Ph. D.-Theses (since 1990) : Summaries

Analysis of nonconducting materials by means of GDMS using the secondary cathode technique. WIM SCHELLES, 1996

Glow discharge mass spectrometry (GDMS) has developed in the past decade as an important tool for (ultra)trace analysis of solid conducting samples. Its main advantages are the low limits of detection (down to sub ppb level), its rather uniform elemental sensitivity and its ability to measure all elements and even isotopes (unlike, for example, X-ray techniques). The sample is exposed to a low pressure discharge and is actually sandblasted, on atomic scale, by particles (ion and neutrals) present in this plasma.
Because the material under investigation acts as a cathode in the discharge, one of the main challenges for glow discharge spectroscopy is to overcome the intrinsic requirement of the sample to be electrically conducting. The above mentioned interesting analytical figures of merit and the increasing need for nonconductor analysis (e.g. ceramics) are major motivations to investigate possible ways to extend the applicability of GDMS towards insulating materials. Up to now, three approaches have been used for nonconductor analysis.
In the case of powder samples, one can bring the nonconductor into a high purity conducting host matrix (e.g. copper powder) and electrodes can be pressed from the resulting powder mix. However, when the sample to be analyzed in investigation is a solid material, like a plate of sintered ceramic material, grinding can be troublesome and, moreover, can cause serious contamination. For these sample types two other approaches, avoiding any sample preparation, are more appropriate.
One method involves the use of a radio frequency powered (rf) glow discharge. The main advantage of the rf source is the possibility to atomize electrically nonconducting solid samples directly, without exposing any auxiliary conductor to the discharge. During the last years, this source has gained, rightly, increasing interest. However, at this moment, no commercial rf glow discharge mass spectrometer is available.
As an alternative for rf GDMS, also direct current (dc) GDMS can be used for the direct analysis of solid nonconductors. This is made possible by the use of a secondary cathode or mask. The flat sample under examination is exposed to the glow discharge through the aperture (3 - 10 mm) of a conducting diaphragm, called the secondary cathode. According to the theory suggested by Milton and Hutton, who first applied this technique successfully for the analysis of glass, a thin conducting film is formed on that part of the nonconducting sample that is in contact with the plasma. This is a consequence of the continuous, in situ, sputter-redeposition of secondary cathode material. When the redeposited layer is thinner than the sampling thickness, it can be penetrated by the bombarding particles. The effective atomization of the nonconducting sample is thus dependent on the appropriate thickness of the conducting layer: a too thin or incomplete layer causes discharge instabilities due to the charge built-up on the sample surface, while an excess of redeposition will prevent the bombarding particles from reaching the nonconductor to be analyzed. Because the relative redeposition is proportional to the discharge pressure, Milton and Hutton considered the pressure as (only) determining parameter in the building of the conducting layer.
In this project, the secondary cathode technique has been further evaluated, based on the limited investigation presented by Milton and Hutton. We have focused on two different aspects of the secondary cathode technique: first the methodological aspects are thoroughly examined; second, several novel applications have been performed with this technique.
It has been shown by Milton and Hutton and confirmed by our own experiments that specific operating conditions are required to obtain successful atomization of a nonconducting sample. Moreover, we have noticed that these optimum conditions seemed not to be the same for all the samples under investigation. An attempt is made to get a better insight in the varying parameters involved in the formation of the conducting layer.
First, some sample characteristics, in particular the electrical resistivity and the specific surface properties have been evaluated. The effect of the electrical resistivity was significant but has only to be taken into account to classify samples as "poorly conducting" or "nonconducting". It was found that samples with a specific electrical resistivity smaller than ca. 5.109 cm could directly be used as cathode to initiate and maintain a stable discharge, without exposing an auxiliary conductor (for example a secondary cathode) to the discharge. The obtained sample signal intensity was, however, found to be reversely proportional to the electrical resistivity of the sample. This means that, for samples with an electrical resistivity close to the edge of 5.109 cm, the sample signal intensity can become extremely low (5.10-16 A). In these cases, the use of a secondary cathode has found to be able to enhance the signal intensity with a factor of 105, to a level that is almost comparable to the intensities reached for conducting samples. Above a critical edge of conductivity (ca. 1.1010 cm) samples become actually completely "nonconducting". This means that no discharge at all can be obtained when samples of this type are used. The use of a secondary cathode is in this case absolutely necessary to create a stable discharge. The difference between "poorly conducting" and "nonconducting" samples has significant consequences. The analysis of poorly conducting samples displays a higher sample signal intensity and a lower contribution of the secondary cathode, resulting in limits of detection that are a factor of ca. 10-20 better than for truly nonconducting samples.
