The presence of colloids in natural waters is one of the major reasons for discrepancies between experimental findings and theoretical predictions of contaminant transport via the water path. Here, particles of 1 nm to 1 µm are regarded as colloidal particles. We developed a scheme of separation and detection for colloidal particles which is applicable to very different natural waters. A minimum of influences on the samples and a maximum of unambiguousness of the measurements is striven for by this scheme. Separation steps are performed as mildly as practicable or, if possible, avoided at all. Separation procedures are centrifugation and mild filtration. Detection methods are scattered light intensity measurements, photon correlation spectroscopy (PCS), ICP-MS, AAS, TOC analysis and imaging by scanning electron microscopy (SEM), i.e., non- invasive as well as invasive methods. An important element of this scheme is the parallel application of as many complementary methods as possible. It allows to achieve validation of the results "against each other" and increases the probability to recognize artifacts.
Our first step in solving a colloid characterization problem is always a non-invasive particle size measurement by PCS on the raw sample. The next operation is a PCS measurement on a 5-µm filtrate of this sample. Filtration through 5-µm Nuclepore filters improves the counting statistics for PCS but has usually only little influence on the colloidal inventory of the sample. Fig. 1 shows PCS results for a water sample from a mine drainage tunnel (colloid concentration 1 mg/L).
Fig. 1: Particle size distribution in water from the drainage tunnel Rothschönberger Stolln (Freiberg) according to PCS. (a) Raw sample. (b) 5-µm filtrate. Colloid concentration: about 1 mg/L. Particles of 50 to 300 nm with the peak maximum at 180 nm are found (c.f. ).
Particles of about 100 nm as those indicated in Fig. 1 can also easily be visualized by SEM when laying on a Nuclepore filter. Examples from a mine drainage tunnel water sample, an acid rock drainage (ARD) sample, a bog water sample and a backwater sample from a sanitary landfill are given in Figures 2 through 5 [1-4].
Fig. 2. SEM micrograph and EDX spectrum of an agglomerate of colloidal particles from the mine drainage tunnel Rothschönberger Stolln (colloid concentration: about 1 mg/L) laying on a 5-µm Nuclepore filter. Scale bar: 2 µm. The filter cake was washed three times, dried and coated with carbon. The aggregate consists of iron/aluminum oxyhydroxide particles of 100 to 300 nm that carry toxic heavy metal contaminants. In the solution, these particles move freely, i.e., the aggregate results from the filtration process (c.f. ).
Fig. 3: SEM micrograph and EDX spectrum of an agglomerate of colloidal particles from acid rock drainage (ARD) from Freiberg (colloid concentration: about 1 g/L) laying on a 5-µm Nuclepore filter. Scale bar: 2 µm. The filter cake was washed three times, dried and coated with carbon. The aggregate is formed of particles of 70 to 250 nm. Chemically, these particles consist of iron oxyhydroxides which also contain toxic heavy metal contaminants. In the solution, these particles move freely, i.e., the aggregate results from the filtration process (c.f. ).
Fig. 4: SEM micrograph of humic particles from the mountain bog Kleiner Kranichsee (colloid concentration: 140 mg/L) laying on a 100-nm Nuclepore filter. Scale bar: 500 nm. The filter cake was washed three times, dried and coated with carbon. The aggregate consists of particles of about 200 nm. In the solution, these particles move freely, i.e., the aggregate results from the filtration process (c.f. ).
Fig. 5: SEM micrograph of inorganic particles from the backwater of a sanitary landfill at Freital, Saxony (colloid concentration: 1 to 2 mg/L) laying on a 15-nm. Scale bar: 2 µm. The filter cake was washed three times, dried and coated with carbon. It consists of particles of 50 to 300 nm. In the solution, these particles move freely, i.e., the aggregate in the center of the figure results from the filtration process (c.f. ).
In all these samples we found good agreement between the PCS results and the SEM particle sizes. Whereas it is not possible to decide from the SEM micrographs alone if the micron-sized aggregates visible on the filters exist also in the solutions, the in-situ measurements by PCS prove that these aggregates are formed by the filtration process. The nanoparticles of the 100- nm size range move usually freely (independently of each other) in the unperturbed solution. The minimum colloid concentration detectable by PCS for particles of about 100 nm is 10 to 100 µg/L. The chemical composition of such particles can be determined by filtration and by ICP-MS and EDX of the filter cakes.
