Abstract
The research presented herein describes the development of a 3D printed, multimodal, miniaturised passive sampler device and its successful application as a surrogate for a freshwater invertebrate species (Gammarus pulex) for qualitative and quantitative biomonitoring of contaminants of emerging concern (CECs). This is the first study to miniaturise, multiplex, and scale up the passive sampling process and utilise modelling techniques to predict the internal concentration of CECs in G. pluex from data collected in situ. This created an opportunity to rapidly estimate and prioritise the chemical risk of pollutants in situ while reducing animal use, increasing sustainability and reducing the disruptive environmental impacts of bioconcentration studies.The increased sensitivity of passive samplers to CECs at environmentally relevant concentrations in water compared to traditional grab sampling was demonstrated in a London urban river using a rapid liquid chromatography-tandem mass spectrometry method (LC-MS/MS) and a machine learning-assisted suspect screening workflow. Almost double the number of compounds identified in the water samples (n = 33) were also detected in the passive sampler extracts (n = 65) including several unique compounds not detected in the water. To overcome the drawbacks of using larger, single-sorbent passive samplers, a miniaturised 3D printed passive sampler device (3D-PSD) was prototyped using a commercially available methacrylate-based polymer resin. The final design consisted of a two-part assembly that connected easily via a friction interface fit and held five separate 9 mm sorbent disks, allowing the device to be multiplexed with different sorbents to increase the number of replicates and the sampled chemical space. It was noted that some CECs absorbed to the 3D-PSD housings, particularly those with a logD value close to or above that of the polymer resin. These compounds were able to be recovered from the resin using organic solvents, presenting an opportunity for the 3D-PSD housing to be utilised as a passive sampler in future work. The sampling rates (Rs) of three different sorbent chemistries (hydrophilic-lipophilic ABSTRACT balanced (HLB), anion, and cation exchange) were determined at the 9 mm disk scale for > 39 CECs. Due to the reduced sampling area of the 3D-PSD, there was a loss of sensitivity compared to standard passive sampler formats, highlighting a drawback of the 3D-PSD, which requires higher instrument sensitivity to compensate.
The 3D-PSD was then applied to field studies in a London urban freshwater river. The pollution impact point was identified as a wastewater plant effluent discharge point using rapid, direct injection LC-MS/MS. Following this, HLB-loaded 3D-PSDs were deployed upstream and downstream of the pollution point to compare the performance of the 3D-PSD in both an impacted and clean environment. Time-weighted average CEC concentrations in water were calculated using the Rs data previously determined and found to be in good agreement with the concentrations measured in water directly. A six-month field study was undertaken where 3D-PSDs containing all sorbents were deployed with co-occurring water and G. pulex collections to create a matched dataset for future modelling. The addition of the cation and anion sorbents increased the chemical space by 23 compounds. Of all compounds detected, imidacloprid (a neonaticide pesticide) presented a medium to high risk across all sorbent phases, was present in the G. pulex samples at every collection and exhibited the highest toxic unit value showing the suitability of the 3D-PSD for this purpose. Data collected during the six-month field study was investigated to determine if there was any relationship between CEC accumulation on the 3D-PSDs and that in G. pulex. Clustering analysis (hierarchical clustering and principal component analysis) revealed that the measured concentration data from 3D-PSDs was more similar to the G. pulex data when compared to those in water and there was a statistically significant relationship between the risk units calculated from the G. pulex and 3D-PSD data. Thereby, the average concentration in the animal could be predicted from the line of best fit with very high accuracy (a mean absolute error of 21 ± 21 ng g-1). The model was validated using literature data and also showed a good predictive relationship when using other data, indicating that the model can be ABSTRACT applied to a variety of studies and environments. Overall, the work presented herein demonstrates that a simple, cost-effective, scalable, and novel passive sampler device could be used for the first time as a surrogate when determining bioconcentrations in invertebrates. Thereby eliminating animal use, increasing sustainability and reducing the disruptive environmental impacts of bioconcentration studies. Therefore, this device, the associated analytical methods and predictive models hold promise as a uniquely scalable solution for simultaneous water monitoring and biomonitoring for the rapid risk assessment and prioritisation of large numbers of CECs in the aquatic environment.
| Date of Award | 1 Oct 2023 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | David Cowan (Supervisor), Stephen Sturzenbaum (Supervisor) & Leon Barron (Supervisor) |