Another sample characteristic that has been evaluated is the surface roughness. It has been demonstrated that the atomization of a sample with a rough surface requires more deposited metal atoms (i.e. a higher discharge pressure) than that of a sample with a smooth surface. This is to be expected, because a rough surface exhibits a larger area exposed to the discharge than a smooth surface.
A third factor that appeared to have a possible influence is generally termed as the affinity between the sample and the deposited metal atoms. It has been shown that samples with a comparable electrical resistivity and a comparable surface roughness can require other discharge conditions to obtain stable atomization of the nonconducting sample. This can plausibly be addressed to differences in surface affinity; analogous observations have been made in the field of plasma deposition techniques.
Second, besides the sample characteristics, also several operating parameters have to be taken into account. The secondary cathode material is of most importance, both because of its sputter yield and because of the impurities present in it. The sputter yield should be low, to prevent an excess of redeposited metal atoms on the nonconducting sample. Only Ta was found to be useful as secondary cathode material for the analysis of truly nonconducting samples. Care has to be taken in choosing the Ta secondary cathode material (manufacturer and even batch).
The secondary cathode geometry can also be considered as a factor determining the final analytical performance of the technique. For the analysis of truly nonconducting samples, the secondary cathode aperture had to be restricted to 4-5 mm, in order to create a stable discharge. The thickness of the secondary cathode is of importance as well. Thin secondary cathodes (ca. 0.25 mm thickness) are preferable to reach low limits of detection, because of the higher sample signal intensity and the lower contribution of the secondary cathode in the mass spectrum. However, because the thickness of the secondary cathode indirectly also influences the crater obtained, deviation of this general rule may be useful in cases of depth profiling.
Evidently, also the influence of the discharge conditions was evaluated. Here, it has to be emphasized that, in contrast to the theory suggested by Milton and Hutton, not only the pressure is of importance, but also the discharge current and the discharge voltage have to be taken into account in the evaluation of the formation of an appropriate conducting layer on the nonconducting sample.
Generally, one can state that several parameters have to be considered together to reach a maximum sample signal intensity and a minimum secondary cathode signal contribution.
After the methodological evaluation of the secondary cathode concept, in order to investigate how to sputter-atomize various sample types, the analytical performance of the technique is described in Chapter 4. The technique has been applied to various nonconducting matrices: glass plates, an iron ore sinter, marble and limestone, ceramic materials (Al₂O₃, ZrO₂ and Macor ceramic), polymers (polytetrafluoroethylene or Teflon, polycarbonate, polyvinylchloride and polyimide) and, after slight adaptation of the technique, also atmospheric particulate matter.
First, it is important to notice that almost all these materials could successfully be analyzed with dc GDMS using the secondary cathode technique. Only the limestone (because of its too rough surface) and the polyvinylchloride (because of its low melting point) could not be analyzed in a stable way. In all other cases, the reproducibility was at the order of 10% RSD. Second, the obtained signal intensities were found to be between 2.10-12 and 5.10-11 A. Given that peaks of ca. 1-5.10-18 A can be distinguished from the background, this results in a practical power of detection between 20 ppb and 1 ppm. However, spectral interferences can, for certain elements, increase the limits of detection. Moreover, sputtering of the secondary cathode material results in a certain blank contribution that can also largely affect the power of detection. Third, it was found that satisfying semi-quantitative data (within a factor of 3) were obtained, based only on one internal standard and no external standards.
Compared to rf GDMS (i.e., the technique that is also capable of handling solid nonconducting samples without grinding them), the above mentioned blank contribution is probably the major drawback. Fortunately, this can only be considered as a limiting factor for about 15 % of the elements. Another obvious disadvantage of the secondary cathode technique is that the discharge conditions cannot be chosen freely and that the required conditions can differ from matrix to matrix. Searching the optimum discharge conditions can demand some trial-and-error experiments but this is a once-only problem. However, for depth profiling purposes, the limited discharge conditions can really expose a problem, because the shape of the crater (ideally rectangular) can hardly be varied. For this application, rf GDMS certainly seems more interesting.
On the other hand, the secondary cathode technique has also a definite advantage in comparison with rf GDMS. The signal intensities do not depend on the thickness of the sample, a problem that has been, up to now, inherent to the use of rf GDMS. Moreover, the rf GDMS equipment is much more complex and more expensive than the use of a simple diaphragm.
Generally, the use of the secondary cathode can be considered as a viable alternative for rf GDMS, unless there are specific requirements. Especially for those researchers who are already using dc GDMS equipment, the secondary cathode technique can be very interesting. It largely extends their area of application without having an extra cost or a significant change of operating procedures.

