Aquatic Colloids Research Articles

Laser‐induced Breakdown Detection (LIBD) for the Highly Sensitive Quantification of Aquatic Colloids. Part I: Principle of LIBD and Mathematical Model

AbstractA new type of Laser‐induced Breakdown Detection (LIBD) is being developed for the direct quantification of aquatic colloids of extremely small sizes at very low concentrations, which are not detectable by the commercial methods currently available. LIBD allows for the detection of colloids in the lower nanometer range. Different statistical and mathematical models are available for the quantification of colloids in aquatic systems. In this double publication (parts I and II), a model based on binomial statistics was employed. This is introduced as the basis of our theoretical descriptions of the experimental data discussed in part II. As a prerequisite, a qualitative and quantitative description of the power density distribution in the focal area of a laser beam is necessary and thus is also presented in this article. Together with the second part, this paper is intended to provide a profound insight into both the theoretical and experimental aspects of the Laser‐induced Breakdown Detection technique, and to show practical applications as well as the latest developments in this field of particle characterization.

Characterization of Freshwater Natural Aquatic Colloids by Atomic Force Microscopy (AFM)

Atomic force microscopy (AFM) has been used to image and quantify riverine colloids in a quantitative and relatively nonperturbing manner. Three main classes of material have been imaged including fibrils (about 10 nm in diameter and 100 nm or more in length), discrete, near-spherical, small colloids primarily below 30-50 nm in diameter, and a surface film, of at least several nanometers thickness, which coatsthe entire mica surface within 30 min of exposure to river water. Colloid structure was found to vary as a function of pH, particularly at high pH. Substantially different structures were observed at high pH values, with the loss of the near-spherical colloids possibly due to rearrangement and aggregation. In addition,film thicknesses of up to 100 nm were estimated on the silicon nitride AFM cantilever after 30 h of deposition in the same water (unperturbed and size fractionated). The observation of these surface films has important implications for understanding the mechanisms by which colloids might bind trace elements. In particular, development of surface coatings implies that binding of pollutants (at least initial surface binding) may be dominated by adsorbed surface layers.

Characterization of sewage plant hydrocolloids using asymmetrical flow field-flow fractionation and ICP-mass spectrometry

Asymmetrical flow field-flow fractionation (AF 4) was applied to characterize aquatic colloids from biological sewage plants and to infer information of colloidal loads, sources, and sinks within the plants, resp. the colloidal interaction with the aqueous phase and the sewage sludge. To characterize the colloids further, especially the distributions of colloid associated heavy metals, the AF 4 system was coupled to an inductively coupled plasma mass spectrometer (ICP-MS). The size distribution is determined by AF 4 with UV absorbance and fluorescence detection after a calibration by monodisperse polystyrene sulfonate standards (PSS). Samples from different sewage plants and from different depths and locations within a plant were compared. The fulvic/humic acid fraction with a particle diameter d p<10 nm appeared to be comparable in all samples and decreases only slightly along the plants, whereas larger colloids with d p>10 nm almost completely passed into the sewage sludge. The concentrations of the initial colloidal heavy metals decreased along the plants.

Assessment of cross-flow filtration for the size fractionation of freshwater colloids and particles

This research has evaluated the ability of cross-flow filtration (CFF) to perform correct size fractionation of natural aquatic colloids (materials from 1 nm to 1 μm in size) and particles (>1 μm) using scanning electron microscopy (SEM) combined with atomic force microscopy (AFM). SEM provided very clear images at high lateral resolution (ca. 2–5 nm), whereas AFM offered extremely low resolution limits (sub-nanometer) and was consequently most useful for studying very small material. Both SEM and AFM were consistent in demonstrating the presence of colloids smaller than 50 nm in all fractions including the retentates (i.e. the fractions retained by the CFF membrane), showing that CFF fractionation is not fully quantitative and not based on size alone. This finding suggests that previous studies that investigated trace element partitioning between dissolved, colloidal and particulate fractions using CFF may need to be re-visited as the importance of particles and large colloids may have been over-estimated. The observation that ultra-fine colloidal material strongly interacted with and completely coated a mica substrate to form a thin film has important potential implications for our understanding of the behaviour of trace elements in aquatic systems. The results suggest that clean, ‘pure’ surfaces are unlikely to exist in the natural environment. As surface binding of trace elements is of great importance, the nature of this sorbed layer may dominate trace element partitioning, rather than the nature of the bulk particle.

