Higher Resolution Mass Analysis in Inductively Coupled Plasma–Mass Spectrometry

A higher resolving power in ICP-MS is used primarily to remove interferences from the measured ion signal, in contrast to the usage of higher resolving power in organic and biological mass spectrometry to establish ion empirical formulas. A practical attainable resolving power of 10,000 suffices for most such purposes in ICP-MS, and is achieved in commercial instruments through proven instrumental means. The ability to "dial in" resolving power as needed in newer trap mass analyzers shows promise for speciation analysis in ICP-MS with chromatographic separation of sample components.

Over the past 10 years, the practical attainable mass resolution in organic and biomolecular mass spectrometry (MS) has increased by approximately a factor of 10. Many instruments continue to operate with essentially unit mass resolution at ion masses of a few thousand. Other instruments can reach a resolution of several tens of thousands in that same mass range, and benchmark experiments showed that resolutions exceeding 1,000,000 could be reached (although perhaps only at lower ion masses, and clean and simple samples). What counts most is what mass resolution can be achieved on a daily basis. The term practical attainable resolution is that level of performance that can be reached without extraordinary measures and without disrupting the workflow in the laboratory. Ideally, higher resolution could be dialed into an analysis as the situation requires, in an analogy to the focus on an optical camera. However, the classical processes for achieving higher resolution in sector-based instruments involve controlling the direction and energy dispersion of a beam of ions emitted from the source, often with the assistance of physical limiters such as ion slits. The physical reconfiguration of the instrument, and confirmation of the desired consequential effect, are procedures that take several minutes even for experienced operators. Peak matching, which involves the creation of an overlap of peak signals corresponding to ions of two different masses on a scope, and measurement of a ratio of accelerating voltages needed to do so, takes even more time and effort. Practical attainable resolution often is below peak instrument performance.

Kenneth L. Busch

The recent increase in attainable resolution has been achieved with the use of Fourier transform–ion cyclotron resonance (FT-ICR) mass spectrometers, and even more recently, orbitrap instruments. In both of these instruments, the mass resolution is established by the length of the observation period for the frequency signal generated by an orbiting cohort of ions trapped within the instrument. The observation period for even very high mass resolution is still only on the order of a few seconds, which is a short enough period that higher resolution mass measurement can be achieved for samples derived from a chromatographic separation. Indeed, with a high enough mass resolution, some of the need for chromatographic separation can be relaxed, and impressive results can be obtained directly for mixtures introduced into the source of the mass spectrometer.

Before we begin a discussion of why higher "mass resolution" is valuable, we should review basic definitions to buttress our understanding of the difference in the terms resolution and resolving power. It is unrealistic to expect that this repetition of the definitions will alter the extended history of casual misuse of these terms, and their not-quite-correct usage in the preceding paragraphs. However, clear and concise language is an integral part of accurate and precise analysis. Resolving power is a measure of the ability of an instrument to differentiate between ions of similar masses. Resolving power is the large number often misidentified as "resolution." Resolving power is defined as m/Δm, where m is the mass of the lower mass ion, and Δm is the mass difference between m and the next higher mass ion from which differentiation is sought. Figure 1 is reproduced from Sparkman's Mass Spec Desk Reference, which provides a more thorough discussion of the use and misuse of these two terms (1). The actual numerical value of resolving power depends upon the degree of differentiation sought, and can be specified at a 10% valley, or at full width half maximum (FWHM). The situation represented in the figure is a theoretical ideal in which two symmetrical peaks are of equal height. The real world usually is not so tidy, but both unequal peak heights and asymmetry can be accommodated within mathematical models to provide a meaningful value for resolving power. Note that resolving power is a unitless number. Resolution is the inverse of resolving power (Δm/m), and is a measure of the observed separation between two adjacent mass peaks. The description of the ion peak separation as resolution comes with units. For example, a resolution can be reported as some parts per million for ions of a specified mass. Resolution can be a small number if the separation of the ion masses is "just barely," or it can be a larger value if the instrumental resolving power is high. Sparkman notes that the MS usage of "resolving power" and "resolution" differs from IUPAC recommendations. The issue eventually will need to be revisited officially as more analyses at higher resolving powers are reported, and the need to compare results grows. Practically, in refereed publications, and even in the title of this article, the tendency is to use "resolution" in place of "resolving power," reflecting the preponderance of the term high resolution in other venues. In the remainder of this column installment, we will strive to use the terms correctly.

Figure 1: Schematic of peaks representing ions of two adjacent masses, and the values used for derivation of the resolving power (m/Δm at a specified valley).

