Characterization of Nanoparticles in Drinking Water Using Field-Flow Fractionation Coupled with Multi-Angle Light Scattering and Inductively Coupled Plasma Mass Spectrometry

Abstract

The current absence of well-established and standardized methods for characterizing submicrometer- and nano-sized particles in water samples presents a significant analytical challenge. With the increasing utilization of nanomaterials, the potential for unintended exposure escalates. The widespread and persistent pollution of water by micro- and nanoplastics globally is a concern that demands attention, not only to reduce pollution but also to develop methods for analyzing these pollutants. Additionally, the analysis of naturally occurring nano entities such as bubbles and colloidal matter poses challenges due to the lack of systematic and consistent methodologies. This study presents Asymmetric Flow Field-Flow Fractionation (AF4) separation coupled with a UV-VIS spectrometer followed by Multi-Angle Light Scattering (MALS) for detection and size characterization of nanometric entities. It is coupled with an Inductively Coupled Plasma Mass Spectrometer (ICP-MS) for elemental analysis. Water samples from different sources, such as untreated mountain spring water, groundwater, and bottled drinking water, were analyzed. The system was calibrated using pure particle standards of different metallic compositions. Our study demonstrates the capability of AF4-UV-MALS-ICP-MS to detect metals such as Al, Ba, Cu, and Zn in particles of around 200 nm diameter and Mg associated with very small particles between 1.5 and 10 nm in different drinking water samples.

1. Introduction

Naturally occurring nano-sized particles are prevalent in water samples irrespective of their origin. However, due to their unique properties, metal, metal oxide, and metalloid nanoparticles are increasingly being applied in industrial applications and potentially reaching water bodies. In terms of production, nano silica (nSiO₂), nano titania (nTiO₂), and nano zinc oxide (nZnO) are in the highest demand. Pulit-Prociak et al. estimated the global production in 2014 to be at least 185,000 tons/year for nSiO₂, 60,000 tons/year for nTiO₂, and 32,000 tons/year for nZnO. Among metallic nanomaterials, nano silver (nAg) is the most produced, with an estimated 135 tons/year in 2014 [1]. In the food industry, nanoparticles such as nAg, nZn, and nCu are employed as antimicrobial agents in packaging [2]. Titanium dioxide (TiO₂) has been used as a food additive. Although its use as a food coloring agent is still permitted in some regions, the European Union banned it in 2022 due to concerns over potential genotoxicity [3]. The paints and coatings industry utilizes various nanoparticles, including nSiO₂, nAg, nCu, and nTiO₂, primarily to delay aging. The natural degradation of white TiO₂ pigment, initially added as microparticles, can cause the release of nTiO₂ into the environment [4]. The cosmetics industry employs metal and oxide nanoparticles, notably nZnO and nTiO₂, in sunscreens for their UV radiation-blocking properties [5]. Additionally, nAg particles are used for their antimicrobial properties in sensor development [6] and the production of electrically conductive materials [7]. The tunability of gold nanoparticles (nAu) makes them valuable for biomedical applications, including treatments against tumor cells and drug delivery [8]. Various types of nanoparticles are explored for environmental remediation, particularly those with adsorbent or catalytic properties, for removing micropollutants, heavy metals, and other contaminants from wastewater and drinking water [9,10]. Materials containing nanoparticles are not always disposed of properly, potentially entering wastewater or surface water directly [11,12]. Toxicological studies have demonstrated that metal or oxide nanoparticles can cause acute toxicity to microorganisms and some aquatic organisms. The effects of long-term exposure to engineered nanomaterials remain largely unknown [13].

The physical and chemical aspects of natural nanometric entities also require further research. DOLLOPs (Dynamically Ordered Liquid-Like Oxyanion Polymers) represent a notable example [14]. Research on DOLLOPs primarily focuses on their formation mechanisms as CaCO₃ nucleation precursors [15,16]. Traditionally, the crystallization of CaCO₃ is described using classical nucleation theory, suggesting the stochastic growth of unstable precursors until a critical nucleus of solid CaCO₃ initiates crystal formation. However, recent theories propose that crystal precursors consist of stable nanometric-sized polymers of CaCO₃, transforming into dense liquid nanodroplets upon initial liquid–liquid separation, gradually aggregating and forming a liquid-like intermediate phase before solid phase precipitation [14], which potentially can agglomerate on the surface of charged nanobubbles [17]. It has been suggested that such nanobubbles can also occur naturally, depending on the pH [18]. A deeper understanding of nano entities such as DOLLOPs could facilitate the development of novel water softening strategies or interventions to address issues related to hard water [19]. Furthermore, proteins, polysaccharides, lipids, and various macromolecules exist at the nanoscale and can adsorb onto metals (e.g., Fe, Mn, Al, Cu, Ni), forming colloids of diverse sizes [20,21,22]. Microorganisms play a significant role in converting metal oxides or ions into sulfide nanoparticles [23]. Silicate colloids, predominantly found in groundwater due to mineral erosion, are recognized for their ability to scavenge and transport toxic substances in aquatic environments [24]. Considering their widespread applications, increasing production rates, and associated environmental hazards, a thorough assessment of the fate and toxicity of nanoparticles in the environment is urgently needed.

