A strategy to use the collision cell as a method development tool for isotope selection in the selected ion monitoring (SIM) operational mode for the quadrupole mass filter.
Since the mid-1980s, ICP-MS instrumentation has been a rapid, sensitive, and robust platform for the elemental analysis of environmental water systems in both qualitative and quantitative determinations. Recent advances have included the use of collision cells inserted into the instrument geometry to increase the selectivity of the measurement by reducing polyatomic isobaric interferences via competing mechanisms described by the first principles of inelastic collisions. This study demonstrates a fundamental strategy to use the collision cell as a method development tool for isotope selection in the selected ion monitoring (SIM) operational mode for the quadrupole mass filter. The isotopic envelope of selenium (Se) was selected for this evaluation.
Inductively coupled plasma–mass spectrometry (ICP-MS) has proven utility in environmental monitoring with the advent of recent advances in commercial instrumentation (1,2). ICP-MS has demonstrated advantages over competing elemental analysis techniques, including better method detection limits, isotopic identification, and low level detection limits as an artifact of the unit resolution operation of modern quadrupole mass analyzers (3,4).
The ICP-MS instrument geometry consists of a region at atmospheric pressure where a sample is continuously introduced via a peristaltic pump with a flow rate on the order of microliters per minute to a nebulizer that creates a fine mist of aerosol droplets amenable to ionization by argon plasma (5). The nebulizer is housed in a Peltier-cooled spray chamber that is chilled to ~2 °C to minimize the formation of oxides. The driving force for the ionization mechanism is the presence of an induced voltage supplied by metallic radio frequency (RF) coils bound to a quartz torch. The ions are entrained in an RF field emitting from an orifice and pulled into a region under vacuum. Commercial design solutions for this step typically use an arduous off-axis path to separate neutral components of the mixture as a means of improving the signal-to-noise ratio (S/N). The vacuum regime is sequential in instrument segments, stepping down from atmospheric pressure (ionization region), to 10-3 Torr (focusing region), and to 10-5 Torr (analyzing region). It is generally accepted that the analyzing region is optimized to favor a mean free path in which ions are focused and accelerated in the z direction of the Cartesian plane when conceptually visualizing the hyperbolic field focusing the ion path through the quadrupole rods in terms of xyz coordinates (6).
This article focuses on the optimization of parameters related to the analyzing region under vacuum. More specifically, parameters related to the separation of polyatomic isobaric interferences that may systematically bias dissolved elemental measurements related to matrix effects of complex environmental water systems. For this reason, digestion procedures and related sample preparation methods will not be discussed. The readers of this work are referred to US Environmental Protention Agency (EPA) Methods 200.8, 1638, 3015, 3050, and 3051 for strategies related to sample preparation. The separation event of interest is that of the gas-phase collision events that occur on-line as a precursor to mass analysis. Because chemical noise is mass selective — and always additive, suggesting falsely high measured determinations (7–9) —strategic isotope selection is critical in minimizing systematic measurement bias.
Selenium (Se) is an analyte of interest for ICP-MS applications related to environmental monitoring. Se naturally occurs in low concentrations and generally is associated with sulfur-containing ores such as pyrite, sphalerite, and chalcopyrite (10). Naturally occurring mineral dissolution, selenium processing, and the use of selenium are the most common routes of liberating it into the aqueous environment as an oxyanion. Oxyanions are negatively charged inorganic compounds with various degrees of oxidation. Examples of oxyanions are COx, SOx, NOx, POx, ClO2, AsOx, SeOx, and so forth. These variations in oxidation effect not only mobility in aqueous systems, but may also effect the level of toxicity. Selenium, for example, possesses a low toxicity in the complexed form; but becomes a significant ecological concern once oxidized to the selenite (Se[IV]) and selenate (Se[VI]) forms, which are also water-soluble. In small quantities, 0.1–0.5 ppm dry weight, selenium is considered a necessary part of life. It is not until levels reach >3 ppm, by dry weight, that it becomes a toxicity concern; however, the bioaccumulative effect has led to the reduction in the EPA's acceptable dischargeable selenium criteria. A 5 ppb limit on total selenium was recommended for fresh water systems in the National Fresh Water Quality Standard, while 50 ppb total selenium is the National Primary Drinking Water Standard (11,12). Industries that are affected by the reduced discharge criteria include minerals and metals mining, oil refining, coal-fired power generation, sulfuric acid manufacturing, and others.