For molecular MS, a higher resolving power can provide an accurate mass measurement for an ion through proper calibration. If the empirical formula of the ion is predictable from other information, then it is a necessary condition that the measured mass corresponds to the calculated expected mass. Incrementally higher resolving power provides greater analytical certainty in the confirmation of the ion identity. Conversely, if the ion mass is known with high accuracy, the empirical formula of an ion can be determined from that value alone for lower masses, or the number of possible combinations of atoms can be narrowed to a few for higher masses. A balance must be struck between the mass measurement accuracy required and the loss of sensitivity that usually accompanies adjustment of the instrument for higher resolving power. Transience of ion signals also can preclude anything but a lower resolving power measurement. The number of possible atomic compositions (empirical formula) for any measured mass increases rapidly as the mass of the ion increases. The accurate mass measurement is therefore most often used for confirmation rather than derivation of an ion empirical formula. It is worth remembering that valuable information can be derived by measurement of the exact mass difference between two ions in a mass spectrum. In this situation, the accurate mass difference establishes the empirical formula of the neutral atomic fragment.

In inductively coupled plasma coupled with mass spectrometry (ICP-MS), the high temperature of the plasma source ensures that molecular species introduced into the source are reduced to simple atomic species, or sometimes diatomic species such as elemental oxides or chlorides. Atomic ion masses are usually distinct, but the isobaric interferences for the diatomic species can be problematic and have been tabulated (2–4). Numerous strategies have been developed to reduce the contribution of these interferences, including modified sample preparation and cleanup, reactive collision cells (5), and spectral peak deconvolution methods (6) that also can be used for other forms of MS. However, the most direct means for reducing isobaric interferences is to increase the resolving power of the mass analyzer. Therefore, in ICP-MS, higher resolving power has a fundamentally different purpose than in organic MS. The purpose is to remove interferences rather than determine empirical formulas. Once isobaric interferences are removed through mass resolution, the isotopic signature of an elemental ion can be measured directly and accurately. This measurement provides direct confirmation of the accuracy and precision of the intensity measurement, and is therefore a basic metric of instrument performance. Additionally, it is only once these accuracy and precision metrics are known that variations in isotopic distributions can be tracked. Such work is distinct from the highly sophisticated analyses completed in isotope ratio MS, which uses dispersed ion beams and separate detectors, and which can track isotopic variations at the part-per-billion level. New applications can be found within the performance envelope lying between unit mass resolution ICP-MS and isotope ratio instruments. For example, higher mass resolving power in ICP-MS will also be needed to provide the requisite accuracy in developing clinical applications (7).

Several manufacturers offer high-resolving power ICP-MS instruments. As an example, the Finnigan Element2 is a high resolving-power (high-resolution) ICP-MS instrument available from Thermo Fisher Scientific (Waltham, Massachusetts). The instrument uses a fast-scanning magnetic sector in a double-focusing BE geometry to achieve a high resolving power. Fast-scanning sector magnets have been used in mass spectrometers used for organic analysis, scanning rapidly over a mass range from about m/z 50 up to about m/z 2500, covering the molecular mass range of interest. For ICP-MS, the mass range of interest is much narrower, corresponding to the mass range of atomic ions. The specifications for the Element2 instrument list a scan speed from m/z 7 to m/z 240 and return in under 150 ms. This scan speed would provide about seven full scans across a 1-s sample bolus, whether it originated in a flow injection analysis or in a chromatographic peak. Although spectral averaging across a narrow peak provides an approximation of the steady-state mass spectrum, in general, the more scans across a peak, the better. The goal is to provide a scan speed sufficiently fast for an accurate measurement of mass spectra as samples are eluted into the ICP source from a chromatographic inlet such as gas chromatography or liquid chromatography, as separation of species via chromatography is used with ICP-MS. It is interesting to remember that early generation sector magnets used in organic mass spectrometers could not scan rapidly, and accelerating voltage scanning could be used across a smaller mass range with a faster effective scan speed. The company literature for the Element2 states that the instrument uses a combination of accelerating voltage and magnetic field strength scanning in its routine scan function.

The topic of interest for this column is the higher resolving power available with the sector instrument for ICP-MS. As in sector instruments for organic MS, a higher resolving power is achieved in the Element2 with a physical slit that narrows the dispersion of the ion beam from the source. The Element2 uses a three-position slit, as shown in Figure 2 (figure adapted directly from the company literature) to provide specified resolving powers of 300, 4000, or 10,000 (described as low, medium, or high resolution). The ion mass for which these resolving powers is achieved is not noted. The slit for resolving power 300 is the widest, and provides the best instrument sensitivity. This low resolving power should provide at least unit mass resolution across the entire mass range of interest, and would be used for relatively clean samples expected to be free from interferences. The medium resolving power is sufficient to prevent most common interferences, and the resolution of 10,000 is suggested for use with complex samples in which the matrix may be unknown or uncharacterized. The highest resolving power is insurance against an unexpected interfering ion signal. The payment for this insurance is a decreased sensitivity. Sensitivity is about an order of magnitude less at resolution 4000 compared with resolving power 300, and almost an order of magnitude lower again at the highest resolving power of 10,000.

Figure 2: Source slit assembly used in a commercial high resolving-power ICP-MS instrument (Finnigan Element2, as described in the company literature).