In this study, we present Asymmetric Flow Field-Flow Fractionation (AF4) separation in combination with a UV-VIS spectrometer, followed by Multi-Angle Light Scattering (MALS) for the detection and size characterization of nanoscale entities. This setup is further connected to an Inductively Coupled Plasma Mass Spectrometer (ICP-MS) to enable elemental analysis of the various fractions. The analysis involves water samples from various sources, such as untreated mountain spring water, treated groundwater, and bottled drinking water, to represent different types of drinking water production plants. The system was calibrated using pure particle standards of various metallic compositions. Our study illustrates the capability of AF4-UV-MALS-ICP-MS to detect metals such as Al, Ba, Cu, and Zn in particles of approximately 200 nm in diameter. Additionally, the system was able to identify Mg associated with very small particles ranging from 1.5 to 10 nm in different drinking water samples.

2. Theory

2.1. Field Flow Fractionation

AF4, a Field Flow Fractionation technique, involves injecting the sample carried by an aqueous mobile phase into a narrow channel [25]. First, a focusing step concentrates the sample against a semipermeable membrane strip, achieved by a perpendicular crossflow. Smaller particles diffuse more easily against the crossflow, positioning them higher in the channel. Subsequently, during elution, the sample flows forward under a laminar flow with a parabolic velocity profile along the channel height. Smaller particles, positioned higher in the channel, emerge earlier. The retention time t_R of the particles is expressible based on their diffusion coefficient D, cross flow V_x, detector flow V_d, and channel height w, as initially described by Giddings and Wahlund [26],

$$t_R={\displaystyle \frac{w^2}{6D}}\ln \left(1+{\displaystyle \frac{V_x}{V_d}}\right).$$

(1)

Additionally, the Stokes–Einstein relation,

$$D={\displaystyle \frac{{\mathrm{k}}_{\mathrm{B}}T}{6\pi \eta r_H}},$$

(2)

describes the diffusion D of spherical particles through a liquid based on the Boltzmann constant k_B, absolute temperature T, dynamic viscosity η, and hydrodynamic radius r_H. The retention time t_R can be directly related to the hydrodynamic radius by inserting the Stokes–Einstein relation into Equation (1),

$$t_R={\displaystyle \frac{w^2\pi \eta r_H}{{\mathrm{k}}_{\mathrm{B}}T}}\ln \left(1+{\displaystyle \frac{V_x}{V_d}}\right),$$

(3)

describing how variations in crossflow can delay or accelerate particle elution. Other factors impacting AF4 operation include the membrane composition, particle surface charge, and mobile phase composition (pH and ionic strength), which influence the interactions between particles and the membrane, as well as electrical double-layer characteristics. Therefore, while primarily a physical separation method, AF4 necessitates consideration of various physicochemical factors to ensure method reliability and avoid potential issues.

After separation by AF4, the sample was first analyzed for UV absorption and subsequently entered a Multi-Angle Light Scattering (MALS) section employing a laser and multiple detectors positioned at various angles around the flow to measure the light scattered by the particles in the stream. MALS is a derivative of the Static Light Scattering (SLS) techniques that assess the angular dependence of light scattering to determine the particle’s radius of gyration (r_g). Equation (4) presents a simplified expression for the relation between the radius and the angular scattering [27],

$${\displaystyle \frac{R_{\theta }}{{\mathrm{K}}^{*}c}}=M_w\left(1-{\displaystyle \frac{16{\pi }^2}{3{\lambda }^2}}\langle r_g^2\rangle {\sin }^2\left({\displaystyle \frac{\theta }{2}}\right)\right),$$

(4)

where R_θ represents the Rayleigh ratio, indicative of the scattered light intensity exceeding that of the pure solvent, as a function of the scattering angle (θ). The optical constant K* is determined by the refractive properties of both the solvent and the sample, while M_w denotes the weight-average molar mass of the sample, c signifies the sample concentration, and λ represents the incident laser wavelength. Equation (4) applies to low concentrations, small particles, and low scattering angles only. For larger particles, the angular dependence must be accounted for differently, including the effects of the particle’s shape, and for higher concentrations, corrections for multiple scattering are necessary.

