The analysis of trace level environmental contaminants in water samples presents a unique challenge to analytical scientists. The complexity of the matrix and the fact that many chemical background species will be present at significantly higher levels than the contaminants we are looking to measure makes quantifying these analytes a challenging task.
High resolution analytical separations are required to resolve trace level analytes from the large number of background species that will inevitably be present. Hyphenated chromatography-mass spectrometry methods like gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-spectrometry (LC-MS) have therefore become ‘gold-standard’ techniques for environmental contaminant analysis. The ability of mass spectrometers to resolve analytes based on mass-to-charge ratio enables stable isotope labelled internal standards to be used which can significantly improve the quantitative performance of the method by correcting for random error in injection volume and for losses that occur during the extraction and clean-up process.
While these techniques are powerful, the data obtained from a LC-MS or GC-MS analysis is only as good as the samples that are introduced into the instruments. If the sample preparation doesn’t extract the analytes effectively, or there are overlapping isomer peaks in the chromatogram, then poor quantitative performance will result. In environmental contaminant analysis we are effectively looking for needles in a haystack, and strategies for analyte preconcentration and clean-up are a necessity to perform quantitative measurements of trace level contaminants.
An analytical workflow for the quantification of trace level environmental contaminants in water samples consists of 4 key steps: sampling, clean-up and pre-concentration, separation and finally, analysis. Each of these steps needs to be planned carefully to achieve accurate analysis. However, there are many options for sample clean-up and separation available to chromatographers and it can be difficult to know where to begin to select the best technique for your application.
Sampling and transport
The first part of a workflow is the taking of samples themselves, transporting them to the lab for analysis and storing them until they can be processed. This sounds straightforward however, all samples will eventually degrade over time and depending on the target compounds you want to study and the matrix they are in, specific sample preservation techniques may need to be employed to prevent losses occurring before samples get to the lab. It’s worth looking at the physical properties of the target analytes here to see what precautions need to be taken.
Some analytes may be photosensitive, for example polycyclic aromatic hydrocarbons will degrade if exposed to sunlight, so samples need to be stored in amber glass to prevent photodegradation from occurring. Acidification of samples is commonly used for analytes such as nutrients and nitrogen compounds, which will kill off any microbes present which could utilise the target analytes and convert them into another form giving an artificially low result.1 Acidification can also be used to ensure that some species like trace metals remain in solution and don’t precipitate out during storage.
Using isotopically labelled internal standards
For a quantitative mass spectrometry analysis, isotopically labelled internal standards can significantly enhance analytical accuracy. While these standards can frequently be expensive to obtain, they can be used to correct matrix effects and random error which occurs in volumetric measurements and sample injections.
Internal standards can be added at different stages during the extraction and clean-up process which will influence what the standard corrects for and how it can be used. Injection internal standards are added after the extraction process has been completed (i.e. immediately before analysis) so the concentration of the internal standard in the sample is known and by comparison of the internal standard peak with the analyte peak will enable quantification. Using a peak ratio will correct for any random error in injection volume as this would affect both the analyte and the internal standard equally.
An injection internal standard will not however correct for sample-to-sample variation which may occur during the extraction process, such as differences in analyte recovery, or matrix effects which may be different from one sample to the next. A surrogate internal standard which is a known amount of an isotopically labelled internal standard before the extraction takes place is required to correct for this variation. An isotopically labelled surrogate standard will have very similar physical properties as the unlabelled analyte. Therefore, its recovery during the extraction process should be the same as that of the analyte enabling an accurate determination of the analyte concentration. To ensure maximum accuracy surrogate and injection standards can be used together, which allows the recovery of the surrogate from the sample matrix to be determined and can be used to correct for any potential differences in extraction efficiency between the surrogate and the analytes.
An example of the use of a surrogate internal standard is illustrated in Figure 1, which shows the analysis of phthalates extracted from seawater using solid phase extraction. A labelled standard (dimethyl phthalate-d4 (DMP-d4)) is added prior to extraction, which allows a peak area ratio against the concentration of dimethyl phthalate (DMP) in the water sample to be determined. The relative response factor of DMP:DMP-d4 was determined by running series of standard solutions prior to analysis and can be used to convert the peak area ratio of DMP: DMP-d4 to give an accurate estimation of the concentration of DMP in the water sample.
Sample clean up and pre-concentration
Once a sample has arrived at the lab and an internal standard is added, analytes then need to be extracted from the sampled material and, if necessary, concentrated to detectable levels. This step is of key importance, it doesn’t matter how good our chromatography and mass spectrometry methods are if the amount of analyte injected is below the mass spectrometer's limit-of-detection. Before performing a clean-up method, filtration through a 0.45um filter is recommended to remove any suspended solids or biological matter from the sample. If the clean-up method being used involves passing the sample through a packed media like a solid phase extraction cartridge, then this step is essential to prevent fouling of the cartridge.