The focus of this article is the optimization of the collision cell before mass analysis; however, it is important to understand that the ionization source plays a role in bond breaking, which is different than the fundamental operation of other atmospheric pressure ionization sources (for example, electrospray ionization  or atmospheric pressure chemical ionization ). As discussed in the previous section, the element Se exists as an oxyanion in solution commonly as selenate (SeO4) or selenite (SeO3). Selenate is more common in environmental waters because of the pH and reducing–oxidizing conditions of ambient water systems.
The driving force for atomization is the extreme temperature of the source and subsequent ionization via charge transfer of the plasma (14):
[SeO4]4- → mixed isotopes Se → mixed isotopes Se+
The argon plasma results when argon flow meets the RF coils and generates an electronic cascade amenable to transferring charge to elemental species in a mixture from a field sample:
Ar → Ar+ + e-
The driving force of the ionization process is the electronic cascade inherent in the argon plasma. It follows that the reaction is not as simple as described above. The presence of ozone is also possible, as well as side reactions with atmospheric contaminants and other ambient analytes.
The optimization of the collision cell is a means to minimize isobaric interferences associated with chemical noise; the instrument ionization parameters are tuned to minimize oxides so the collision cell is not the means of converting selenate and selenite to the selenide ion.
At a basic level, collision cells are an on-line tool for chemical separations. In some applications, it has previously been reported that collision cells have minimized the need for sample preparation and in some cases the need for chromatographic separation before MS analysis (15). In "tandem-in-space" mass spectrometer instrument configurations (where no ion storage is needed), a continuous ion beam containing the sample mixture is transmitted over a fixed distance and subjected to physical collisions or chemical reactions (as ion–molecule interactions) in an RF-only multipole device. When the gas introduced is helium (He), the mechanisms induced in the sample mixture that degrade the ion mixture are collision induced dissociation or collisionally activated dissociation; kinetic energy discrimination due to collisional cooling; and scatter in transmission along the z-axis.
In this experiment, collision cooling is of particular interest and is commonly used to promote chemical separations by kinetic energy discrimination in a mixture. A narrow spread of kinetic energies for a given mass range is generated from the mixture of an ion beam post ion source creating a viable milieu for isobaric interferences to affect the measurement. Polyatomic ions typically have higher cross-sectional areas than the isobaric elemental counterparts. It follows that polyatomic ions encounter more collisions over a fixed distance in the collision cell. The increased number of collisions results in a decreased kinetic energy. The collision cell exit potential can be optimized to discriminate or reject the lower kinetic energy polyatomic ions. A recent study by Sakata and colleagues demonstrated this technique in application (17).
The quadrupole mass analyzer can be designated to monitor a specific isotope of the element of interest using selected ion monitoring (SIM) mode. The fundamentals of SIM have been described elsewhere (18), but briefly, the applied RF and DC voltages generate the mass-selective hyperbolic field to filter an ion population that correlates to a selective mass-to-charge ratio (m/z). The RF and DC voltages operate at a fixed ratio. To capture the isotopic envelope of an element, the power supply can be set to oscillate with RF and DC ratios selective for the multiple m/z filtering events. The isotopic envelope for Se is quite complex, as described in Table I: 74Se, 76Se, 77Se, 78Se, 80Se, and 82Se. The recommendation from several EPA documents related to analytical measurements by ICP-MS steers the analyst in the direction of the isotope at m/z 82. This is interesting as it has been well documented that interferences from 82Kr are common in Ar Dewars. Correction equations are often applied to mitigate this issue. It should also be noted that the EPA-recommended isotope, 80Se, is not the most abundant in nature. That isotope is plagued with its own mass-selective contaminants that complicate direct measurement strategies.