A significantly more difficult problem is the non-invasive or little-invasive determination of particles of only few nanometers in diameter. The minimum concentration detectable by PCS is much higher for such particles because these particles scatter very little light. The presence of only few larger particles prevents PCS because they optically mask the small particles. The small particles can be unmasked by removing the larger particles using filtration or centrifugation. An unmasking experiment on an ARD sample (colloid concentration about 1 g/L) is shown in Fig. 6 (cf. ).
Fig. 6: Particle size distribution in an acid rock drainage (ARD) sample from Freiberg (colloid concentration: about 1 g/L) according to PCS. (a) Raw sample. (b) 5-µm filtrate. (c) 400-nm filtrate. (d) 50-nm filtrate. Filtration through sufficiently small filter pores results in the unmasking of the ultrafine colloid particles (c.f. ).
Light scattering intensity measurements, PCS, and chemical analyses by ICP-MS, TOC analysis, ion chromatography etc. should be combined to identify the colloid inventories and the colloid composition/mineralogy of such complex colloid mixtures in centrifugation or filtration experiments, and one should always be aware of the artifacts that can be caused by centrifugation and especially by filtration (self-coagulation, clogging, adsorption etc.). Fig. 7 demonstrates the particle size characterization of organic particles in a bog water sample down to about 1 nm by ultrafiltration (concentration of the organic particles: about 14 mg/L) .
Fig. 7: Ultrafiltration of a water sample from the bog Kleiner Kranichsee, Saxony. Colloid concentration: about 14 mg/L. Decreasing amounts of humic material are passing through the ultrafilters with decreasing molecular weight cut-off. The sums of the filtrate and the retentate concentrations show that the recovery of the ultrafiltration process is reasonable (the poorest recovery was found for the 30-kD ultrafilter). Cf. .
Also extremely fine iron oxyhydroxide particles in ARD (colloid concentration about 1 g/L)  and fine particles in wood degradation products and in lignin solutions (colloid concentration about 2 g/L)  could be characterized by ultrafiltration. The visualization of particles of less than 10 nm in size is possible by transmission electron microscopy (TEM) and atomic force microscopy (AFM). However, sample preparation is tedious for these methods; the techniques, although challenging and very promising, are hardly suited as routine techniques. An example is given in Fig. 8. It shows the result of an experiment in which we could visualize the individual molecules of a humic acid solution spin-coated onto mica by AFM (collaboration with the Dresden University of Technology). SEM is not suited to visualize such particles because of lack of resolution power.
Fig. 8: AFM image of humic acid deposited on mica by means of spin-coating. Concentration of the spin-coated humic acid solution: 200 mg/L, pH value: 11.3. Scan size: 1 µm x 1 µm. Elongated agglomerates (A - 250 nm x 60 nm) and disk-like agglomerates (B - 50 nm diameter) are visible. Height of the deposits: 1.5 to 2 nm. Periodicity of the substructure: 12 to 14 nm. The agglomerates are monolayer clusters of 'humic acid subunits'. The 'subunits' could be identified as the individual humic acid molecules. AFM image taken by M. Mertig, Dresden University of Technology.
A real challenge to our separation and detection scheme is the investigation of extremely fine low-concentration colloids (few nanometers and few mg/L). Particle size determination by PCS fails in such cases because of poor counting statistics due to low scattered light intensities and because of masking problems. Ultrafiltration is often disturbed by adsorption problems at the filter membranes. For inorganic particles (Fe and Al oxyhydroxide particles), we reached the lowest detection limits by centrifugation. Using centrifugal accelerations of up to 46 000 g and centrifugation times of up to 10 h, we were able to classify colloidal particles down to about 5 nm in diameter at concentrations of about 0.5 mg/L . Fortunately, low-concentration colloids do usually not play an important role in contaminant transport via the water path as has for instance been demonstrated for groundwaters from crystalline rock formations in Switzerland . The surface area of the particles of low-concentration colloids is too low in comparison to the available rock surface to take real influence. Extremely fine particles in particular tend to be low-concentrated in the nature due to high instability if they are not stabilized electrostatically (as it is the case in humic acid solutions or in ARD solutions). This instability is caused by the very fast coagulation kinetics of environmental particles of the lower nanometer size range .
|Institute - Zänker, Harald; Hüttig, Gudrun; Richter, Wolfgang; Brendler, Vinzenz; 18.1.2000|