Weathering of monuments by natural and air pollution factors: a study using trace-, isotopic- and micro-analysis. KARLIEN TORFS, 1997

Stable isotope measurements are a useful tool in determining the source of a certain element. Sulphur isotope analyses are regularly utilized to identify the origin of sulphate in rainwater and aerosols and of atmospheric SO₂, but only seldom to explain stone deterioration phenomena.
Various kinds of samples were taken in the region around Antwerp, Belgium : in the centre of the city, in the harbour (north of the centre) and in the surrounding area. Stone crusts were gathered, atmospheric SO₂ was sampled, rainwater was monitored and fume exhausts of several anthropogenic emitters were studied (a power plant, oil refineries, ...). The isotopic composition was measured using an isotope ratio mass spectrometer, after converting the samples into SO₂.
Relatively large variations of the ³⁴S/³²S ratio have been found in gypsum crusts on monuments situated in the centre of Antwerp, the suburbs and the surrounding area. More heavy sulphur is encountered in the centre of the town. This variation may be caused by the importance of other SO₂ sources resulting in an enrichment of ³⁴S in the weathering crusts in the centre of the city. The variation in S isotopic composition of the crusts coincide geographically with the variation of SO₂ in the atmosphere, but generally. The petrochemical industry in the harbour is supposed to be the major SO₂ emitter in the north of Antwerp, while in the southern region, the local brickyard industry influences the isotopical composition of SO₂ and the weathering layers on stone.

Characterization of individual aquatic suspended particles using Electron Probe X-ray Microanalysis and Scanning Transmission Electron Microscopy. WENDY JAMBERS, 1997

