamber where it is mixed with O3 and NH3. The 1% O3 is supplied by an ozonator which is equipped with the GPD. The required NH3 gas was obtained from a 2−4% ammonia solution prepared from a 20% NH3 solution (Ultrapur, Kanto Chemical Co. Inc., Japan) by bubbling Ar through a PFA bottle. Though the NH3 gas could be obtained by a gas cylinder, e.g., NH3 gas diluted in Ar, we chose the NH3 solution due to its convenience for use as well as flexibility of changing the concentration. The reaction with O3 and NH3 provokes the formation of particles, which are online transported into the ICPMS via GED (J-Science Labo. Co. Ltd., Kyoto, Japan).3−5 Because the GED consists of two concentric glass tubes with pores of 0.07 μm in diameter, it acts as a membrane. Since the introduced ambient air and the remaining gases from the reaction (O3 and NH3) were exchanged by Ar gas across the membrane, the particles converted by GPD and stabilized in Ar gas were introduced directly into the ICPMS. Because of the absence of ambient air, O3, and NH3, the ICP can be operated at parameters commonly used in ICPMS. The mixed metal carbonyl gases of Cr(CO)6, Mo(CO)6, and W(CO)6 used for the demonstration of the GPD-GED-ICPMS were produced by a metal standard gas generator (MSG2) which was also developed by J-Science Labo. Co. Ltd. (Kyoto, Japan).12,13 The MSG2 contained the solid form of metal carbonyls and sublimated gases were used for the demonstration. All measurements were carried out using an Agilent 7500cs ICPMS (Agilent Technologies Inc., Tokyo, Japan) without the use of the collision gas. The operating conditions of the different instruments are summarized in Table 1.

shielded 1600 W 15.0 L min−1 1.0 L min−1 ∼1100 mL min−1 (MFC-4 in Figure 2) spectrum 53 Cr, 95 Mo, 97 Mo, 98 Mo, 182W, 183W, 184W 100 ms none

■

RESULTS AND DISCUSSION Reaction Mechanisms for Gas to Particle Conversion with O3 and NH3. The oxidation for metal carbonyl gas by O3 for the gas to particle conversion within the GPD can be described as follows: 2Cr(CO)6 + 5O3 → Cr2O3 + 12CO2 Mo(CO)6 + 3O3 → MoO3 + 6CO2

W(CO)6 + 3O3 → WO3 + 6CO2

The NH3 gas supplied was considered to generate particles of ammonium nitrate (NH4NO3) when reacting with O3 as described below. 2NH3 + 4O3 → NH4NO3 + H 2O + 4O2

The metal oxides generated with ozone form agglomerates with the NH4NO3 particles. In the reaction chamber the agglomerate grow to a particle size that cannot pass through the porous membrane of the GED and remain in the Ar carrier gas entering the ICPMS. The NH4NO3 nanoparticles acting as carrier are readily dissociated within the ICP ion source without forming any additional spectroscopic interference when compared to the introduction of nitric acid blank solutions. Operating Conditions and the Efficiency of the Gas to Particle Conversion. Various operating conditions of the GPD 10026

dx.doi.org/10.1021/ac502168h | Anal. Chem. 2014, 86, 10025−10029

Analytical Chemistry

Technical Note

Figure 2. Schematic diagram of the experimental setup for optimizing the operating conditions of the GPD device. The total Ar gas flow rate of MFC 1, 2, and 3 was fixed at 300 mL min−1 to keep the concentration of metal carbonyl gas constant for both the direct introduction to ICPMS and the introduction to ICPMS via GPD-GED.

Figure 3. Signal intensities and ratios of 98 Mo for both introduction systems as a function of the flow rate of Mo(CO)6. The flow rates for both 1% O3 and NH3 gases for gas to particle conversion in GPD were 25 mL min−1 each and the concentration of the NH3 solution was 2%.

2). The ratios of Mo obtained from signals when introducing the sample through the GPD-GED divided by the signals measured by direct introduction are also shown in Figure 3. The flow rates of 1% O3 and NH3, which were generated by an ozonator and the bubbling of 2% NH3 solution, respectively, were 25 mL min−1 each. As can be seen in Figure 3, similar intensities of Mo were observed for the direct gas introduction. At 100 mL, 25% lower intensity for the GPD-GED approach was observed when compared to the direct sample introduction, which decreased further with increasing flow rates. In order to improve the ratio, we studied the operating conditions of GPD in more details. Figure 4 shows the results of the ratios of Mo obtained under different operating conditions. We increased the flow rates for both 1% O3 and NH3 as well as the concentration of NH3 solution from 2% to 4%. It could be seen in Figure 4, the ratio improved when compared to the results shown in Figure 3. The highest intensity was reached for flow rates of 100−150 mL min−1 and 50 mL min−1 for 1% O3 gas and NH3, respectively, when the reaction chamber designed (see Figure 1) was used for gas to particle conversion. Because the lower flow rate showed