There are many options available for sample clean-up and the selection will depend on the structure of the analytes, the composition of the matrix and the concentration at which you expect the analytes to be present. For environmental water samples, clean-up will usually involve either liquid/liquid extraction or solid phase extraction (SPE). SPE can make use of a range of media such as cartridges, discs or solid phase microextraction fibres. The QuEChERS method (Quick, Easy, Cheap, Effective, Rugged, Safe) uses a combination of LLE and SPE methodologies, where a LLE separation is followed by dispersive SPE, where sorbent particles are added to the samples to bind matrix interferences. The particle-bound interferences are then precipitated out using centrifugation or removed via filtration to obtain a liquid extract/supernatant which contains purified analytes for GC-MS or LC-MS analysis.2
All these methods have their own specific factors which impact the efficiency of analyte extraction from water. For a liquid/liquid extraction, the pH of the liquid sample has a dramatic effect on the recovery of the target analytes. Ionised analytes are highly polar and will remain in the aqueous water layer rather than being extracted into an organic solvent. So, looking at the structure of your analytes and seeing what functional groups are present is necessary. Analytes with carboxylic acid groups for example, require pH control, acidifying the sample (ideally two pH units below the pKa of the analyte) ensures the analyte remains neutral and will extract more efficiently. This effect is demonstrated in an article by Gonzalez-Barriero et al.3 which used LLE (and SPE) to study perfluorinated alkylated substances in tap-water samples. Eleven different PFAS were analysed in this study and a comparison of LLE with and without pH control was performed. Shorter chain PFAS showed very poor recovery from tap water at ambient pH as they were ionised and highly polar. However, when sulfuric acid was added, reducing the pH to four, PFAS that were neutralised dramatically improved their recovery, with all PFAS except perfluorohexanoic acid (PFHxA) showing satisfactory recovery of 70% or greater at this pH, with pFHxA yielding 64% recovery (vs 0% at neutral pH).
SPE methods use solid stationary phase sorbents through which a liquid sample is passed. Analytes can be concentrated on the stationary phase and eluted through stepwise addition of an elution solvent. SPE methods are particularly useful when a high level of pre-concentration is required, large volumes of aqueous sample can be passed through a SPE cartridge and then eluted in small volumes of solvent to achieve high concentration factors. The selection of sorbent in SPE can be confusing, so it's worth consulting SPE manufacturers to see if they have recommendations for a particular application. A standard C18 cartridge may work for most analytes but if you have ionisable analytes for example, better extraction efficiency may be obtained with an ion-exchange sorbent. As with LLE, pH control is vital, with a reversed-phase cartridge like C18, the retention of polar analytes can be improved by ensuring they pass through the cartridge as neutrals.
An alternative method that can be used for volatile analytes is headspace GC-MS. This method works by placing the water sample in a partially filled sealed vial and then gently heating until volatile analytes have equilibrated into the headspace above the sample. The sample vapour can then be injected into a GC-MS system leaving nonvolatile background species behind in the liquid sample. Headspace analysis works very well in combination with sample preconcentration using solid phase microextraction (SPME) which uses a sorbent coated fibre which is introduced via a hollow needle into the headspace above the sample. Analytes in the headspace can then concentrate up on the fibre before being introduced into the heated injector of the GC-MS which will thermally desorb them for analysis.
Chromatographic methods such as gas chromatography (GC) and liquid chromatography (LC) are staple techniques in environmental analysis. GC uses long, narrow bore silica capillaries (30-60m), where the inner wall is coated with a layer of a stationary phase. The chemistry of the stationary phase layer can be tailored to the analytes under investigation. Analytes are introduced into a heated injector (~250 °C) where they flash volatilise before being directed by an inert gas flow onto the column which is held in a temperature-controlled oven at a relatively low temperature (40-50°C) at the start of an analysis. The majority of analytes will condense on the front of the column at this temperature. The oven temperature is then ramped to create a temperature gradient. Low-boiling point species will elute from the column first with higher-boiling species eluting later in the analysis, and species with similar boiling points are resolved based on the strength of interaction with the stationary phase.