A review of literature has shown that micro-analytical techniques have hardly been used to characterize individual suspension in aquatic environments. The majority of the publications discuss the results of analyses of samples collected, in the frame of the GEOSECS project, in the Pacific and Atlantic Ocean and of samples collected in different rivers in Western Europe. However in some continents, like South-America and Africa, and in large regions, like the Middle-East and the former USSR, individual aquatic particles have never been characterized.
The majority of these single particle analyses have been performed using automated electron probe X-ray micro-analysis (EPXMA) and scanning electron microscopy combined with energy dispersive X-ray detection (SEM-EDX). Thanks to this automation, large numbers of particles can be analysed in a reasonable period of time, resulting in an improved accuracy for the characterization of the particles. However, these techniques can only give information about the chemical composition of the particles. This limitation was circumvented by using manual EPXMA and scanning transmission electron microscopy - selected area electron diffraction combined with EDX (STEM-EDX-SAED) in addition to the automated EPXMA measurements. These additional analyses reveal information about the morphological and crystalline structure of the particles. Another problem encountered during automated EPXMA measurements is the difficulty to analyse organic particles. This can be solved by staining the samples with ruthenium tetroxide (RuO₄), which is a strong oxidizing agent that reacts with most organic compounds. This staining results in an increase of the effective average atomic number of the elements in the organic particles and thus a higher BSE signal, necessary for detection during automated EPXMA.
These automated analyses result in huge data matrices, which have to be reduced for interpretation. This reduction was preformed using factor analysis and hierarchical clustering combined with selection rules. In earlier studies, hierarchical clustering was always combined with a secondary clustering of the groups resulting from the first one, but this combined clustering is very time consuming. Tests on the classification using clustering followed by selection rules or using only selection rules have proven that both give comparable results. Selection rules can only be applied when some information about the chemical composition of the data is available, but this information can be obtained using the relative X-ray intensities of common minerals present in suspension and the results of the primary clustering. This combination of hierarchical clustering and selection rules can be used to reduce the data matrices of samples without prior knowledge of the chemical composition of the samples. The data of the automated analyses were also used to obtain size distributions of the different particle types and to calculate the TSP values of the samples.
Before STEM-EDX-SAED could be used for the characterization of suspended particles, the camera constant necessary to calculate the lattice spacings of the crystalline structure, and the imaging system had to be calibrated using standard materials. Because this technique had never been used for the characterization of suspension present in seawater, also the sample preparation had to be optimised. Studies on freshwater samples have revealed that embedding of the particles in the hydrophilic Nanoplast resin and direct centrifugation of the particles onto the electron microscopy (EM)-grid are the preparation techniques which induce minimal physicochemical and physical transformations of the particles. Because Nanoplast embedding produced the smallest numbers of artefacts, this approach was the first to be tested on North Sea samples. Unfortunately, sea salt crystallized not only as individual particles, but also as coatings on natural suspended particles, resulting in interferences in the diffraction patterns of these particles. Crystallized sea salt was also present after direct centrifugation, but the amount of coated particles was limited. Thus direct centrifugation onto an EM-grid is the best preparation technique for STEM-EDX-SAED analyses of suspended particles present in seawater.
Suspended particles, collected during the different seasons at ten locations and three depths in the Southern Bight of the North Sea, have been individually characterized using EPXMA and STEM-EDX-SAED. The composition of the suspended particles was relatively constant in place and time. Only during high primary production a different composition was obtained. All suspension samples were dominated by aluminosilicate, Ca-containing and Si-containing particles (mean abundance of over 95%). The aluminosilicates were mainly of mineral origin; diffraction analyses revealed the presence of biotite and chloritoid, but also a limited number of anthropogenic fly-ash particles were detected. Ca-containing and Si-containing particles had a dominant mineral source, but biogenic particle were also important contributors to these particle types, especially during periods of primary production. The majority of the Fe-rich particles were found to have a mineral source, while Ti-rich particles were identified as polycrystalline rutile and brookite, which are both of anthropogenic origin. Staining of the samples with RuO₄ has confirmed the presence of an organic coating on all suspended matter in natural waters. However, the amount of organic matter present in the particles varied with the seasons. Biogenic and trace metal particles always contained large amounts of organic material, but during primary production in spring Fe-rich, Ca-rich, Si-rich and aluminosilicate particles also hold large amounts of organic matter. This Ru-staining not only revealed that only half of the organic particles are detected during automated analyses of unstained samples, but it also showed that some trace metal rich particles and aluminosilicates have too low backscatter electron (BSE) signals to be detected in the unstained samples. It can thus be concluded that the total suspended matter can only be characterized after staining of the sample. A detailed study of the trace metal containing particles revealed that the majority were natural Mn-containing particles, but also significant contributions of Cr-containing and Ni-containing particles were detected. The Cr-containing, Ni-containing and Zn-containing particles were mainly associated with organic particles, while manganese was mainly present as a coating on minerals. Comparison of the total suspended particulate matter (TSP) of the Ru-stained and unstained samples revealed that the amount of organic particles was small during winter and summer and large and more variable during primary production. Due to the influence of wind force and wind direction, the total amount of TSP was three times higher in autumn and winter compared to summer and spring.
During the sampling campaigns on the North Sea also rainwater samples were collected. The particulate phase was characterized using EPXMA, while the dissolved fraction was analysed using inductively coupled plasma - optical emission spectroscopy (ICP-OES), ICP-mass spectrometry (ICP-MS) and ion chromatography (IC). The suspension was dominated by natural aluminosilicates, organic and Si-rich particles, but also relatively high abundances were found for the Fe-rich and Ti-rich particles, which both mainly originated from anthropogenic sources. For two successive samples, collected in the port of Ipswich, a significant decrease was noticed in the TSP value, particles size and abundance of coarse aluminosilicate particles. During the first shower, the coarse particles, mainly present underneath the cloud, are washed out, resulting in a dominance of smaller rainout particles during the second shower. Above sea, the amount of TSP is much smaller and more variable than above land and the differences in washout and rainout between successive samples is less visible. This is most likely due to the constant input of sea-spray particles to the air and thus during rainshowers also to the rainwater. The trace metal containing particles were dominated by Cr-containing particles, which mainly seem to originate from the surface micro-layer of the North Sea. The importance of the micro-layer as a source of trace metals was confirmed during the analyses of the dissolved fraction, where a correlation was found between high concentrations of trace metals and of marine compounds like chlorine, sodium, sulphur and potassium in the dissolved fraction of the collected rainwater.
Comparison of the suspended particles in rainwater and aerosols collected above the North Sea and suspended particles collected in the North Sea have revealed that the air is an important source for trace metal containing and Ti-rich particles present in North Sea suspension. It also revealed that the large quantities of Ca-containing particles present in the North Sea suspension cannot originate from the air, and because only small concentrations were reported to be present in riverine suspension, they have to originate from marine production and from the erosion of the chalk cliffs of Dover and Calais.
EPXMA has also been used to characterize suspension and sediments from Lake Baikal and the Kara Sea, both located in Siberia. In Lake Baikal, suspension samples have been collected at the inflow of the tributaries and in the central part of the southern basin. The samples were dominated by natural aluminosilicate and Si-rich particles, but traces of pollution were also visible, especially in the southern basin. This pollution is not only produced by the large pulp mill in Baikalsk, at the southern coast of the lake, but also by the industry near Irkutsk and Ulan-Ude. In the northern basin, pollution was visible at the Tiya River, which contained large abundances of Fe-rich particles that most likely originate from debris dumping during the construction of the Baikal-Amur railroad.
The influence of the Yenisey River on the composition of the suspension and sediments of the Kara Sea was tested using suspension and sediment samples collected from the estuary to 400 km in open sea. These samples were dominated by aluminosilicate, Si-rich and Fe-rich particles. Some of these Fe-rich particles also contained manganese and zinc and are thus most likely of anthropogenic origin. The large amounts of Ti-rich particles present in these samples are also though to have an important anthropogenic source. Only for the Fe-containing aluminosilicates a decreasing trend was detected going from the river to the open sea, which points to the Yenisey as the major contributor of these particles. Due to the large numbers of overlapping particles (the filters were heavily and inhomogeneously loaded) and the presences of crystallized salts, especially sea salt (the filters were not rinsed with de-ionized water), no extra information could be obtained about the trace particles and TSP of the suspension samples. Although the information gained from these samples is limited, the results are still presented, because these are the first results ever to be published on the analyses of individual aquatic particles of a region that is one of the main inputs of water to the Arctic.
In general it can be stated that suspension and sediments present in lakes and seas are dominated by natural materials originating from erosion and biogenic production, like aluminosilicates and Si-rich particles, but the traces of anthropogenic inputs are also visible, even in remote areas like Siberia.