Partition of Endocrine-Disrupting Chemicals between Colloids and Dissolved Phase As Determined by Cross-Flow Ultrafiltration

Cross-flow ultrafiltration (CFUF) was developed first for the isolation of natural colloids and subsequently for determining the partition of selected endocrine-disrupting chemicals (EDCs) between river colloids and dissolved phase. In this study, a 1-kDa Millipore Pellicon 2 cartridge type CFUF system was validated using a range of molecular probes spiked in natural waters. Results show that good retention (>80%) of high molecular weight (HMW, >1 kDa) molecules and low retention of low molecular weight (LMW, <1 kDa) molecules can be achieved at high concentration factor (cf) values in sampling mode or over long time scales in recirculation mode. The interactions between aquatic colloids and EDCs were studied by mixing EDCs, water, and colloids previously isolated by CFUF for a certain duration, followed by the separation of the target compounds between the truly dissolved and colloid-bound phases by CFUF and analysis by gas chromatography-mass spectrometry (GC-MS). The kinetics of EDCs binding to colloids were relatively rapid, reaching equilibrium within 5 min. The mass balance of chosen EDCs through CFUF system was fully investigated, with good recovery for the relatively polar EDCs such as estrone and 17beta-estradiol. On the basis of EDC sorption by colloids, the partition coefficient normalized to colloidal organic carbon content (Kcoc) was 8.85 x 10(3), 1.50 x 10(4), 8.85 x 10(3), 4.87 x 10(4), and 1.59 x 10(4) mL/g for bisphenol A, estrone, 17beta-estradiol, 17alpha-ethynylestradiol, and 16alpha-hydroxyestrone, respectively, which are comparable with the values reported in the literature. In addition, it has been shown that the Kcoc values of EDCs were relatively independent of their octanol-water partition coefficient (Kow) values, suggesting the important role of different binding mechanisms other than nonspecific hydrophobic interaction between EDCs and natural colloids. As the CFUF-GC-MS method can be used to quantify very low concentration of pollutants and is not limited to fluorescent compounds, it has the potential to be a widely applicable separation/analytical tool for determining the partition of organic pollutants between colloidal materials and dissolved phase.

Visualisation of natural aquatic colloids and particles ? a comparison of conventional high vacuum and environmental scanning electron microscopy

The applicability of environmental scanning electron microscopy (ESEM; imaging of hydrated samples) and conventional high vacuum scanning electron microscopy (SEM; imaging of dried samples at high vacuum) for the observation of natural aquatic colloids and particles was explored and compared. Specific attention was given to the advantages and limitations of these two techniques when used to assess the sizes and morphologies of complex and heterogeneous environmental systems. The observation of specimens using SEM involved drying and coating, whereas ESEM permitted their examination in hydrated form without prior sample preparation or conductive coating. The two techniques provided significantly different micrographs of the same sample. SEM provided sharper images, lower resolution limits (10 nm or lower), but more densely packed particles, suggesting aggregation, and different morphological features than ESEM, suggesting artefacts due to drying. ESEM produced less easily visualised materials, more complex interpretation, slightly higher resolution limits (30-50 nm), but these limitations were more than compensated for by the fact that ESEM samples retained, at least to some extent, their morphological integrity. The results in this paper show that SEM and ESEM should be regarded as complementary techniques for the study of aquatic colloids and particles and that ESEM should be more widely applied to aquatic environmental systems than hitherto.

Measuring multimodal size distributions of aquatic colloids at trace concentrations

Many applications in research as well as industry require a highly sensitive detection of particulate matter in water or process liquids. We present a technique to measure the distribution of colloid sizes between 20 and 100nm in situ at ultratrace concentrations. The method is based on laser-induced breakdown detection, which has been applied for trace detection of colloids and determination of the weighted mean size. Using a refined method of data evaluation, we are now able to measure the number density of inorganic colloids in six size classes between 20 and 100nm simultaneously below ppb concentration.

