Contemporary Trends in Ion Chromatography

Photo Credit: Melamory/Shutterstock.com

Ion chromatography coupled to mass spectrometry (IC-MS) does not follow the same rules as coupling other modes of liquid chromatography to mass spectrometry (LC–MS). Leon Barron from the Analytical & Environmental Science Division, at King’s College London, UK, reveals some of the misconceptions surrounding IC and discusses contemporary trends and novel applications for this inventive technique. - Interview by Kate Mosford

Q. You did your PhD using ion chromatography coupled to mass spectrometry (IC-MS). Are there any misconceptions about coupling IC with MS?

A: When I started working with IC-MS in 2001, there were only a few published applications (1–4), but thankfully a more sizeable body of knowledge now exists that has enabled IC-MS to be used more routinely. However, the main misconception many non-experienced users have is that IC-MS follows the same rules as coupling other modes of liquid chromatography (LC) to MS, such as reversed-phase LC.

There are also several different modes of IC, all of which carry their own considerations for coupling to MS detectors. Firstly, reversed-phase separations normally use organic solvents in the eluent to elute less polar solutes. This also helps the desolvation process for electrospray ionization, for example, and generally makes coupling relatively more straightforward. On the other hand, IC is often performed under aqueous conditions and elution is achieved by exchange with an eluting ion species. Injected sample ions are therefore separated based on their size and charge. As the retention mechanism is different, IC does not generally require an organic solvent in the eluent. The challenge for IC-MS, and especially for trace analysis, is achieving acceptable sensitivity using nonvolatile IC eluates containing a range of inorganic and organic ions. This is generally overcome in a number of ways, but the most prominent approach is by post-IC infusion of organic solvent prior to an electrospray source. This obviously adds cost and complexity to the configuration for routine operation.

Secondly, sensitivity is often affected by the presence of relatively high concentrations of the eluent ion species. Suppressors can be used, in a similar manner to how they are used with conductivity detection, to overcome this problem for ion exchange chromatography and especially when using hydroxide-based eluents where they can be converted to water. Electrospray sources can add more back pressure compared to conductivity cells, and problems can arise with the stability of membrane-based suppressors over lengthy periods of time, but the designs are becoming more and more robust. Packed-bed suppressors are an alternative and are more pressure resistant, but need periodic regeneration. Multiple switchable packed-bed suppressors are now available that can now be regenerated automatically on-line. The trend for suppressed ion exchange chromatography (as the dominant mode of IC) coupled to MS is the requirement for three pumps: one for the analytical flow; one for suppressor regeneration; and one for post-IC organic solvent modification. This is perhaps more complex than most users anticipate when first considering IC-MS!

Other options for IC-MS do exist, which all have their pros and cons. For example, coupling IC to inductively coupled plasma mass spectrometry (ICP-MS) is more straightforward, but is a harder ionization technique than electrospray ionization coupled to MS (ESI-MS). This makes it more suited to elemental analysis and is predominantly used for metals-based applications. Quantitatively, IC-ICP-MS suffers from some interference issues, which need to be carefully minimized. As another alternative mode of IC, ion interaction chromatography coupled to MS has been successful and is more akin to the traditional LC–MS configuration because suppressors or auxiliary pumps may not be needed. An eluent that contains the ion interaction reagent can also contain some organic solvent, so desolvation in the ion source can be improved. The problem for this mode of IC is that not all ion interaction reagents are compatible with MS (for example, quaternary ammonium species). Unfortunately, the capacity and selectivity for larger numbers of analytes in complex samples using MS-compatible ion interaction reagents tends to be rather limited using this mode of IC.

Q. Why has there been a recent increase in coupling LC to high‑resolution accurate mass spectrometry (HRAMS) and where is the niche for IC‑HRAMS? What can this technique offer that other techniques cannot?

A: The coupling of many different modes of chromatography to HRAMS has predominantly enabled flexible options for full data-capture analysis in targeted, suspect screening and nontargeted type applications. As mass accuracies of less than 5 ppm are now achieved routinely, there is less of a need to preselect a precursor compound m/z, such as in selected reaction monitoring. Alongside a more traditional targeted approach, the main advantage is that the data can be reviewed, further mined, or compared (holistically or in part) after the analysis has been completed. This can, for example, enable the retrospective identification of new compounds, including their precursors or transformation products. Such methods are often used quantitatively as well as qualitatively, which is excellent. High mass accuracy datasets can also be used in a nontargeted manner to differentiate a chemical signature from that of another sample, but with increased confidence when operated at high resolution (5,6). This means that HRAMS gives some extra flexibility than some other mass analyzers. However, in my experience these MS detectors are not necessarily the most sensitive and triple quadrupoles tend to still offer the best performance, especially for trace quantitative analysis. That said, some of the newer HRAMS instruments include a quadrupole stage that allows selected reaction monitoring to be performed if higher sensitivity and selectivity is needed. For me, it has been a fine balance between the sensitivity you need for a particular application and how much you want to know about your analytes or sample matrix.