2.2. Calcium Carbonate Nucleation

Calcium carbonate plays a crucial role in biomineralization processes and geosciences in general, forming both biological (reefs and ocean sediments) as well as geological scales that bind a significant amount of CO₂, affecting ocean chemistry and contributing to water hardness. It serves as a model for studying both classical [28] and nonclassical [29] mineral crystallization and has been researched for over a century. Despite this, early crystallization stages, particularly prenucleation, remain poorly understood. Amorphous calcium carbonate (ACC) is recognized as an intermediate in mineralization processes [30,31,32,33], with evidence suggesting that multiple ACC species exist. ACC, a transient precursor in biomineralization, is preceded by species forming immediately after ion contact, before nucleation, as postulated [34] and suggested by modeling [35]. Recent studies [14] reveal stable prenucleation ion clusters, even in undersaturated solutions, characterized by equilibrium thermodynamics and a multiple-binding model [15]. These clusters, composed of alternating calcium and carbonate ions, exhibit a dynamic topology of chains, branches, and rings. The existence of dynamic, flexible, and hydrated precursors explains the formation of liquid-like amorphous states and the non-classical growth behavior of ACC. These dynamically ordered liquid-like oxyanion polymer clusters (DOLLOPs) are crucial in calcium carbonate nucleation and may also influence the crystallization of other minerals.

Figure 1 shows a schematic of the different mechanisms and stages of calcium carbonate precipitation.

Simulations [36] suggest that water plays a key role in ACC nucleation, since their hydration is driven by thermodynamic stability, with water incorporation increasing as particle size grows. Although the enthalpy decreases consistently with increasing water levels in ACC particles, accounting for the entropic penalty of water removal from the bulk reveals a free energy minimum. The preferred stoichiometry of ACC depends on particle size; as the particles grow, the water content per formula unit of calcium carbonate also increases. This effect can lead to a drier core and wetter shell structure.

Overall, experimental and simulation data on the precipitation of calcium carbonate supports a nonclassical nucleation and growth mechanism. The transition from amorphous (ACC) to crystalline CaCO₃ (Calcite, Aragonite, or Vaterite) may follow a classical or non-classical nucleation process triggered by environmental factors, representing a second growth stage.

A simple, general definition of water hardness is the amount of dissolved calcium and magnesium in the water, and for both ions, the nucleation pathways mentioned above might occur. Naturally, a determination of the concentration of both of these elements is of interest in the context of the analysis presented. However, the ICP-MS operating conditions in STD mode made detecting calcium problematic due to the isobaric interference of ⁴⁰Ar⁺ on its most abundant isotope, ⁴⁰Ca. Polyatomic interferences of ⁴⁰Ar¹⁶O⁺ and ⁴⁰Ca¹⁶O⁺ also occur with the main iron isotope, ⁵⁶Fe [37]. These interferences prevent the signals of these metals from being measured accurately, which is why we focused our analysis on the ions that could be detected without any interference as a compromise for maintaining the higher sensitivity by applying STD mode over KED mode (which would have made the calcium detection plausible). Since the aqueous chemistry of magnesium and calcium ions concerning scaling is comparable and it is known that these ions can crystallize together as calcium magnesium carbonate also on the nanoscale [38], and because the CaCO₃ compound is used as a representative system for studying the processes of nucleation and crystallization in general [39,40,41,42,43], we chose to use magnesium as a tracer for the possible presence of (calcium magnesium carbonate-)DOLLOPs. We are planning to apply different analysis conditions in the ICP-MS, such as collision/reaction cells, or to test alternative plasma conditions, to avoid these interferences in a subsequent work in the future.

In the present work, we present experimental data corroborating the existence of both magnesium-containing DOLLOPs and amorphous magnesium(calcium)carbonate in natural waters, as the presence of magnesium can be correlated to scatterers of a few or a few hundred nanometers in size, respectively.