Detection of aquatic colloids in drinking water during its distribution via a water pipeline network

Laser-induced Breakdown Detection (LIBD) is a highly sensitive method for the direct detection of nano-particles (colloids). During the detection process plasmas are generated on single particles by a focused laser beam, the resulting plasma light emissions are detected optically. The method is based on the difference in breakdown thresholds of liquid and solid matter, it is lower for solid material. The laser pulse energy is adjusted precisely so that in the pure liquid no breakdown events occur, and only in the presence of colloids is the breakdown threshold in the focal volume exceeded. The spatial distribution of several thousand recorded plasma flashes within the focal volume reveals the mean particle diameter. The evaluation of the number of breakdown events per number of laser pulses results in a breakdown probability, together with the particle size the concentration is calculated using specially-designed computer software. Compared to conventional laser light scattering methods the LIBD is approximately 6 orders of magnitude more sensitive for particles smaller than ca. 0.05 pm. Together with laser light obscuration the LIBD technique has been used successfully for the quantification of aquatic nano-particles during drinking water processing and its distribution via a pipeline network of nearly 1,700 km total length. In addition, the particulate content of several brands of mineral water has been investigated.

Size fractionation of aquatic colloids and particles by cross-flow filtration: analysis by scanning electron and atomic force microscopy

The suitability of the combined application of environmental scanning electron microscopy (ESEM), scanning electron microscopy (SEM) and atomic force microscopy (AFM) for the evaluation of the ability of cross-flow filtration (CFF) to perform adequate size fractionation of freshwater colloids and particles was examined. ESEM and SEM imaging provided reference images of the CFF-generated fractions and, in estimating the experimental cut-off diameter of the membrane, provided evidence that separation was not consistent with nominal pore sizes of the membranes. However, analysis of the images showed that size distribution of CFF-generated fractions and the estimated cut-off diameter of the membranes were dependent on the advantages and limitations of the two imaging techniques. With both ESEM and SEM, best estimates of size cut-offs were lower than the nominal pore size of the membrane in the case of 0.45 μm membranes, but roughly accurate in the case of 0.1 μm pore size membranes. The results also suggested that the effectiveness of CFF may benefit from a pre-separation step using a minimally perturbing technique such as split thin-flow fractionation. AFM demonstrated the presence of colloids smaller than 50 nm in all fractions including the retentates, showing that CFF fractionation is not fully quantitative and not based on size alone. The results indicate that previous studies investigating trace element partitioning using CFF may need re-evaluation as the importance of particles and large colloids may be over-estimated.

Measurement of the size and structure of natural aquatic colloids in an urbanised watershed by atomic force microscopy

Atomic force microscopy (AFM) in tapping mode was used to determine the conformation of humic substances and aquatic colloids from rivers in an urban catchment in the West Midlands, U.K. Hurnie macromolecules were shown to have a size of about 1–3 nm in agreement with the literature, indicating that the preparation methods and the AFM were both performing satisfactorily. Three types of natural aquatic colloids were observed by AFM. Firstly, a surface coating about 1–5 nm thick, likely composed of organic and oxide material flattened by drying and interaction with the AFM tip. Secondly, small irregular, globular material between 1 and 70 nm in size, again most likely made of oxide and organic material. Lastly, fibrillar material was present which was 1–10 nm in diameter and 10–1000 nm in length. Most likely this material was microbially produced (muco-) polysaccharides. Size distributions of colloids from all samples, regardless of sample site and sample preparation, indicated colloids with a fairly low polydispersity and with particle numbers dominated by material < 10 nm.

Aquatic colloids relevant to radionuclide migration: characterization by size fractionation and ICP-mass spectrometric detection

The application of the flow-field flow fractionation (FFFF) combined with on-line ICP-mass spectrometry (ICP-MS) to the characterization of aquatic colloids is described. The capabilities and drawbacks of the technique are discussed with the aid of two examples. (1) The size distribution of smectitic colloids dispersed from a natural bentonite is determined by FFFF and laser light scattering (LLS) and compared with the size information obtained by ICP-MS detection. Due to the pronounced size dependency of the scatter light intensity, the LLS detection tends to overestimate the larger sized particles. Therefore, the FFFF-ICP-MS fractogram delivers more reliable size information. (2) Groundwater humic/fulvic colloids and the humic matter extracted from the corresponding sediment, both taken from the Gorleben aquifer (Lower Saxony, Northern Germany), are analysed by FFFF-ICP-MS. The location of REE, U, Th and Ca in different colloid size fractions appears to be very similar in both samples. The element specific fractograms suggest in agreement with earlier studies the location of Th and the REE mainly in inorganic colloids>17 nm containing also Fe and/or Al. U and Ca appear to be distributed between larger colloids and the fulvic/humic acid fraction with a size <3 nm. The results demonstrate that even in groundwater/sediment systems with high humic/fulvic content, inorganic colloids may play an important role as carrier for polyvalent metal ions. The consequences of this finding on the applicability of in-situ K d values for U, Th and REE, taken as naturally abundant representatives of the nuclear waste derived actinides, for the assessment of laboratory sorption data is discussed.