Microanalysis of indoor and marine aerosols and cloud particles. LIEVE DE BOCK, 1998

This study reports on the micro-analytical characterisation of three different environmental particles: indoor aerosols, marine aerosols and cloud particles. Insight in the possibilities and applications of micro-beam techniques, in the field of single particle analysis, was acquired by a literature search of the period between January 1990 - January 1997. The conclusion of the review stated that micro-analysis on the single particle level has indeed been able to reveal unique information, inaccessible by bulk analysis. However, since each of those techniques has still its own drawbacks and no technique is capable of providing the complete characterisation of the sample, the combination of several complementary techniques to solve the problem is still required. Moreover there seems to exist a tendency towards particle analysis in the on-line configuration, in which particles are collected immediately into the instrument, eliminating typical problems associated with off-line techniques.

Multivariate techniques like cluster and factor analysis have become essential for the processing and interpretation of huge data sets, obtained by fully automated analytical techniques. The most important theoretical aspects of hierarchical cluster analysis (HCA) and principal factor analysis (PFA) were highlighted in this work. The combined approach of HCA and PFA, in the Integrated Data Analysis System (IDAS), for the processing of single North Sea aerosol data sets, appeared to be a remarkable technique to uncover the real intrinsic structure of the data set. The stability of the PFA procedure, as well as the reproducibility of the sampling campaign, was confirmed.

To elucidate the soiling of paintings in the Correr Museum in Venice, a combination of EPXMA, SEM-EDX, STEM-EDX, Nano-Probe, SPM and EDXRF was successfully applied for the complete characterisation of the individual indoor aerosol particles. Multivariate techniques performed on the EPXMA and SEM-EDX data revealed three aerosol sources and six to eight particle types. SEM-EDX and SPM mappings showed that the large number of elements present in particularly the Ca,Si-rich particles (diameters >8 mm), could be attributed to coagulation processes between Ca,Si-rich particles and a series of small particles. Nano-probe data of aerosol particles (diameters <100 nm) revealed that the in- and outdoor aerosol compositions were identical, in this size fraction. Information on the distribution and semi-quantification of trace elements inside the giant Ca,Si-rich aerosol particles was achieved by SPM measurements. A comparison between the in- and outdoor aerosol compositions suggested further the existence of an indoor Ca-rich aerosol source: probably the deterioration of wall plaster and building materials, and the strong accumulation of Ca-rich particles inside the museum. An evaluation of the composition of the particle size fraction, considered as potentially harmful for the paintings, revealed that sometimes almost all of these particles can be identified as a threat. Finally, the overall results clearly stressed the fact that the indoor air composition should be seen as a complex function of several factors. A controlled environment inside the museum will be necessary to avoid, on the one hand the infiltration of outdoor particles and, on the other hand the accumulation of particles in the museum. An adequate coating of the interior walls in the Correr Museum, to prevent particle production by deterioration of wall plaster and building materials, should be considered. Implementation of the suggested actions will finally lead to a better preservation of the visual qualities of the paintings in the Correr Museum in Venice.

Individual North Sea aerosol particles, sampled between 1993 and 1995, were studied by EPXMA and SEM-EDX in combination with multivariate techniques. Hierarchical cluster analysis revealed nine major particle types. All particle types were discussed in view of characteristic sampling periods which were classified, according to the air mass back-trajectories, in three categories: marine influenced periods, periods influenced by Scotland-England and by France. Air masses influenced by England seemed to contain important quantities of aluminosilicate particles and some Fe-rich particles, both enriched with S. The presence of gypsum particles on the other hand was more pronounced in air masses influenced by the south of England. Finally the main difference between the terrestrial influence of France and Scotland-England was encountered in the abundances of Fe-rich and gypsum particles. Additionally these French air masses were also characterized by a small number of heavy metal containing particles. Seasonal variations were investigated by comparing the composition of aerosol samples collected during the summer and the winter seasons. Air masses influenced by Scotland - E-England revealed higher abundances of aluminosilicate particles during the summer, whereas Pb-rich particles and Organic- and S-rich organic particles appeared to be more abundant during the winter season. The major difference between the summer and the winter marine air masses was found in the abundance of transformed sea salt particles. Finally, in a pure marine environment a higher organic particle abundance was found during the summer season. The identification of the particle origin was possible by performing principal factor analysis on the aerosol data of all five North Sea sampling campaigns. Two giant aerosol sources and three sub-micrometer particle sources were identified.

