Latest Advances in the Analysis of Complex Environmental Matrices
A report from a meeting held on Friday 22nd February 2019 at The Royal Society of Chemistry, Burlington House, London.
Over 60 delegates and speakers attended this one day meeting which was jointly organised by the Environmental Chemistry Group (Dr Roger Reeve), the Separation Science Group of the Analytical Division (Dr Lee Williams), and the Water Science Forum (Professor Graham Mills). It was the fourth in a series of biennial meetings looking at a wide range of instrumental analytical techniques involved in the analysis of environmental matrices, including sampling, clean-up, analysis (often by multi-dimensional chromatographic techniques and mass spectrometry), and computational data processing. Eight presentations were made by representatives from across academia and industry. The keynote lecture was given by Dr Emma Schymanski (University of Luxemburg) on the use of environmental informatics to identify unknown chemicals and their effects. There was also an exhibition by instrument manufacturers and laboratory consumables suppliers.
Dr Roger Reeve (Environmental Chemistry Group) opened the meeting, welcomed the delegates and chaired the first session. The first two lectures focussed on highlighting the need for computational techniques in chromatographic data analysis in compound identification and effect prediction. The first was given by Dr Leon Barron (King’s College London) on ‘Mixing high resolution chemical analysis and machine learning in ecotoxicology for aqueous invertebrates’. Samples from water, fish (tissue and plasma/bile) and invertebrates (whole body) were analysed by liquid chromatography two dimensional mass spectroscopy (LC-MS/MS), suspect screening and machine learning (ML) by liquid chromatography high resolution mass spectroscopy (LC-HRMS), and metabolomics investigation by hydrophobic interaction chromatography (HILIC-HRMS). LC-HRMS was found to markedly aid new compound identification, while ML provided a good prediction of retention time on liquid chromatography stationary phases, which could then be developed into bioconcentration factors. ML algorithms have previously been used to reveal underlying trends for predictions in toxicology, but improvements in the quality and availability of data sets are needed for future development. As such, ML can support toxicological studies, but is no panacea.
Anthony Buchanan (SepSolve Analytical) gave the next talk, entitled ‘Enhanced confidence in river water quality monitoring using passive sampling and GC×GC-TOF-MS (two dimensional gas chromatography, time of flight mass spectroscopy) with tandem ionisation’. In this talk, he outlined current analytical challenges, including ‘priority’ substances and emerging pollutants. He discussed how grab sampling may miss pollution events and, as such, passive sampling may be more suitable. The four week example investigation used a silicone rubber passive sampler for hydrophobic substances found at eight locations across a river system. Flame retardants, musks, and pharmaceuticals were detected. Gas chromatography was conducted with a non-polar column for the first dimension and a polar column for the second dimension. Mass spectrometry was conducted via fast switching between harder and softer ionisation energies to improve confidence in compound identification, as required by legislation. This produced two MS datasets from a single run with no added analysis time.
After the coffee break, the session was chaired by Dr Lee Williams (University of Sunderland). He introduced Dr Caroline Gauchotte Lindsay (University of Glasgow), presenting ‘Micro- and nanoplastic pollution of freshwater and wastewater treatment systems’. Dr Gauchette Lindsay’s group has been investigating the size and morphology of plastics including both fibrous and globular nanoparticles. Sediments under investigation were first pretreated, then examined under a scanning electron microscope (SEM) with energy dispersive spectroscopy (EDS). Care had to be taken to avoid particle fragmentation throughout pretreatment processes. Fibres comprised approximately 88% and 95% of all plastic pieces. Characterisation of water samples by Fourier transform infrared attenuated total reflectance spectroscopy (FT-IR) again showed fibres were dominant, although they proved challenging to accurately quantify, with plastics (58% polypropylene and 5% polyethylene) and cellulose (11%) identified, while the remainder was of unknown origin.