were examined using mixed metal carbonyl gases of Cr(CO)6, Mo(CO)6, and W(CO)6 from the metallic standard gas generator (MSG2) as functions of flow rates of O3 and NH3 gases. Figure 2 shows the schematic diagram of the experimental setup for the optimization of the operating conditions of the GPD. Four mass flow controllers were installed to investigate the influence of different gas settings on the particle conversion and excitation within the ICP. The Ar gas flow rate of mass flow controller-1 (MFC-1) was fixed at 50 mL min−1, and the Ar gas flow rate of MFC-2 was changed in order to examine the reaction between the metal carbonyl gas and O3 as well as NH3 in the reaction chamber of the GPD. The total flow rate of Ar was fixed at 300 mL min−1 at the outlet of the GED which allowed the concentration of the metallic gas introduced into ICPMS to be constant by combining MFC-2 and MFC-3. The MFC-4 supplied Ar was used to feed optimum amounts of carrier gas to the ICP in order to achieve maximum intensities of Mo. Figure 3 shows the intensities of 98 Mo measured by the direct introduction of metal carbonyl gas into ICPMS or the introduction through GPD-GED to ICPMS, as a function of the flow rate of metal carbonyl gas to the GPD (MFC-1 + MFC10027

dx.doi.org/10.1021/ac502168h | Anal. Chem. 2014, 86, 10025−10029

Analytical Chemistry

Technical Note

detection limits obtained by GPD-GED-ICPMS were 4−5 orders of magnitude lower than those generally obtained by other analytical techniques such as gas chromatography with an electron capture detector (GC-ECD) or by Fourier transforminfrared spectroscopy (FT-IR) providing limits of detection of a few nL L−1.19−23 The achieved LODs were also 1−2 orders of magnitude lower than those values reported for GC-ICPMS providing limits of detection of a few pL L−1 (a few ng m−3).14−16 The different LODs between the GPD-GED-ICPMS and GCICPMS might be due to different sensitivities of ICPMS instruments. Since the GPD-GED-ICPMS can achieve the direct measurement of metal carbonyl gases as well as the removal of reaction and ambient gases from metal carbonyl gases, the technique opens new capabilities for the online analysis of metallic gas in industrial processes, which includes real-time analysis and monitoring of ultratrace metallic gases in ambient air or other gases such as H2 and CO2.

Figure 4. Ratios of 98 Mo observed under different reaction conditions within the GPD as a function of the gas flow rate of Mo(CO)6. The ratios were obtained by the signal intensities generated by GPD-GED divided by the signal intensities measured using direct sample introduction.

■

better efficiencies than the higher, the length of the reaction chamber should be further optimized. Calibration Curve and Detection Capability. In order to determine the figures of merit for metal carbonyl gas detection, either the linearity of the calibration curve or the detection limit (LOD) were evaluated. Figure 5 contains three calibration curves for three Mo isotopes in dependence on the concentrations of Mo(CO)6 gas. The linear regression coefficients (r2) obtained were better than 0.9999. The calibration curves for Cr(CO)6 and W(CO)6 showed similar linear regression coefficients (r2) of more than 0.9998. The better regression coefficients on Mo(CO)6 was attributed to the specific optimization of ICPMS for Mo. Though the linearity of the calibration curve displayed in Figure 5 was about 2−3 orders of magnitude, the linearity of 5−6 orders of magnitude (up to 100 nL L−1) could be expected because 100 nL L−1 of Mo(CO)6 corresponded to ∼0.0007 mg L−1 of MoO3 was sufficiently low with respect to NH4NO3 particles generated (roughly estimated as ∼0.9 mg L−1) if 1% O3 introduced into the GPD showed 100% reaction to form the particle. The LOD for Cr(CO)6, Mo(CO)6, and W(CO)6 estimated were based on the equation of “3σ/slope” which was derived from 3 times the standard deviation of the background (3σ) and the slope (cps/[pL L−1 or ng m−3]) of the regression curve and were 0.15 pL L−1 (0.32 ng m−3), 0.02 pL L−1 (0.07 ng m−3), and 0.01 pL L−1 (0.07 ng m−3), respectively. The

CONCLUSION

A novel gas to particle conversion-gas exchange technique for the direct analysis of metallic gas in ambient air by ICPMS is proposed and its reaction efficiency and detection capability were demonstrated using metal carbonyl gases, such as Cr(CO)6, Mo(CO)6, and W(CO)6. A reaction mechanism is proposed and an optimization strategy was developed to gain insights into the reaction and detection separately. The limits of detection were 4−5 orders of magnitude lower and similar or slightly better compared to those of conventional analytical techniques such as GC-ECD as well as FT-IR and GC-ICPMS, respectively. The results indicate that the ambient sampling for direct measurements by ICPMS is one of the powerful techniques for metallic gas. The technique is considered to be applied for other metallic gas such as Fe(CO)6, Ni(CO)6, Hg, and AsH3. Since ICPMS is not a handy instrument, the direct measurement of environment air from distant place is limited. The sampling method using a Tedlar bag is considered to compensate the limitation. Though several limitations are pointed out, GPD-GED-ICPMS will allow real-time analysis and monitoring of ultratrace metallic gases in ambient air as well as other gases with respect to quality managements on working and living environments, semiconductor industry, nuclear energy, engine exhaust measurements, and many others.