Characterization of Gorleben groundwater colloids by atomic force microscopy.

Groundwater colloids from the Gorleben site (Lower Saxony, Germany) are characterized in the presence of Eu(III) by tapping-mode atomic force microscopy (AFM) with phase contrast imaging. Using a liquid cell the method allows investigations of samples being in contact with aqueous solution. This ensures that complex structures are kept in their native hydrated state. Different types of colloids and aggregates are found by AFM, e.g., spherical particles, fibrous structures, and structures which appear to be hollow. A partial coating of the edges of clay particles with humic colloids can be assumed from phase contrast images. Therefore, aquatic colloids and their aggregates found in Gorleben groundwater can be characterized as a complex mixture of components, which may influence the migration of groundwater contaminants in different processes.

Characterization of aquatic colloids by a combination of LIBD and ICP-MS following the size fractionation

SummaryHumic colloids in deep groundwater are characterized in order to ascertain how heavy metal ions of chemical homologue to actinides are associated in different size fractions. The colloid size fractionation is made by two different methods: flow field flow fractionation (FFFF) and size exclusion chromatography (SEC), which is followed by analysis of chemical composition using UV spectroscopy for organic components and ICP-MS for inorganic components. Relative number density of humic colloids following the size fractionation is determined by laser-induced breakdown detection (LIBD). For the appraisal of colloid size change upon metal ion complexation, purified Aldrich humic acid is loaded with the Eu3+ion on increasing the concentration and the size change is then determined by LIBD. Humic colloids are interacted with radioactive tracers,155Eu(III) and228Th(IV), to appreciate their sorption behaviour onto different colloid size fractions and thus to compare with natural humic colloid-borne elements of M(III) and M(IV).Whereas the organic component of humic colloids is found in the size range about 3 nm, the inorganic components of actinides homologues, M(III) and M(IV), are observed in the fractionated size range from 10 nm to 35 nm. This observation leads us to presume that the inorganic composites, composed of heavy metal elements, are peptised by humic/fulvic acid. The laboratory traced elements,155Eu(III) and228Th(IV), are quantitatively sorbed only onto the organic component. This fact infers the complexation of traced metal ions with humic components of colloids without visible interaction with the inorganic components. Once purified humic acid is complexed with Eu(III) in an excess amount, the size increases over 200 nm and the resulting products undergo flocculation or precipitation. The results of the present experiment indicate that the natural humic colloid-borne heavy metal ions appear in different chemical states as compared to the same but laboratory traced metal ions. As a result, the laboratory simulation entails a better understanding of the underlying molecular level chemical reactions, which are important to the assessment of the colloid-borne actinide migration.

Laser induced breakdown detection for the assessment of colloid mediated radionuclide migration

Colloids play an important role in the transport of pollutants in the environment. Harmful substances can undergo transport over large distances if bound to colloids in aqueous surrounding. One important example is the migration of Pu(IV) at unexpectedly high rates over several miles at a Nevada nuclear detonation test site. For long term safety assessments of radioactive waste repositories, it is hence crucial to know about the amount, size distribution and chemical composition of colloids in the ground water. Standard methods (e.g. light scattering) can be applied for high concentrations and large sizes of particles. Colloids smaller than 50 nm, however, are detected with very low efficiency. Laser induced breakdown detection (LIBD) can fill this gap. A new instrumentation is presented, which as compared to previous instruments, opens up a much wider operational dynamic range, now covering three orders of magnitude in size (5–1000 nm) and seven orders of magnitude in particle concentration (1 ppt – several ppm). The technique is based on plasma formation on colloidal particles inside the focus of a pulsed laser. The plasma plume is detected by three-dimensional optical observations and by means of its shock wave with a piezo-detector. For mathematical modelling, detailed knowledge on the photon fluence distribution in the focal region is indispensable. For the first time a true Gaussian TEM00 mode has been achieved in the focus of a LIBD apparatus and great care has been taken to guarantee long-term stability of the optical parameters. Automated control of the laser pulse energy and beam shape is introduced to allow routine reproducible measurement. The apparatus combines acoustic detection with three-dimensional optical monitoring of the focal region with two CCD cameras placed perpendicular to each other in order to gain additional size information. The breakdown events are systematically characterized with respect to the number density and size of aquatic colloids as a function of the laser pulse energy. Whereas the threshold energy (irradiance) only depends on the colloid size, the breakdown probability at higher pulse energies is a direct function of the number density of colloids. A correlation of the two facts allows the speciation of the colloidal size distribution.