Table I: The isotopes of selenium related to decision making in ICP-MS method development
An ICP-MS system equipped with a helium collision cell was configured to simultaneously execute multiple stages of the SIM operational mode corresponding to selected isotopes of Se. Each standard was measured with no-gas mode and gradient conditions of collision cell pressure to optimize the kinetic energy discrimination mechanism of separation. The candidate isotopes were evaluated based on measured concentrations of Se due to signal contribution from known interferences and measured values for a parts-per-billion Se standard solution prepared with a relatively higher concentration of known interferences.
Reagents and Standards
Selenium standards were obtained as certified reference material from external vendors. The selenium standard used for the calibration curve was Spex CertiPrep CLSE 2-2y 1000 mg/L prepared in 2% nitric acid. A synthetic Se interference solution was prepared to mimic the known interferences for Se described in Table I.
The ICP-MS system was an Agilent 7700x equipped with a helium collision cell to remove polyatomic isobaric interferences (Figure 1). Samples were introduced to the nebulizer spray chamber by an ASX-500 series autosampler (Cetac Technologies). Software settings for instrument configuration were as follows: RF power, 1550; RF matching, 1.80; sampling depth, 8.0; carrier gas, 0.45; nebulizer pump, 0.10; S/C temp, 2; dilution gas, 0.65; extract1, 0; extract2, -160; omega bias, -90; omega lens, 8.0; cell entrance, -30; cell exit, -60; deflect, 0.0; plate bias, -60; collision cell He, 3.0; oct bias, -18; oct RF, 180; energy discrimination, 4. The CE entrance was -50, and the CE exit was -70. The He gas flow was varied.
Figure 1: Surface plot of measured isotopic abundance normalized to theoretical isotopic abundance for Se versus variable collision pressure.
The quantification of isotope specific determinations of Se was achieved using an internal standard calibration model. The internal standard was Y monitored at m/z 89. The analyte signal was normalized to the measured signal of an internal standard within injection. The internal standard was continuously introduced and mixed with the standard solutions on-line to minimize error associated with sample manipulation and preparation. The prepared internal standard concentration was 100 ppb. It has previously been reported that the selection of an appropriate internal standard is more dependent on m/z vs. ionization energy (19).
The instrument performance was verified using Agilent tuning solution diluted to 10 ppb before measurement. The startup internal diagnostic procedure verified the daily performance of the plasma source (torch axis setting and plasma correction), the collision cell, the quadrupole mass analyzer, and the electron multiplier. The quadrupole mass analyzer had a mass range of 2–260 m/z. The daily performance was verified by comparing the peak width at 50% of the peak height vs peak width at 10% for m/z 7, 89, and 205. The extended linear dynamic range of the pulse–analog (PA) electron multiplier detector was verified by a PA standard solution. The standard solution was diluted 1:100 before measurement.
Qualitative Analysis: The Concentration Dependence of the Capture of the Isotopic Envelope for Se
The isotopic envelope of an element encompasses the qualitative identification of the ion by m/z and the semiquantitative relative abundance obtained as a measure of detector response. In this study, the isotopic envelope was used to predict the optimal isotope selective collision cell parameters by comparing measured relative response with the theoretical relative abundance (Table II). The capture of the theoretical value suggests a measurement free of mass-selective interferences. A graphic representation of the data was prepared by normalizing the relative error to the theoretical value with a value of 1.0 translating to a 1:1 ratio of measured abundance:theoretical abundance (Figure 1). The working range for Se isotopes generally is 3–6 mL/min of He gas. Recovered percentages may be explained as a function of attenuation of the ion transmission because of the collisional cooling in the collision cell. The isotope at m/z 74 is relatively low in abundance and may favor a mechanism of scatter in the presence of the inelastic collision events. The concentration of the standard solution was 1 ppm Se prepared in 2% nitric acid.