Figure 5. Calibration curves observed for different isotopes of Mo(CO)6 by GPD-GEDICPMS. The flow rates of the sample gas introduction as well as 1% O3 and NH3 gases for GPD were 100 and 50 mL min−1, respectively, and the concentration of the NH3 solution was 4%. 10028

dx.doi.org/10.1021/ac502168h | Anal. Chem. 2014, 86, 10025−10029

Analytical Chemistry

■

Technical Note

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Inductively Coupled Plasma Mass Spectrometry; Montaser, A., Ed.; Wiley-VCH, Inc.: New York, 1998. (2) Inductively Coupled Plasma Mass Spectrometry Handbook; Nelms, S. M., Ed.; Blackwell Publishing Ltd.: Oxford, U.K., 2005. (3) Utani K.; Nishiguchi K. Gas exchange method and gas exchange device as well as analytical system. Japan Patent Kokai 2006-170659, June 29, 2006. (4) Utani K.; Nishiguchi K. Sample introduction system. Japan Patent Kokai 4462575, February 26, 2010. (5) Nishiguchi, K.; Utani, K.; Fujimori, E. J. Anal. At. Spectrom. 2008, 23, 1125−1129. (6) Suzuki, Y.; Sato, H.; Hikida, S.; Nishiguchi, K.; Furuta, N. J. Anal. At. Spectrom. 2010, 25, 947−949. (7) Kovacs, R.; Nishiguchi, K.; Utani, K.; Günther, D. J. Anal. At. Spectrom. 2010, 25, 142−147. (8) Suzuki, Y.; Sato, H.; Hiyoshi, K.; Furuta, N. Spectrochim. Acta, Part B 2012, 76 (10), 133−139. (9) Tabersky, D.; Nishiguchi, K.; Utani, K.; Ohata, M.; Dietiker, R.; Fricker, M. B.; Maddalena, I. M. D; Koch, J.; Günther, D. J. Anal. At. Spectrom. 2013, 28 (6), 831−842. (10) Dorta, L.; Kovacs, R.; Koch, J.; Nishiguchi, K.; Utani, K.; Günther, D. J. Anal. At. Spectrom. 2013, 28 (9), 1513−1521. (11) Ohata, M.; Nishiguchi, K.; Utani, K. Bunseki Kagaku 2013, 62 (9), 785−791. (12) Utani K.; Yokoyama K. Standard gas for trace element analysis in gas and analytical method using the standard gas. Japan Patent Kokai 8-5524, January 12, 1996. (13) Utani K.; Yokoyama K. Standard gas for trace amount of metal analysis, preparation device and procedure for the standard gas. Japan Patent Kokai 9-61314, March 7, 1997. (14) Feldmann, J. J. Anal. At. Spectrom. 1997, 12, 1069−1076. (15) Pecheyran, C.; Quetel, C. R.; Lecuyer, F. M. M.; Donard, O. F. X. Anal. Chem. 1998, 70, 2639−2645. (16) Feldmann, J. J. Environ. Monit. 1999, 1, 33−37. (17) Izaki R.; Nakamura H. Analytical method and device for metal carbonyl compounds. Japan Patent Kokai 2003-66019, March 5, 2003. (18) Utani K.; Nishiguchi K. Measurement procedure and measurement system for target gas. Japan Patent Kokai 2013-205241, October 7, 2013. (19) Basic Gas Chromatography, 2nd ed.; McNair, H., Miller, J., Eds.; John Wiley & Sons, Inc.: Hoboken, NJ, 2009. (20) Kitamaki, Y. AIST Bull. Metrol. 2006, 4 (3), 227−236. (21) Brody, S. S.; Chaney, J. E. J. Gas Chromatogr. 1966, 4, 42−46. (22) Atar, E.; Cheskis, S.; Amirav, A. Anal. Chem. 1991, 63, 2061− 2064. (23) Shimosaka, T.; Kato, K. Bunseki Kagaku 2013, 62 (1), 11−17.

10029

dx.doi.org/10.1021/ac502168h | Anal. Chem. 2014, 86, 10025−10029