A mobile laser-induced breakdown detection system and its application for the in situ-monitoring of colloid migration

Abstract A mobile Laser-Induced Breakdown Detection (LIBD) system is developed for the field study of the aquatic colloid migration. The principle of LIBD is based on the plasma generation at breakdown of colloids in the laser focus region, which is then observed by either photo-acoustic detection or optical monitoring. The method provides information on the number density and average size of colloids. The present instrumentation built up particularly for the field study is applied for the aquatic colloid migration in a natural granite fracture. The study is performed within the Colloid and Radionuclide Retention project (CRR) in order to investigate the role of colloids on the radionuclide migration. For this purpose, bentonite colloids dispersed in granite groundwater are injected into a rock fracture zone with a dipole distance of 5 m. After passing through the fracture, the groundwater is guided into a flow-through detection cell of LIBD for the quantification of colloids. Another experiment is performed with the same colloids traced by tetra- and tri-valent metal ions, Th(IV), Hf(IV) and Tb(III). The recovery of colloids as found by LIDB appears to be about 55±5% of the injected colloid mass concentration, which is corroborated by ICP-MS analysis of Al present in bentonite colloids. An average size of recovered colloids is slightly decreased from the value of before injection and hence the relative number density increased in the course of the rock fracture migration. Recovered fractions of colloid-borne trace elements in the extracted groundwater are 78±8% for both tetra-valent elements and 33±3% for trivalent element of the originally traced concentrations. Plausible explanations are given for the different recoveries of heavy metal ions.

Laser‐induced breakdown spectroscopy for soil diagnostics

SummaryLaser‐optical measurements and fibre optics are potentially attractive tools for applications in soil science because of their great sensitivity and selectivity and their capabilities for on‐line and in situ analysis. We have investigated laser‐induced breakdown spectroscopy (LIBS) for the quantitative detection of metal ions on the surface of natural soil samples from two sites (Hohenschulen and Oderbruch, Germany). The LIBS technique allows the spatially resolved investigation of adsorption and desorption effects of ions in soil. A frequency doubled (532 nm) and Q‐switched Nd:YAG laser with a pulse duration of 8 ns is focused on the soil surface and induces a plasma. Typical power densities are 150 mJ mm−2. The plasma emission is recorded in time and spectrally resolved by a gateable optical multichannel analyser (OMA). A delay time of about 500 ns between laser pulse and OMA gate was used to resolve single atomic and ionic spectral lines from the intense and spectrally broad light that is emitted by the plasma itself. The dependency of the LIBS signal of a single spectral line on the amount of water in the sample is investigated in detail. The results indicate that quenching of water in the plasma plume reduces the line intensities, while the interaction with aquatic colloids increases the intensity. The two processes compete with each other, and a non‐linear correlation between measured line intensities and the amount of water in the sample is obtained. This is verified by a simple computer simulation and has to be taken into account for the quantitative interpretation of LIBS signals, e.g. when absolute concentrations are estimated. In the present investigation natural calcium concentrations &lt; 2 μg kg−1 were measured with the LIBS technique in the samples for the two test sites. In addition, measurements were made with dry and water‐saturated BaCl2 mixed soil samples, and no significant difference in the detection limit for barium was obtained.