Table II: Qualitative determination of the optimal collision cell pressure for isotope-specific determination of Se
Quantitative Analysis, Part I: Evaluation of the Background in Synthetic Laboratory Blanks
A synthetic interference solution was prepared with 2% nitric acid diluent to the following concentrations of interference analytes described in Table I: 280 ppm Br- (28,000:1 interference to analyte), 280 ppm SO3- (28,000:1 interference to analyte), 400 ppm Ca2+ (40,000:1 interference to analyte), and 400 ppm Cl- (40,000:1 interference to analyte). The synthetic interference solution was measured under gradient conditions of the collision cell pressure indirectly monitored in milliliters per minute. The results are detailed in Figure 2. The measurements were performed by sampling six replicates of the same vial under each of the instrument conditions. It was interesting to observe the improvement in background counts for all isotopes of Se. The isotope corresponding to m/z 80 showed the most significant contribution to background in the presence of interferences. Under instrument conditions of relatively low collision cell pressure, m/z 80 would make the worst possible candidate for SIM (Figure 2a). Figures 2b and 2c sequentially exclude the ion most susceptible to interferences, leaving m/z 77, 78, and 82 as candidate ions over specific ranges of operational parameters. Before measurement, we can predict these three candidate ions will provide viable quantitative data as the environment is most selective versus mass-selective interferences.
Figure 2: Monitoring the isotopic envelope response for Se in a blank synthetic interference diluent (a) all isotopes (b) graphed without m/z 80 (c) graphed without m/z 80 and 76.
Quantitative Analysis, Part II: Evaluation of the Recovery of a Selenium Standard Prepared in a Synthetic Interference Solution
Calibration curves and recovery of fortified standards: A total of 48 calibration curves (six isotopes × eight cell pressure settings) was generated for each isotope with corresponding instrument conditions described above. Solution standards of Se were diluted serially from a vendor-prepared 10,000 ppb stock solution. Solution standards were prepared to final concentrations of 1, 10, 100, and 1000 ppb in 2% nitric acid. The synthetic interference solution was used as diluent for a selenium standard to evaluate the collision cell variables applied to the instrument configuration. The Se concentration of the sample solution was prepared to a final contribution of 10 ppb by serial dilution of the 1000 ppb calibrator into the synthetic interference solution as diluent. Table III details the observed background-corrected counts for the isotopic envelope of Se illustrating the inverse relationship of sensitivity and selectivity (that is, more sensitivity may mean more interferences). Table IV illustrates the quantitative result of decision making related to isotope selection and collision cell parameters.
Table III: Background-corrected counts for a 10 ppb Se standard prepared in the synthetic interference diluent
To further improve the optimization, a second experiment was performed that targeted m/z 77 and 78. Because the natural abundance of Se for m/z 78 is higher than 77, it was anticipated that 78 would outperform 77 at lower concentration levels. The collision cell pressure corresponding to a flow rate of 4 mL/min (the middle of the 3–5 mL/min working range illustrated in Table II) was selected. The integration times were increased for this experiment. The analyte was held at 5.01 s, and the internal standard was 0.501 s. The same calibrators and fortified standards were used. No significant difference was observed in the final determination of Se for either isotope. A similar investigation was performed that focused on the salient parameters that affect the energy of the ion beam at the entrance and exit of the collision cell as well as the voltages that facilitate bias in kinetic energy discrimination. Again, no significant difference was observed in the final determination of Se for either isotope. Further evaluation should include a concentration dependence study of the interfering ions on the collision cell performance.
Table IV: Simultaneous determination of a 10 ppb standard prepared in a synthetic interference solution as diluent under variable collision cell pressure
The application of collision cooling results in signal attenuation of the analyte ion transmission (7). Strategies to mitigate this phenomenon relate to signal enhancement before the collision cell. It has previously been reported that the addition of low percentages of methanol (v/v) has improved the charge transfer process related to the atmospheric pressure ionization mechanism. It is anticipated that the selection of an isotope with lower natural abundance with higher rates of ionization efficiency would provide lower detection limits than an isotope with a higher natural abundance in the presence of chemical noise. In terms of S/N, the detection limits could be improved by increasing S. Toward improved measurement accuracy at lower levels, the next step of this optimization would involve detailed method development studies focused on the atmospheric pressure variables in the analytical process (for example, optimal dilution in the formation of aerosols to increase ionization efficiency, candidate formulations of mixed plasmas to minimize artifacts that induce plasma cooling).