3. Materials and Methods

3.1. AF4-UV-MALS-ICP-MS Instrumentation and Software

Size fractionation of aqueous samples was performed using an Eclipse Neon (Wyatt Technology, Goleta, CA, USA) equipped with a short separation channel (153 × 22 mm) set at a fixed height of 250 μm. The focusing position of particles in the channel was consistently set at 25% of the channel length for all experiments. A 1260 Infinity II HPLC system (Agilent Technologies, Santa Clara, CA, USA) with a pump, degasser, and auto-sampler was utilized to pump the mobile phase and sample injections into the AF4 channel. The mobile phase, filtered using VacuCap^® filters (Pall Corporation, New York, NY, USA) with 0.1 μm Supor^® membranes, was supplemented with an inline filter housing a 0.1 μm PVDF membrane (Merck Millipore, Burlington, VT, USA) post-HPLC pump. The separation channel was equipped with a 10 kDa Polyethersulfone (PES) membrane (Microdyn Nadir^®, Wiesbaden, Germany) and was optimized for a multi-step consecutive method with 0.2 mM NH₄HCO₃ as the mobile phase and a sample volume of 750 μL. All samples were filtered using 1.2 μm glass fiber filters (VWR Scientific, Radnor, PA, USA) prior to the analysis.

A UV-VIS Variable Wavelength Detector (VWD), 1260 Infinity II (Agilent Technologies), was positioned as the first detector post-separation channel, and the UV absorption of the fractionated solute was measured at 280 nm using a deuterium lamp as the light source. A Dawn Neon™ MALS device (Wyatt Technology, Goleta, CA, USA) followed, for the determination of the particle radii. Data from light scattering and UV absorption measurements were processed using Astra 8.2.0 software (Wyatt Technology). Radius determination was based on the Debye fitting formalism [44] with a 2nd-degree polynomial fit.

For online metal analysis with AF4-MALS, an iCAP^® Q ICP-MS (Thermo Scientific, Waltham, WA, USA) was integrated as a third detector alongside the setup. To enhance sample ionization into the plasma, 4% HNO₃ was introduced into the flow. Yttrium (10 ppm) served as an internal standard to monitor signal variability due to pressure differences or nebulization inconsistencies. Both acid and internal standards were added via the ICP-MS’s built-in peristaltic pump using solvent flex 3-stop tubing (Meinhard, Golden, CO, USA) with an internal diameter of 0.38 mm. The instrument operated in standard (STD) mode with no reaction or collision cells. The interface between MALS and ICP-MS is illustrated in Figure 2 and Figure 3, while the ICP-MS operating parameters are summarized in

Table 1. Operating parameters of the ICP-MS coupled to the AF4-MALS.

Plasma Torch RF Power/W	Plasma Torch Nebulizer Flow/L min−1	Plasma Torch Auxiliary Flow/L min−1	Plasma Torch Coolant Flow/L min−1	Quadrupole Dwell Time/s	Peri-Pump Flow Rate/mL min−1
1550	1.07	0.8	14	0.01	~0.19

.

Pure nanoparticle dispersion standards were measured to assess the functionality of the AF4-UV-MALS-ICP-MS connection. Details of this procedure are provided as Supplementary Information in Section S1. The results of these experiments were used to optimize the measurements of the drinking water samples.

3.2. Water Sample Analysis

Water samples from undisclosed drinking water companies were analyzed for the following parameters: TOC and IC were determined using a Shimadzu TOC-L analyzer (Shimadzu, Kyoto, Japan). ICP-OES analysis was carried out using an ICP-OES iCAP PRO (Thermo Scientific). pH and EC were measured using an HQ40d multimeter (Hach, Ames, IA, USA) with IntelliCAL PHC101 and CDC401 probes, respectively. Turbidity was measured with a turbidimeter (2100Q IS, Hach) in Nephelometric Turbidity Units (NTUs). Color was measured using a UV-1800 (Shimadzu) spectrophotometer and evaluated according to the Hazen scale. Nanometric particle presence was determined using online ICP-MS (iCAP^® Q ICP-MS (Thermo Scientific). Isotopes tracked during runs were the following: ⁷Li, ⁹Be, ¹¹B, ²³Na, ²⁴Mg, ²⁷Al, ²⁹Si, ³¹P, ³²S, ³⁹K, ⁴⁴Ca, ⁴⁸Ti, ⁵²Cr, ⁵⁵Mn, ⁵⁷Fe, ⁵⁹Co, ⁶⁰Ni, ⁶³Cu, ⁶⁶Zn, ⁷⁵As, ⁷⁷Se, ⁸⁵Rb, ⁹⁰Zr, ⁹⁵Mo, ¹⁰⁷Ag, ¹¹¹Cd, ¹¹⁸Sn, ¹³⁷Ba, ¹³⁹La, ¹⁴⁰Ce, ¹⁸²W, ¹⁹⁷W, and ²⁰⁸Pb. The samples were analyzed for the concentration of these isotopes, and the overall concentration of these elements was measured. Sample information is provided in

Table 2. List of samples provided by three undisclosed drinking water companies, with their total initial volume and information on their origin.