Dr John Quick (ALS Environmental Ltd) continued with ‘Exploring the advantages of automated sample preparation and GC-TOF for semi-volatile organic compound and pesticide analysis in environmental waters’. In this talk, he outlined the advantages of automation, including increased accuracy and precision, lower sample volumes, and fewer consumables required with associated benefits in cost and health and safety. For alkylphenols and ethoxylates, a sample was simultaneously extracted and derivatised using hexane/pentafluorobenzoyl chloride. In a second study, automation of the classical technique of liquid-liquid extraction (dispersive liquid-liquid micro-extraction) was used for semi-volatile organic compounds and pesticides.
The morning finished with the keynote lecture presented by Associate Professor Dr Emma Schymanski (Luxemburg Centre for Systems Biomedicine) on ‘Environmental informatics to identify unknown chemicals and their effects’ (see Figure 1). European, US and worldwide community initiatives to help connect chemisty and toxicity with environmental observations were discussed.
Over 60 delegates and speakers attended this one day meeting which was jointly organised by the Environmental Chemistry Group (Dr Roger Reeve), the Separation Science Group of the Analytical Division (Dr Lee Williams), and the Water Science Forum (Professor Graham Mills). It was the fourth in a series of biennial meetings looking at a wide range of instrumental analytical techniques involved in the analysis of environmental matrices, including sampling, clean-up, analysis (often by multi-dimensional chromatographic techniques and mass spectrometry), and computational data processing. Eight presentations were made by representatives from across academia and industry. The keynote lecture was given by Dr Emma Schymanski (University of Luxemburg) on the use of environmental informatics to identify unknown chemicals and their effects. There was also an exhibition by instrument manufacturers and laboratory consumables suppliers.
Dr Roger Reeve (Environmental Chemistry Group) opened the meeting, welcomed the delegates and chaired the first session. The first two lectures focussed on highlighting the need for computational techniques in chromatographic data analysis in compound identification and effect prediction. The first was given by Dr Leon Barron (King’s College London) on ‘Mixing high resolution chemical analysis and machine learning in ecotoxicology for aqueous invertebrates’. Samples from water, fish (tissue and plasma/bile) and invertebrates (whole body) were analysed by liquid chromatography two dimensional mass spectroscopy (LC-MS/MS), suspect screening and machine learning (ML) by liquid chromatography high resolution mass spectroscopy (LC-HRMS), and metabolomics investigation by hydrophobic interaction chromatography (HILIC-HRMS). LC-HRMS was found to markedly aid new compound identification, while ML provided a good prediction of retention time on liquid chromatography stationary phases, which could then be developed into bioconcentration factors. ML algorithms have previously been used to reveal underlying trends for predictions in toxicology, but improvements in the quality and availability of data sets are needed for future development. As such, ML can support toxicological studies, but is no panacea.
Anthony Buchanan (SepSolve Analytical) gave the next talk, entitled ‘Enhanced confidence in river water quality monitoring using passive sampling and GC×GC-TOF-MS (two dimensional gas chromatography, time of flight mass spectroscopy) with tandem ionisation’. In this talk, he outlined current analytical challenges, including ‘priority’ substances and emerging pollutants. He discussed how grab sampling may miss pollution events and, as such, passive sampling may be more suitable. The four week example investigation used a silicone rubber passive sampler for hydrophobic substances found at eight locations across a river system. Flame retardants, musks, and pharmaceuticals were detected. Gas chromatography was conducted with a non-polar column for the first dimension and a polar column for the second dimension. Mass spectrometry was conducted via fast switching between harder and softer ionisation energies to improve confidence in compound identification, as required by legislation. This produced two MS datasets from a single run with no added analysis time.