For IC-MS in general, a large proportion of applications focus on a small number (<20) and size range of molecules (<250 Da). As with other modes of LC, coupling IC to HRAMS means that more ions can be acquired across the run allowing new compounds to be identified retrospectively. Higher-energy collisional dissociation can also be used in the same run to gather fragment information if needed. This could be quite beneficial for applications where targeted IC-MS is normally used to determine traces of toxic contaminants in drinking water, such as disinfectant byproducts (7). Only a small proportion of disinfectant byproducts have been characterized and the use of HRAMS is likely to help rapidly identify other related contaminants that may be toxic at similarly low concentrations. One other potentially important niche area for IC‑HRAMS in my opinion lies in the “–omics” field and particularly so for metabolomics. In 2011, Burgess and colleagues (8) were the first to show that IC-HRAMS could fill obvious gaps left by hydrophilic interaction liquid chromatography (HILIC) and reversed‑phase LC separation modes for the detection of organic acids and sugar mono-/di-phosphates. Limits of detection could be achieved at 0.01–100 pmol levels on-column for 34 such analytes when IC was run in capillary scale.

Q. What are the benefits of using IC in forensic analysis?

A: In forensic science, IC has long been a technique of choice together with capillary electrophoresis (CE) for low‑order explosives analysis in particular (9,10). Low‑order explosives are those that generally deflagrate (burn rapidly) rather than detonate. Even though these explosions propagate slower, they can still inflict significant damage as seen in Oslo in 2011 and Boston in 2014. Many such homemade or improvised explosive devices (IEDs) contain inorganic low‑molecular‑weight salts such as ammonium nitrate or potassium perchlorate and are amenable to separation by either CE or IC, as opposed to HILIC, normal-phase LC, or reversed-phase LC. In the absence of confirmatory detection techniques (these offer the structural and elemental information necessary for identification of a compound), both CE and IC are normally used together to orthogonally confirm the presence of explosives using conductivity, electrochemical, and UV–vis detection. They are particularly good analytical techniques for complex samples and especially for post‑blast analysis where some spectroscopic techniques by themselves may lack the selectivity and sensitivity required for mixtures.

Guidelines proposed by the Scientific Working Group for Forensic Explosives Analysis (SWGFEX) include CE, IC, and IC‑MS as techniques of choice for explosives (11). We recently performed an in-depth review of IC-MS and how it could be used in the forensic explosives analysis arena (12). Few direct applications of IC‑MS to explosives detection actually exist despite a large body of literature on other related applications with common analytes like oxyhalides and nitrogen ion species (13). This prompted us to perform the first research for explosives detection using IC–HRAMS, while at the same time trying to simplify the traditional complexity of coupling the two techniques. For this we studied the effect of organic solvents in the IC eluent on separation selectivity and any resultant gains in MS sensitivity. We found that IC selectivity could be tailored using relatively small amounts of acetonitrile or methanol in the eluent (30–40%), but which was also enough to maximize ESI‑HRAMS sensitivity. This removed the need for a third auxiliary organic solvent infusion pump. Along with excellent selectivity for 11 organic acids and inorganic ions, the method was quantitative over multiple orders of magnitude and sensitive at a level similar to traditional IC-MS/MS configurations. This approach highlighted the value of HRAMS even further in that it could differentiate between perchlorate (m/z 98.9491) and the bisulphate anion (m/z 98.9555), which was an artifact of chemical suppression and which would have otherwise interfered with its detection. We went on to use the approach to detect a black powder substitute in a latent human fingermark deposit (14).

Q. You used anion exchange chromatography to investigate gunshot residue in three ammunition types. Can you tell us the purpose of this research, the technique you developed for this application, and what were the main challenges you had to overcome from an analytical perspective?

A: This was the first research paper of a series using IC and IC-MS for energetic materials analysis performed during Elizabeth Gilchrist’s PhD thesis. The purpose was to determine whether ammunition types used could be identified based on their post-discharge residues (15). The objectives of the work were to develop a quantitative anion exchange chromatography method to characterize the range of anions present in gunshot residue and then to potentially differentiate three ammunition types from each other based on occurrence or quantity of specific anions present. The method was developed using an IC resin composed of a hyperbranched anion exchange polymer electrostatically attached to sulfonated core particles.