Despite the successful applications of FTIR in different fields, several problems were encountered in a case study concerning the FTIR-microscopic analysis of giant marine aerosol particles. In conclusion it appeared to be more interesting to combine EPXMA and FTIR for the analysis of particles with a higher concentration of organic material. In a second case study, the giant aerosol particle composition variations in air masses crossing the North Sea, were evaluated. After crossing a large distance above the North Sea, the abundance of marine crystallization products increased at the first two samples stages of the May impactor while the aluminosilicate abundance on the first three sample stages decreased. In some of the samples the transformed sea salt abundance decreased or even disappeared at higher stages in favour of pure NaCl. No significant decrease in particle diameter could be observed. The relation between the giant particle composition and their shape was obtained successfully by non-automated EPXMA. All particle shapes matched with those described in the literature, only for NaCl and CaSO₄ unexpected shapes were detected.

To acquire a better understanding of the mechanisms responsible for the modification of marine stratocumulus clouds by emissions from ships, the composition of aerosols (>0.2 mm), collected during the MAST experiment, in different reservoirs (below-cloud ambient aerosol, ambient cloud droplet residual particles, ship plumes below-cloud, ship track cloud droplet residuals and above-cloud aerosol) was investigated by SEM-EDX. Conclusive evidence was found that combustion-derived particles emitted from a ship stack can result in cloud droplet formation in a ship track. The occurrence of Si-containing particles, in the below-cloud plume and ship track residual particle samples from a diesel-powered ship, provide the direct lick between aerosols emitted from the ship's stack to an effect in cloud. Additionally, it was found that a large fraction of the ambient aerosols were composed of organic material or organic material associated with chlorine. The number fraction of these organic or organohalogen particles was largest in clean marine boundary layers, and decreased with increasing pollution levels. The fraction of transformed sea salt increased with pollution level in the MBL, reflecting the increased anthropogenic impact in the more polluted boundary layers.

Evaluation and optimization of "Grazing Emission X-ray Fluorescence" spectrometry. MARTINE CLAES, 1999

Grazing-emission X-ray fluorescence (GEXRF) spectrometry is a recent XRF development based on the total reflection phenomenon using a wavelength-dispersive (WD) detection device. The idea and the technique are so new that only a prototype GEXRF-module exists at present and, at the start of this Ph. D.-work only, a few publications were available in literature.

The combination of grazing-emission XRF techniques with WD detection is based on the work of Becker et al., who stated in 1983 that XRF spectrometry can be made surface sensitive not only by grazing incidence of the exciting radiation, as in total reflection XRF (TXRF), but also by grazing emission of the characteristic radiation, i.e. by detecting only that part of fluorescence radiation that is emitted at grazing (otherwise known as "glancing") angles. The small solid angle of detection is compensated for by irradiating the sample directly with the uncollimated polychromatic radiation from the X-ray tube. Becker's results imply that GEXRF can be used for applications similar to those of (variable angle) TXRF, i.e. determination of the composition, thickness and density of thin layers and the analysis of elemental trace amounts on flat surfaces.

Two main advantages of GEXRF can be distinguished. First, the principal idea of GEXRF implies that the emitted characteristic radiation is only measured under an angle of a few minutes above the sample. Thus, the scattered radiation generated in the support material cannot be detected and the signal of a sample deposited on such a flat support can be measured in a background-free way, leading to extremely low detection limits (down to the sub-ppb level). Secondly, because of the WD detection, GEXRF also lends itself for the analysis of low-Z elements, which is an advantage over conventional TXRF.

Originally, the GEXRF unit was designed to be an accessory to a normal WD-XRF instrument, and it was expected to be especially powerful for ultratrace measurements, as are often required in environmental research projects. The aim of this work was to investigate how effective this system was in practical situations for the analysis of elemental trace amounts, to investigate ways in which the lower detection limits could be minimized and to optimize and evaluate the GEXRF technique for environmental sample analysis.

Grazing-emission XRF spectrometry utilizes WD-detection, as a result of which the range of elements accessible for grazing X-ray techniques is considerably extended towards the lighter elements. Because this softer characteristic radiation is more strongly absorbed in the matrix, it was of interest to investigate the extent of the linear response for micro- and trace analysis using GEXRF. This was done by calculating the fluorescence intensities as a function of layer thickness and composition, which is explained in more detail in Chapter II. The large path length of the emerged beam with photons of lower energy can cause matrix effects which could be harmful to the quantitative determination of elements. There are some conditions that have to be fulfilled to be able to apply the methods of quantification. The specimens are limited in thickness and consequently in area-related mass or covering of the carrier. Therefore, the critical thickness of a sample layer for which matrix effects can be considered as absent, was determined by studying the absorption-enhancement effects in thin sample layers for different analytes in three kinds of matrices, taking into account the interferences of the fluorescent radiation above the substrate.