With freedom from isobaric interferences, and the ability to certify instrument performance through observation of the correct elemental isotopic envelope, the confidence limits (error bars) on a set of replicate elemental quantitation values can be determined. Then, elemental distributions in samples can be ascertained reliably, and the distributions of all elements can be queried. This is a complex multidimensional data set that can be used in interesting new applications. For example, Norris and colleagues (8) used the trace elemental profiles of feathers from the Western sandpiper to distinguish bird populations that inhabit different winter sites. High resolving-power ICP-MS was used to determine the concentrations of 41 trace elements in the feathers sampled from known discrete populations of these birds. Statistical analyses using linear discriminant analysis identified 15 elements for which the measured concentrations evidence one of the five distinct wintering sites. Those five sites are indicated in Figure 3a. Three sites are quite a distance from one another; the last two sites are distant from the other three but only 3 km from each other. Still, even these two sites could be distinguished based upon the elemental distributions. The distributions for six elements at each of the five sites is depicted in Figure 3b. Note that the differences between the sites exceed the confidence limits established for each measured value. Without the confidence limits, the comparisons would lose meaning. This analytical work complements the genetic information that can be indicative of the differences in breeding populations of the birds. The ICP-MS analysis can be automated easily and can handle large numbers of easily collected samples; additional applications in similar areas can be expected to appear in the future. Compilations of high resolving-power ICP-MS publications, talks, and posters through 2004 are available on the web (9).

Figure 3: (a) Map of the winter-resident locations of five separate populations of Western sandpiper; these grounds are designated by letter abbreviations S, L, B, R, and A. (b) Plots of the concentrations of six elements in the feathers of these populations, showing that the populations can be distinguished.

In organic MS, if moderate resolving power is good, then higher resolving power is thought to be still better, as long as the loss in sensitivity is not too severe. The situation is somewhat different in ICP-MS, where resolving power sufficient to remove the isobaric interferences is all that is required. The ICP itself acts to reduce other potential interferences to a well-known population such that a resolving power of 10,000 is all that is needed. Higher resolving powers in research instruments for ICP-MS have been demonstrated, but more as a demonstration of feasibility. FT-ICR instruments and orbitrap instruments do offer considerable promise in ICP-MS because of the manner in which higher resolving power is achieved — that is, by longer observation of the frequency signal from an ion population in stable orbit inside the instrument. The Orbitrap in particular represents an instrument in which the requisite resolving power can simply be accessed as the analysis requires it. Consider an experiment in which distributions for a variety of elements are to be determined, as in the example given earlier for the analysis of feathers drawn from migratory bird populations. An exploratory scan at a moderate resolution establishes what elements are present, and whether there is any indication of an isobaric interference based upon deviations from expected isotopic intensities. If there is no indication of interference, a random higher-resolution scan can be programmed into the analysis series to act as a quality control check. If there is an indication of an isobaric interference, the higher-resolution scan is selected automatically. This decision can be made in real time even as samples are introduced over a few seconds, and eventually might be fast enough so that speciation through chromatography, with higher resolving power confirmation of results, is achieved. Applications of hybrid trap–Orbitrap instruments to ICP-MS also can be foreseen, if the need for the analytical results justifies the relatively high initial cost of such instruments. A similar argument that invoked the high cost of MS as a detector for ICP was invoked as the first ICP-MS instruments were constructed. The informing power of the results soon rendered that argument obsolete.

Kenneth L. Busch is often reminded that new significant accomplishments in analytical sciences are usually the result of dedicated problem-solvers diligently searching for solutions, rather than developers of solutions waiting patiently for problems to arrive at their laboratory door. An "open-door" policy is best matched to an open mind and a fearless willingness to explore. Knock on KLB's electronic door at wyvernassoc@yahoo.com

References

(1) O.D. Sparkman, Mass Spec Desk Reference, Second Edition (Global View Publishing, Pittsburgh, Pennsylvania, 2006).

(2) S.H. Tan and G. Horlick, Appl. Spectrosc. 40(4), 445–460 (1986).

(3) S. Nelms, Ed., Inductively Coupled Plasma Mass Spectrometry Handbook (Blackwell Publishing, Ames, Iowa, 2005).

(4) M. Plantz, "Common Molecular Ion Interferences in ICP-MS," found at https://www.varianinc.com/media/sci/apps/icpms06.pdf.

(5) D.R. Bandura, V.I. Baranov, and S.D. Tanner, Fres. J. Anal. Chem. 370(5), 454–470 (2001).

(6) J. Mejda and J.A. Caruso, J. Am. Soc. Mass Spectrom. 15, 654–658 (2004).

(7) E.E.M. Brouwers, M. Tibben, H. Rosing, J.H.M. Schellens, and J.H. Beijnen, Mass Spectrom. Rev. 27, 67–100 (2008).

(8) D.R. Norris, D.B. Lank, J. Pither, D. Chipley, R.C. Ydenberg, and T.K. Kyser, Can. J. Zool. 85, 579–583 (2007).

(9) See http://www.hricpms.com/pdfs/2004Jan_journals.pdf and http://www.hricpms.com/pdfs/MCICPMSTalksPosters.pdf.