This resin shows particular selectivity for perchlorate, but has been used in the past for explosives analysis (16). After we developed and validated the method, we found that the resulting traces from the three ammunition types were relatively complex, but also different. Up to 15 low-molecular-weight anion peaks were observed in all cases for n = 30 replicate firings of each type. Some anions were present in one ammunition, but not in another. For example, cyanate was present in residue from a 9-mm ammunition type and nitrite was present in two types of larger rifle ammunition. Furthermore, the quantities of each ion or their ion ratios could also be used to differentiate the three ammunition types from each other. In comparison to explosives analysis, which only usually contains a small number of components amenable to IC, such a useable and sufficiently complex signature was very exciting for us. In particular, because this represented some of our earlier work on energetic materials (17), the need for structural information became very evident early on. As IC separates by size and charge, it is rare that species elute before the small, singly charged fluoride ion. However, a more rapidly eluting species seemed to be indicative of gunshot residue here. Further work is needed, especially using IC-HRAMS, to help identify it and any other anions or cations in gunshot residue. Furthermore, new approaches to gunshot residue analysis are now needed, as traditional approaches, such as scanning electron microscopy with energy dispersive X-ray spectroscopy for barium-antimony-lead composite particle detection become forensically more limited because of phasing out of lead-based primers.

Q. You also developed a capillary‑scale ion chromatography method with suppressed conductivity to look at gunshot residue, sweat, and human fingerprints. What were the advantages of this chromatographic technique for this application and what were your findings?

A: We ran this study in collaboration with a leading vendor using a capillary IC system. We were one of the first groups to use this technology, which was a really great opportunity to see whether we could perform more sensitive analysis of samples of limited size or availability as is often the case in forensic science (18). In terms of performance, we could detect up to 1800-fold lower mass on column than we could with 2-mm bore IC (0.3–26.2 pg) for 17 anions (19). This meant that we could reduce our injection volume by 100-fold to 0.4 µL to even allow replicate injections of very small samples, in this case, extracts of single latent human fingermarks. Even more exciting was that we could detect traces of gunshot residue indicative species and contact with a black powder substitute in a fingermark in five successively deposited impressions on a glass surface. With respect to sweat analysis, we were able to identify smokers from traces of benzoate and thiocyanate, the latter of which we have shown in the past was a useful marker of smoking behaviour. Interestingly, given that this was capillary‑scale IC operating at a low flow rate of 11 µL/min, the system was left running for 93 days uninterrupted without any real loss in performance, which meant that we could run samples at any point without lengthy startup or equilibration times.

References

1. R. Loos and D. Barceló, Journal of Chromatography A938(1–2), 45–55 (2001).

1. A.F. Loughlin, G.L. Skiles, D.W. Alberts, and W.H. Schaefer, Journal of Pharmaceutical and Biomedical Analysis 26(1), 131–142 (2001).

1. B. Divjak, M. Novic, and W. Goessler, Journal of Chromatography A862(1), 39–47 (1999).

1. G. Socher, R. Nussbaum, K. Rissler, and E. Lankmayr, Journal of Chromatography A912(1), 53–60 (2001).

1. C. Bignardi et al., Journal of Chromatography A1372, 133 (2014).

1. D. Vuckovic et al., Analytical Chemistry83, 1944 (2011).

1. L. Barron et al., Talanta 69, 621 (2006).

1. K. Burgess et al., Rapid Communications in Mass Spectrometry25, 3447 (2011).

1. G.W. Dicinoski et al., Analytical Letters 39, 639 (2006).

1. M. Calcerrada et al., TrAC - Trends in Analytical Chemistry75, 75 (2016).

1. TWGFEX Laboratory Explosion Group, Standards & Protocols Committee, Recommended Guidelines for Forensic Identification of Post-Blast Explosive Residues (Revision 8) (2009).

1. L. Barron et al., Analytica Chimica Acta806, 27 (2014).

1. G.H.L. Lang et al., Journal of Forensic Sciences 54, 1315 (2009).

1. E.S. Gilchrist et al., Analytica Chimica Acta865, 83 (2015).

1. E. Gilchrist et al., Forensic Science International221, 50 (2012).

1. C. Johns et al., Journal of Chromatography A1182, 205 (2008).

1. E. Gilchrist, F. Jongekrijg, L. Harvey, N. Smith, and L. Barron, Forensic Science International 221(1–3), 50–56 (2012).

1. E. Gilchrist et al., Analyst 137, 1576 (2012).

1. E. Gilchrist, N. Smith, and L. Barron, Analyst137(7), 1576–1583 (2012).

Leon Barron is a senior lecturer in forensic science in the Division of Analytical and Environmental Sciences at King’s College London. He has been actively engaged in research into separation science for over 15 years. This includes the development of new stationary phases and sorbents as well as their integration into miniaturized platforms for sampling, preparation, and analytical separations. He also leads projects on analytical method development for trace analysis using advanced sample extraction techniques, followed mainly by IC-MS and LC–MS analysis. Applications of his research lie across the forensic and environmental science fields. One of his main interests lies in using new analytical technologies to understand the sources, fate, and effects of emerging contaminants such as pharmaceuticals, illicit drugs, and personal care product residues in our environment and in humans. He also works closely with police and with government security agencies on the development of high resolution screening methods for large numbers of inorganic and organic explosive traces. He was awarded his B.Sc. (Hons) in 2001 and Ph.D. in 2005 at Dublin City University, Ireland.

E-mail:leon.barron@kcl.ac.uk