The linear concentration range of GEXRF when applied to micro- and trace analysis was investigated. It was assumed throughout this study that in micro- and trace analysis the only reflecting optical interface present is that between the sample and the sample carrier. For an optimum signal-to-noise ratio in the measurement, the use of a sample carrier of higher mass density than the expected layer density is recommended. Linearity is ensured for ultra-thin (< 1 nm) layers, e.g. the sub-monolayer amounts encountered in silicon surface contamination analysis. For thicker layers, two effects cause non-linearity: oscillations in the calibration curve due to standing wave fields and negative deviations from linearity due to absorption of radiation by the matrix. The latter can be largely compensated for by the addition of an appropriate internal standard. Oscillations in the calibration curve are more complex. However, in real-life samples prepared by evaporation of aqueous droplets, oscillations in the calibration curve are likely to be less important. In preparing these samples, the exact morphology of the dried residue is not well controlled and generally a distribution of particulates of different sizes is deposited. Here, oscillations will be smoothed due to the averaging over the particle size distribution.

The possibility of using polycarbonate as an alternative to silicon and quartz glass as sample carrier in GEXRF has been investigated. The roughness and flatness were tested by means of a profilometer, from which it was concluded that polycarbonate suits as a cheaper alternative for silicon and quartz glass discs. Furthermore, a pretreatment of the carriers was applied to obtain an acceptable homogeneous distribution. Electron MicroProbe Analysis (EMPA) applications showed that this pretreatment leads to a better homogeneity of the sample deposit on the carrier. Sample preparation procedures for solid samples, aerosol samples and surface and thin layer samples are also investigated.

Further an attempt has been made to get a better idea about the analytical features of the grazing emission technique. While it is certainly true that no one technique can ever be expected to offer all of the features that a given analyst might desire, the GEXRF method has good overall performance characteristics. In contrast with a conventional TXRF set-up, all elements from Z > 4 (Be) can be determined in a WD-grazing emission mode if the instrument is equipped with suitable analyzing crystals. The good sensitivities for light elements allow the use of minute sample amounts for a non-destructive analysis. The improved peak-to-background ratio, due to geometrical optimizations, leads to correspondingly lower detection limits. Instrumental detection limits for chemical analysis within the range of 40 pg to 3 ng can be realized. This corresponds to a concentration of between 400 ng/l to 3 µg/l, if a sample volume of 100 µl is examined. The counting statistical error was calculated to be typically about 2 %. The accuracy is determined to some extent by matrix effects and by extraneous additions and losses before or during the measurement step. For the WD detection in the grazing emission technique, this accuracy is estimated to be better than 10%. Since the present GEXRF prototype is a sequential rather than a simultaneous instrument, the main disadvantage at the moment is the long analysis time when many elements have to be determined.

After the characteristic evaluation of the grazing emission device, the analytical performance of the technique is illuminated. The technique has been applied to various materials, among which water samples (dripping water, reference water standard, mineral water and seawater), biomaterials and pigment and aerosol samples. It should be pointed out that all these materials could successfully be analyzed with GEXRF. Only aerosol samples (because of geometrical problems) could not be analyzed in a stable way. In all other cases the reproducibility was about 10%. Given the prototype status of this grazing exit technique, the detection limits for light elements (sub-nanogram level) obtained for environmental sample analysis can be improved. However, for the high-Z elements, WD-GEXRF is expected not to be better than its related TXRF method.

The competitiveness of GEXRF is compared with efficient and well-established methods of X-ray fluorescence and other techniques for micro- and trace analysis. It is clear that there exists a wide variety of XRF techniques for trace analysis and that several techniques are quite competitive; in general, the choice of method to be used should be made for each case separately. The most important XRF-competitors of GEXRF are TXRF, synchrotron source spectrometers (SSXRF) and particle induced X-ray emission (PIXE). Two things that each of these special systems have in common are a very high sensitivity and ability to work with extremely low concentrations and/or small specimens. It is clear from the obtained results during this dissertation that the perspectives for trace analysis of liquid samples with GEXRF, as a macro-analytical tool, are not very promising since matrix effects are difficult to avoid. A second reason to pass GEXRF for trace analysis is the non-competitiveness of GEXRF or any other XRF technique with ICP-MS. However, if elemental traces in small volumes e.g. in biological tissues, have to be determined, the GEXRF technique has potential that ICP-MS does not have. Hence, this will be the application field for further optimization of the current GEXRF instrumental set-up.

The most promising features of GEXRF, however, are in the field of surface and thin-layer analysis. Trace contaminations on flat surfaces, like silicon wafers, can be determined and depth profiling can characterize stratified near-surface layers. The method is especially suitable for analysis of layers of nanometer thickness deposited on wafers.

Future developments to increase the sensitivity of GEXRF are situated in the field of synchrotron radiation. Designing a laboratory microprobe GEXRF spectrometer with simultaneous detection remains a great challenge. The application of synchrotron radiation to GEXRF analysis can inherently lead to better detection limits for low-Z elements in the fg region. However, the major problem of accessibility to such storage rings and experimental stations still persists.