Ultrafiltration behavior of major ions (Na, Ca, Mg, F, Cl, and SO 4) in natural waters

Aquatic colloids, including macromolecules and microparticles, with sizes ranging between 1 nm to 1 μm, play important roles in the mobility and bioavailability of heavy metals and other contaminants in natural waters. Cross-flow ultrafiltration has become one of the most commonly used techniques for isolating aquatic colloids. However, the ultrafiltration behavior of chemical species remains poorly understood. We report here the permeation behavior of major ions (Na, Ca, Mg, F, Cl, and SO 4) in natural waters during ultrafiltration using an Amicon 1 kDa ultrafiltration membrane (S10N1). Water samples across a salinity gradient of 0–20‰ were collected from the Trinity River and Galveston Bay. The permeation behavior of major ions was well predicted by a permeation model, resulting in a constant permeation coefficient for each ion. The value of the model-derived permeation coefficient ( P c) was 0.99 for Na, 0.97 for Cl, and 0.95 for F, respectively, in Trinity River waters. Values of P c close to 1 indicate that retention of Na, Cl, and F by the 1 kDa membrane during ultrafiltration was indeed minimal (<1–5%). In contrast, significant (14–36%) retention was observed for SO 4, Ca, and Mg in Trinity River waters, with a P c value of 0.64, 0.82, and 0.86 for SO 4, Ca and Mg, respectively. However, these retained major ions can further permeate through the 1 kDa membrane during diafiltration with ultrapure water. The selective retention of major ions during ultrafiltration may have important implications for the measurement of chemical and physical speciation of trace elements when using cross-flow ultrafiltration membranes to separate colloidal species from natural waters. Our results also demonstrate that the percent retention of major ions during ultrafiltration decreases with increasing salinity or ionic strength. This retention is largely attributed to electrostatic repulsion by the negatively charged cartridge membrane.

Determination of colloid size by 2-D optical detection of laser induced plasma

A new method is presented for the size determination of aquatic colloids, which is based on the selective plasma generation (breakdown) on individual particles by a focused laser pulse. For this purpose, the breakdown events of particles are monitored by 2-D optical detection instead of commonly used photoacoustic detection. The length of the breakdown volume determined for the 2-D projection of spatially distributed plasma events generated on colloidal particles is proportional to the particle size. Based on this fact, an average diameter of a given colloidal dispersion can be determined. The method is calibrated by well-characterised hard-sphere polystyrene reference particles of different diameters.

Laser-induced breakdown detection (LIBD) of aquatic colloids with different laser systems

The Laser-Induced Breakdown Detection (LIBD) is a new, sensitive method for the quantification of aquatic colloids of lower range nanometer size in very low concentration, which can not be detected by conventional light scattering methods. The method is investigated by two different laser setups: an excimer pumped dye laser system and a Nd-YAG laser system. In this paper these two different setups for the LIBD are described and compared. Polystyrene reference particles with diameter from 1000 nm down to 19 nm are used for the investigation. The experimental data can be described by a model based on binomial statistics. The Nd-YAG laser system leads to a significant technical improvement in colloid measurement and the limit of detection for particle concentration is better than for the excimer-dye laser system. The model calculation on the experimental data results in a minimum detectable particle radius of r min=0.6±0.3 nm, which is significantly lower than the value of r min=9.8±2.0 nm found for the excimer-dye laser system. Such an improvement makes LIBD an even more powerful analytical tool for the investigation of aquatic colloids in the size range <50 nm.

Fractionation and composition of colloidal and suspended particulate materials in rivers

The association of pollutants (nutrients, heavy metals and organic compounds) with colloidal and suspended particle matter (SPM) plays a dominant role in determining their transport, fate, biogeochemistry, bioavailability and toxicity in natural waters. A scheme for the fractionation and composition of colloidal and SPM from river waters has been tested. All four separation methods, i.e. sieving, continuous flow centrifugation, tangential flow filtration, sedimentation field-flow fractionation, were for the first time used to separate five size particulate fractions from river. Significant (gram) amounts of colloidal material (<1 μm) in three size ranges, nominally 1–0.2, 0.2–0.006 and 0.006–0.003 μm were obtained. The separation scheme was able to process large samples (100 l), within reasonable times (1 day) and the apparatus was portable. The aquatic colloid size was also characterized with high resolution by using sedimentation field-flow fractionation technique. The mass-based particle size distribution for the water sample showed a broad size distribution between 0.05 and 0.4 μm with the maximum around 0.14 μm. There was a systematic increase in the content of organic carbon (estimated by loss on ignition), Mg, Ca, Na, Cu and Zn with decreasing particle size, highlighting the importance of the colloidal (<1 μm) fraction. It was concluded that the colloidal Cu and Zn concentrations in rivers might be much higher than those reported before.