After the coffee break, the session was chaired by Dr Lee Williams (University of Sunderland). He introduced Dr Caroline Gauchotte Lindsay (University of Glasgow), presenting ‘Micro- and nanoplastic pollution of freshwater and wastewater treatment systems’. Dr Gauchette Lindsay’s group has been investigating the size and morphology of plastics including both fibrous and globular nanoparticles. Sediments under investigation were first pretreated, then examined under a scanning electron microscope (SEM) with energy dispersive spectroscopy (EDS). Care had to be taken to avoid particle fragmentation throughout pretreatment processes. Fibres comprised approximately 88% and 95% of all plastic pieces. Characterisation of water samples by Fourier transform infrared attenuated total reflectance spectroscopy (FT-IR) again showed fibres were dominant, although they proved challenging to accurately quantify, with plastics (58% polypropylene and 5% polyethylene) and cellulose (11%) identified, while the remainder was of unknown origin.
Dr John Quick (ALS Environmental Ltd) continued with ‘Exploring the advantages of automated sample preparation and GC-TOF for semi-volatile organic compound and pesticide analysis in environmental waters’. In this talk, he outlined the advantages of automation, including increased accuracy and precision, lower sample volumes, and fewer consumables required with associated benefits in cost and health and safety. For alkylphenols and ethoxylates, a sample was simultaneously extracted and derivatised using hexane/pentafluorobenzoyl chloride. In a second study, automation of the classical technique of liquid-liquid extraction (dispersive liquid-liquid micro-extraction) was used for semi-volatile organic compounds and pesticides.
The morning finished with the keynote lecture presented by Associate Professor Dr Emma Schymanski (Luxemburg Centre for Systems Biomedicine) on ‘Environmental informatics to identify unknown chemicals and their effects’ (see Figure 1). European, US and worldwide community initiatives to help connect chemisty and toxicity with environmental observations were discussed.
She outlined the need for target, suspect and non-target screening to identify species of interest in rivers and wastewaters, and identified 364 targets in Swiss wastewater using solid phase extraction (SPE) and LC-MS/MS. Suspect screening involved the selection of candidates based on known and calculated physical properties. Several mass spectral libraries were needed to support this work, as each had unique entries. As such, the NORMAN network was set up within the EU for exchanging information, and validating and harmonising data on emerging environmental substances between research groups. Associate Professor Dr Schymanski described using the NORMAN suspect list exchange and its chromatography data bank (Digital Sample Freezing Platform).
|
Lunchtime provided an opportunity for delegates to network and inspect the sponsors’ exhibition. David Sharpe and Guy Jones (RSC) gave a short presentation on ChemSpider. This database of chemical information on small molecules amalgamates >250 data sources and is curated to ensure quality representation. It was used by a number of the presenters.
The afternoon session was chaired by Professor Graham Mills (University of Portsmouth). Professor Alistair Boxall (University of York) discussed ‘Temporal and spatial variations in pharmaceutical concentrations in an urban river system’. In his work, inputs to the Ouse River system (Yorkshire) were studied across 11 sites, and 41 pharmaceuticals monitored. A detailed model was produced predicting how concentrations varied over time at each of the sites. This model can be successively upgraded as more local details of the river system are introduced. Pharmaceutical monitoring was extended to a global network and correlation of results suggested a greater problem of high concentrations of pharmaceuticals in lower income countries.
Dr Katie Read (University of York) spoke next, introducing ‘GC-TOF for remote monitoring – Cape Verde atmospheric laboratory’. Much of the talk discussed problems of atmospheric analysis in a laboratory located in the middle of the Atlantic (see Figure 2), despite using standard methodologies. Monitoring covered the whole range of atmospheric constituents from major components to species at parts per trillion levels. Air masses from a variety of origins and effects were observed from Africa, Europe and North America in addition to the pristine Atlantic air masses.
The afternoon session was chaired by Professor Graham Mills (University of Portsmouth). Professor Alistair Boxall (University of York) discussed ‘Temporal and spatial variations in pharmaceutical concentrations in an urban river system’. In his work, inputs to the Ouse River system (Yorkshire) were studied across 11 sites, and 41 pharmaceuticals monitored. A detailed model was produced predicting how concentrations varied over time at each of the sites. This model can be successively upgraded as more local details of the river system are introduced. Pharmaceutical monitoring was extended to a global network and correlation of results suggested a greater problem of high concentrations of pharmaceuticals in lower income countries.