Company	Sample Label	Information on the Origin of the Water
A	A1	Treated groundwater (Cascade aeration → Pellet softening → Multilayer filtration)
	A2	Treated groundwater (Limited aeration → Rapid marble filtration → Aeration → Rapid sand filtration)
B	B1	Treated groundwater (Cascade aeration → Rapid sand filtration)
	B2	Treated groundwater (Cascade aeration → Rapid sand filtration)
C	C1	Bottled drinking water
	C2	Untreated mountain spring water (source for C1)

and

Table 3. Total organic carbon (TOC), inorganic carbon (IC), common metal ions, pH, conductivity, turbidity, and color of water samples A1, A2, B1, B2, C1, and C2.

Analysis	A1	A2	B1	B2	C1	C2
TOC/mg L−1	3.6	2.3	1.6	3.1	1.0	3.0
IC/mg L−1	35.3	29.1	54.1	39.4	34.9	36.8
Al3+/µg L−1	6.3	<5.00	5.32	13.7	<5.00	<5.00
Ba2+/µg L−1	14.8	15.3	32.6	17.4	56.3	56.7
Cu2+/µg L−1	2.8	2.19	68.9	3.87	27.4	30.1
Mg2+/µg L−1	6560	4360	9620	9510	>10,000	>10,000
Sr2+/µg L−1	144	129	391	289	167	174
Zn2+/µg L−1	2.1	1.93	2.32	2.02	5.45	2.77
pH	7.68	7.61	7.82	7.90	7.85	7.63
EC/µS cm−1	355.7	402.0	542.3	514.3	377.0	394.0
Turbidity/NTUs	1	1	0	1	0	2
Color (Hazen)	4	4	10	7	1	4

.

4. Results and Discussion

The light-scattering fractograms of all samples measured with AF4-MALS are presented in Figure 4.

Across all samples, late-eluting particles, which have relatively large radii (≥90 nm), dominate the signals. The intensity of the peaks varies between samples, which can be attributed to variations in the particle concentration and differences in particle sizes. In one instance (untreated mountain spring water, C2), a population of scatterers eluted early, exhibiting a strong scattering signal. Given the fact that the elution time is proportional to the size of the particles and that the intensity of Rayleigh scattering is proportional to the sixth power of the radius, this signal either indicates a very large number of very small scatterers or larger scatterers that elute early. If the former was true, the fitting algorithm should have provided a size distribution of the peak (red dots above), which was not plausible here, therefore indicating that the scatterers might not be solid particles but (intrinsic) nanobubble-associated compounds with large slipping planes [18].

In general, the samples show a wide variation in their particle size distribution, as illustrated in Figure 5. Some samples, such as B2 (treated groundwater—cascade aeration → rapid sand filtration), display a broad range of particle sizes, ranging from 90 to 240 nm for the radius. Other samples, such as A1 (treated groundwater—cascade aeration → pellet softening → multilayer filtration), yield a narrow distribution with radii between 250 and 270 nm only.

The combination of the AF4-UV-MALS with the ICP-MS allowed an elemental analysis of the different size fractions. Of the different isotopes tracked, ²⁴Mg, ²⁷Al, ⁸⁸Sr, and ¹³⁷Ba yielded signals associated with specific elution times in most samples. Additionally, ⁶³Cu and ⁶⁶Zn were detected at longer elution times in two samples. Fractograms alongside ICP-MS counts for the specified isotopes are presented in Figure 6, Figure 7 and Figure 8.

The fractograms generally indicate that most samples exhibit prominent peaks for ²⁷Al and ¹³⁷Ba eluted at approximately 27 to 30 min, aligning with the elution of the larger-sized population. These data suggest that Al and Ba are likely associated with nanometric particles and contribute to their composition, possibly bound to or incorporated in natural organic matter. This hypothesis is supported by the TOC analysis (the data are reported in the Supplementary Information, Section S2), which showed organic carbon concentrations ranging from 1.00 to 3.6 mgˑL⁻¹ in most samples. Furthermore, UV absorption (280 nm) peaks coinciding with late LS peaks further suggest the presence of organic matter at these retention times. However, possible signal interferences in the ICP-MS analysis should also be considered. For instance, for ²⁷Al, polyatomic interference with ¹²C¹⁵N⁺ ions [37] is possible. Given that the TOC analysis confirmed the presence of organic matter, additional analyses would be required to exclude this potential interference, which is beyond the scope of the present work.