Method development decisions related to targeted elemental analysis in complex mixtures begins with the isotope selection to minimize mass-selective interferences. This decision making cannot be achieved by intuition alone, as the most abundant isotope occurring in nature may not be the choice that leads to the method with the lowest detection limit associated with the element of interest. The presence of polyatomic isobaric interferences as chemical noise requires additional resolution elements in the analytical process. It follows that on-line parameters associated with recent advances in commercial ICP-MS instrumentation must be evaluated in accordance with first principles. It was determined that the most abundant element was not the best choice for a selective, robust analytical method. It should also be noted that there was more than one correct combination of parameters that can provide a similar conclusion.
The authors of this work gratefully acknowledge the work of Olaronke Olubajo for technical support related to MassHunter software and instrument diagnostics.
(1) S. D'Ilio, N. Violante, C. Majorani, and F. Petrucci, Anal. Chim. Acta 698(1-2), 6–13 (2011).
(2) S.D. Tanner, V. Baranov, and D. Bandura, Spectrochimica Acta Part B 57, 1361–1452 (2002).
(3) R. Thomas, Practical Guide to ICP-MS: A Tutorial for Beginners (CRC Press, 2008).
(4) F. McLafferty, Tandem Mass Spectrometry (John Wiley & Sons, 1983).
(5) G. Holland and D. Bandura, Plasma Source Mass Spectrometry: Current Trends and Future (RSC, 2005).
(6) E. Hoffmann and V. Stroobant, Mass Spectrometry: Principles and Applications (John Wiley & Sons, 2007).
(7) K. Busch, Spectroscopy 17(10), 32–37 (2002).
(8) K. Busch, Spectroscopy 18(2), 56–62 (2003).
(9) K. Busch, Spectroscopy 18(5), 52–55 (2003).
(10) D.J. Adams and P. Pennington, Proceedings of the SME Annual Meeting, Denver, Colorado, Preprint 05-53 (2005).
(11) EPA, National Recommended Water Quality Criteria, 2011. http://water.epa.gov/scitech/swguidance/standards/current/index.cfm (accessed May 2011).
(12) EPA, 2001, Selenium Treatment/Removal Alternatives Demonstration Project; Mine Waste Technology Program Activity III, Project 20; MSE Technology Applications, Inc.: Butte, MT. EPA/600/R-01/077.
(13) R. Cole, Electrospray Ionization Mass Spectrometry, Fundamentals, Instrumentation and Applications (John Wiley, 1997).
(14) J. Watson, Introduction to Mass Spectrometry, Third Edition (Lippincott-Raven Publishers, 1997).
(15) R. Yost and C. Enke, Anal. Chem. 51, 1251A–1264A (1979).
(16) G.D. Rayson and G.M. Hieftje, Spectrochimica Acta Part B: Atomic Spectroscopy 41(7), 683–697 (1986).
(17) N. Yamada, J. Takahashi, and K. Sakata, J. Anal. At. Spectrom. 17, 1213–1222 (2002).
(18) F. Kero, R. Yost, and R. Pedder, "Quadrupole Mass Analyzers: Theoretical and Practical Considerations," Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics (John Wiley & Sons, 2005).
(19) F. Vanhaecke, H. Vanhoe, R. Dams, and C. Vandecasteele, Talanta, 39(7), 737–742 (1992).
(20) M. Shah, S. Kannamkumarath, J.C.A. Wuilloud, R. Wuilloud, and J.A. Caruso, Anal. At. Spectrom. 19, 381–386 (2004).
Frank A. Kero, Lucas R. Moore, and Jeff Malson are with Kemira, Atlanta Analytical Services in Atlanta, Georgia. Please direct correspondence to: firstname.lastname@example.org.