Optimising GC methods is relatively simple and usually involves setting a temperature gradient which allows analytes enough time to separate. However, there are other concerns which can arise, some analytes may exhibit poor thermal stability and decompose in the injector, or they may have reactive groups (carboxylic acids, amines etc.) will give poor peak shapes due to them having poor solubility in the stationary phase or by causing them to bind unpredictably to the injection liner or fused silica column. To resolve these effects, a chemical derivatisation step such as methylation or acetylation can be performed to ‘block’ reactive groups, improving peak shapes, and enhancing the volatility and thermal stability of the analyte.4
In contrast to GC, mobile phase composition plays a central role in the LC separation mechanism. LC columns are typically 5-25 cm long and packed with particles (3-5 µm HPLC, <2µM UPLC) which can be coated with a stationary phase. As with SPE there are many LC stationary phases and the selection depends on the analytes properties, as the stationary phase should have similar properties to the analytes in order to retain them well. The non-polar C18 phase is the most popular as most compounds will show some retention on this phase. However other phases may give better performance for certain analytes for example phenyl stationary phases work well with aromatic species like PAHS.5
Once analytes are introduced onto the column, they separate based on their partitioning between the liquid mobile phase and the particulate stationary phase. Elution can be controlled using a solvent gradient. In a reversed-phase separation increasing the percentage of organic solvent in the mobile phase will increase the solubility of non-polar analytes in the mobile phase, reducing their partitioning into the stationary phase and causing analytes to be eluted more quickly. Therefore highly polar species with a high affinity for the aqueous mobile phase will be eluted rapidly, with larger non-polar species being eluted later. Control of mobile phase pH again is again vital, if the pH of the mobile phase is close to the pKa of an ionisable analyte then the analyte can exist as both ionised and neutral forms giving rise to peak broadening/splitting. Therefore, adding a pH modifier can help improve retention, prevent broadening, and enhance ionisation in LC-MS ion sources.
GC-MS analysis primarily uses the electron ionisation source, which causes extensive fragmentation, that allows fragment ion libraries such as the NIST main library to provide preliminary identification of unknown species. In a targeted quantitative analysis involving the use of internal standards, the best form of GC-MS experiment to use is a selected ion monitoring experiment where for each analyte and standard a minimum of two fragment ions (a quantifying ion and a qualifying ion) are monitored. By monitoring a limited number of ions, many background interferences can be removed from the chromatogram and the sensitivity of the method is enhanced as the mass spectrometer can spend a greater proportion of its duty cycle transmitting analyte and standard ions.
In LC-MS analysis, electrospray and atmospheric pressure chemical ionisation sources are widely used, with these approaches intact molecular ion species are most often observed, so libraries are not as well developed. The best type of mass spectrometer for quantitative analysis at trace levels is the triple quadrupole (QQQ) mass analyser, although other analysers such as orbitraps can also be used. In a QQQ, multiple reaction monitoring (MRM) can be employed where target analytes and standards are isolated in the first quadrupole, fragmented in the second and then two or more known fragment ions are transmitted in the third quadrupole. This highly specific approach enables unrivalled sensitivity and has been used to detect many environmental contaminants at ultra-trace levels.6
1 S˙liwka-Kaszyńska M., Kot-Wasik. A., Namieśnik J., (2003) Preservation and Storage of Water Samples, Critical Reviews in Environmental Science and Technology, 33:1, 31-44, DOI: 10.1080/10643380390814442
2 Perestrelo R., Silva P., Porto-Figueira P., et al. (2019) QuEChERS - Fundamentals, relevant improvements, applications and future trends, Analytica Chimica Acta, 1070, 1-28, DOI: 10.1016/j.aca.2019.02.036
3 González-Barreiro C., Martínez-Carballo E., Sitka A., et al. (2006) Method optimization for determination of selected perfluorinated alkylated substances in water samples. Analytical and Bioanalytical Chemistry 386, 2123–2132 DOI: 10.1007/s00216-006-0902-7
4 Ranz A., Maier E., Motter H., Lankmayr, E. (2008), Extraction and derivatization of polar herbicides for GC-MS analyses. Journal of Separation Science, 31: 3021-3029. DOI: 10.1002/jssc.200800290
5 Letzel T., Pöschl U., Wissiack R., Rosenberg M., Niessner R., (2001) Phenyl-Modified Reversed-Phase Liquid Chromatography Coupled to Atmospheric Pressure Chemical Ionization Mass Spectrometry: A Universal Method for the Analysis of Partially Oxidised Aromatic Hydrocarbons. Analytical Chemistry, 73:1, 1634-1645. DOI: 10.1021/ac001079t
6 Xu S., Knoerr S., (2020) Quantitative analysis of trace- and ultratrace-level Dimethylsilanediol (DMSD) in water, soil, sediment and biosolids by liquid chromatography-triple quadrupole tandem mass spectrometry, International Journal of Environmental Analytical Chemistry, 100:3, 241-253, DOI: 10.1080/03067319.2019.1636039
Dr Jim Reynolds, from Lincoln in the UK, studied Biochemistry at Nottingham Trent University in the UK and received his PhD in Analytical Chemistry which focussed on mass spectrometry ion source development in 2003. After working in several different positions, he is now a Reader in Analytical Chemistry in the Department of Chemistry at Loughborough University. He has worked in several areas of research including trace level analysis both for environmental and clinical applications such as exhaled breath analysis. He also has a keen interest in the development of ambient ion sources for in-situ mass spectrometry and ion mobility spectrometry to enable field-based measurements.