At lower retention times, a consistent peak of ²⁴Mg is observed in all samples. A comparison with the blank (Figure 9) reveals a sharp void peak and a distinct ²⁴Mg peak. This peak could be indicative of magnesium-containing DOLLOPs. It is interesting to note here that sample C2, untreated mountain spring water, shows the presence of magnesium at the same early dilution times where it yielded a MALS peak probably associated with bubbles (see also Figure 3). Assuming that the magnesium peak stems from DOLLOPs containing magnesium, this result could be an indication that DOLLOPs agglomerate on charged nanobubble surfaces, as suggested recently [17]. It should be mentioned here that Calcium ions, which should be predominantly present in this peak if that was the case, could not be measured in the current configuration due to interference with the ⁴⁰Ar⁺ isotope from the argon used as the ICP plasma gas.

However, it is challenging to actually determine whether this peak is associated with nanometric particles or ionic Mg²⁺. The absence of an LS peak at this retention time in most samples suggests against the nanoparticle hypothesis, but the higher sensitivity of ICP-MS compared to LS detectors leaves open the possibility of very small particles or magnesium-containing DOLLOPs eluting undetected by MALS unless they agglomerate on charged bubbles. The ICP-MS configuration available could not distinguish between ionic and nanoparticle signals, limiting the ability to ascertain the nature of the ²⁴Mg peak.

A similar trend is observed for ⁸⁸Sr, suggesting that Strontium ions might be involved in the DOLLOP formation as well. An example from sample B1, treated groundwater, is shown in Figure 10.

The intensity of the ⁸⁸Sr signals compared to the blanks was generally less pronounced than for ²⁴Mg. Additionally, the ⁸⁸Sr peaks appeared at slightly later elution times. In some cases (e.g., samples A2, aerated and treated groundwater, and C1, bottled drinking water), sharp peaks were observed alongside larger particles, similar to the signals for ²⁷Al and ¹³⁷Ba. The ⁶³Cu and ⁶⁶Zn signals for sample C1, bottled drinking water, are shown in Figure 11 and Figure 12.

Both metals appeared at later retention times, coinciding with the elution times of ²⁷Al and ¹³⁷Ba whenever detected. Quantitative analysis via ICP-OES (Table 3) confirmed the presence of all mentioned elements in the samples. However, in some cases (specifically, ²⁷Al in samples A2, treated and aerated groundwater, and C1, bottled drinking water), signals detected by online ICP-MS were below the detection limit of ICP-OES. This raises the possibility of polyatomic interference for the ²⁷Al signal, but it could also be due to the superior detection limits of ICP-MS compared to ICP-OES, allowing only the online ICP-MS to detect aluminum.

5. Conclusions

AF4-MALS is a powerful tool for the separation and characterization of submicrometric and nanoparticles in environmental water samples. Its operational flexibility allows for optimization to specific samples, accommodating unique characteristics such as concentration, surface charge, and size distribution. Although method development can be time-consuming due to the numerous variables involved, the potential for customization makes AF4-MALS a highly adaptable technique. Coupling AF4-MALS with detectors such as ICP-MS significantly enhances its effectiveness, enabling the precise identification of metallic nanoparticles and metal-associated colloidal matter. In this study, online ICP-MS provided valuable insights into the interactions of metallic particles as they traversed the AF4 channel, effectively differentiating particles by detecting specific metal isotope signals. While MALS accurately measures particle sizes, limitations can arise with very small particles (<10 nm) due to the small scattering intensity of these sizes. In general, our study successfully demonstrates the capability of AF4-MALS-ICP-MS to detect metals such as Al, Ba, Cu, Zn (associated with larger particles around 200 nm), and Mg (associated with very small particles between 1.5 and 10 nm) in different drinking water samples, thereby shedding light on the distribution and composition of (natural) nanometric components of drinking water. Additional Total Organic Carbon (TOC) analysis indicates the presence of natural organic matter, which is physically or chemically binding to these metals.

Overall, this study highlights the robust capabilities of AF4-MALS, particularly when coupled with ICP-MS, in providing comprehensive insights into the nature and interactions of nanoparticles in environmental water samples. The technique’s adaptability and precision underscore its potential for broader applications in environmental monitoring and analysis.