Dr Katie Read (University of York) spoke next, introducing ‘GC-TOF for remote monitoring – Cape Verde atmospheric laboratory’. Much of the talk discussed problems of atmospheric analysis in a laboratory located in the middle of the Atlantic (see Figure 2), despite using standard methodologies. Monitoring covered the whole range of atmospheric constituents from major components to species at parts per trillion levels. Air masses from a variety of origins and effects were observed from Africa, Europe and North America in addition to the pristine Atlantic air masses.
Challenges included dust, salt, volcanic rock and corrosion in the aggressive environment. In situ measurement and sampling with subsequent laboratory analysis were used. Calibration posed difficulties. Few, if any, methods are available to generate known samples of analytes at the low concentrations needed and in controllable real world matrices. Generation devices (e.g. permeation devices) were used where species were known to be unstable in cylinders (such as ozone/ oxygenated VOCs), with a blender to dilute to the specified concentration range. Existing GC-MS and new GC-TOF-MS analytical systems were compared. The results suggested that 50% more ozone was destroyed in this region than predicted by climate models, which may be due to catalytic destruction from natural halogens emitted by sea spray.
|
Professor Gary Fones (University of Portsmouth) presented ‘Liquid chromatography/quadrupole time of flight mass spectroscopy screening of polar pollutants sequestered by passive sampling devices at the river catchment scale’. The potential for passive samplers used for long-term trend monitoring was discussed.
Their advantages include locating sources of pollution and generating more robust time-integrated data compared to spot sampling, or ‘snapshot’ type grab sampling (see Figure 3), which may not be representative of conditions where concentrations of pollutants fluctuate or are not homogeneous. Since achieving legislated detection limits may be complicated when only low volume spot samples are collected, a polar version of the Chemcatcher passive sampler was chosen for polar pesticide analysis at 14 sites in the River Rother catchment system (West Sussex) over a two week sampling period. Liquid chromatographic analysis was used with Quadrupole Time of Flight (QTOF), selected for accurate mass, fragment and isotope information and a high sensitivity coupled with a large dynamic range, allowing 51 compounds to be detected.
|
The final session of the day was a presentation by Wai-Chi Man (ThermoFischer Scientific), introducing the ‘Power of ion chromatography with mass spectrometry for polar pesticides in water’. Ion chromatography can be a suitable method for separating highly polar species as well as ionic species in water, including polar pesticides and disinfection chemicals such as haloacetic acids and inorganic oxyhalo- species. Ms Man demonstrated that sensitivity and selectivity could be increased by coupling the chromatograph with a mass spectrometer, whereupon polar pesticides in food and environmental samples could be determined well below Environmental Protection Agency (EPA) regulatory limits with no chemical derivatisation necessary and low chemical noise. She finished by asking whether slow separation of all possible components (including unexpected ones), or more rapid analysis with peak overlap of the main analyte with less likely species was better. The answer depends, of course, on the application, but reminds us of the multiple and sometimes conflicting factors that affect the selection of an appropriate analytical technique.
Synopses were prepared by Roger Reeve, Lee Williams and Graham Mills. The industry meeting sponsors were LECO, JSB, SepSolv Analytical and Shimadzu. Financial support is also acknowledged from Environment Sustainably and Energy Division of the RSC.
Keynote presentation slides are available online:
www.slideshare.net/EmmaSchymanski/rsc-environmental-cheminformatics-to-identify-unknowns-feb-2019
Synopses were prepared by Roger Reeve, Lee Williams and Graham Mills. The industry meeting sponsors were LECO, JSB, SepSolv Analytical and Shimadzu. Financial support is also acknowledged from Environment Sustainably and Energy Division of the RSC.
Keynote presentation slides are available online:
www.slideshare.net/EmmaSchymanski/rsc-environmental-cheminformatics-to-identify-unknowns-feb-2019