Environmental analysis for conservation of museum objects and prehistoric cave paintings. KRISTIN GYSELS, 2000

This study is focussed on environmental analysis in view of conservation of our cultural and natural heritage. Two major issues of this work are concentrated on the preservation of museum objects on the one hand and prehistoric paintings in karstic caves on the other hand. A third topic deals with a field evaluation of aerosol sampling techniques for coarse particles. These play a major role in deposition processes and are thus very important in view of conservation of e.g. the North Sea environment.

In Chapter 1 of this dissertation, the environmental influence on preservation of museum objects and cave paintings is discussed. In the case of museums, several factors (micro-climate, gaseous, particulate and microbiological pollution) are important, of which only the influence of aerosols on the conservation of works of art is discussed extensively in this work. Chapter 1 starts with an overview of the sources, concentration and behaviour of indoor aerosols. The situation in museums is described more specifically. Soiling mechanisms and potentially harmful compounds are discussed as well. Principally, all atmospheric particles can be considered harmful because they can cause significant soiling when they are deposited onto the works of art. However, some compounds are more harmful because they can also cause chemical damage. From this point of view, especially S-rich, Fe-rich and organic particles are important.

Preservation of cave paintings is also discussed in Chapter 1. Aerosol pollution is less important in this case. The presence of visitors, releasing heat, water vapour and CO₂ has a major influence. In this way, the natural cave environment, which has been built up over the centuries, is disturbed and the CaCO₃/Ca(HCO₃)₂ equilibrium in the micrometric water film covering the paintings is disturbed. These mechanisms are discussed in more detail in Chapter 1. Control mechanisms caves can apply to minimise the effects of visitors are mentioned as well.

Chapter 2 gives an overview of the analytical techniques used in this work. For the analysis of the bulk aerosol concentrations, EDXRF and PIXE were applied. On a rather extensive amount of atmospheric and aquatic suspended matter samples, automated single particle analysis was performed with either SEM-EDX or EPMA. Dissolved compounds in the water samples were determined by IC, and ICP-AES or AAS. Soil samples were analysed by EDXRF, or EPMA and DRIFT. The main advantages and disadvantages of these techniques, in view of the types of samples analysed in this work, are discussed in Chapter 2.

The results of the characterisation of the indoor aerosol in four European museums are described in Chapter 3. The selection of these four museums was based on their architecture, climate and environment. The Correr Museum in Venice, Italy, the Sainsbury Centre for Visual Arts in Nowich, UK, the Kunsthistorisches Museum in Vienna, Austria and the Koninklijk Museum voor Schone Kunsten, Antwerpen, Belgium were chosen for a closer investigation. Bulk aerosol concentrations, as well as the composition of the individual particles, were determined. Comparison with the composition of the outdoor aerosol provided more information about the possible sources of the indoor particulate matter. Moreover, dry deposition samples were collected in order to determine what amount of particles could actually be deposited onto the works of art.

In the Correr Museum (CM) in Venice, degradation of wall plaster was found to be the major source of indoor particulate matter. This led to high concentrations of Ca-rich and Ca-Si particles. The influence of the outdoor atmosphere was not negligible either. Significant amounts of sea salt and S-rich particles were present in the indoor environment. No indoor sources were identified in the Sainsbury Centre for Visual Arts (SCVA) in Norwich. The outdoor air appeared to be the major source of indoor aerosols. Mainly S-rich particles were identified in winter, while in summer, sea salt particles were most abundant in the indoor air. The direct outdoor influence was minor in the Kunsthistorisches Museum (KHM) in Vienna. Construction works, carried out on the museum courtyard, generated many Ca-rich and Ca-Si particles, which ended up in the galleries through the air-conditioning system. In winter, a new elevator was constructed inside the museum, leading to even higher concentrations of these particle types. In the Koninklijk Museum voor Schone Kunsten (KMSK) in Antwerp, the outdoor influence was again larger. In winter, restoration and construction works constituted an additional indoor source of Ca-rich and Ca-Si particles. Along with sea salt, these were the main particle types identified in this season. In summer, S-rich particles were most abundant. The summer abundances of Ca-rich particles remained low, even though the museum is situated in a limestone building. Interior wall plaster and construction works apparently have a far greater influence on indoor Ca-concentrations.

In each museum, the influence of visitors was clearly reflected on the soil dust concentrations. Even though soil dust is a particle type of an outdoor origin, it was found to enter the museums in another way than other particles, which also have an outdoor source. Indoor relative abundances were usually higher than outdoors. Soil dust particles are most probably introduced by visitors and can later be resuspended. This process is even more efficient in the presence of a carpet. This hypothesis is supported by the clear influence of the number of visitors and the presence of carpets on the concentrations of soil-derived elements (Al, Si, Ti and Mn) in the four selected museums.

From the point of view of conservation, the situation was worst in the CM. Concentrations of potentially damaging S-rich and Fe-rich particles were higher than in the other museums. Moreover, significant soiling could be caused by the numerous Ca-rich and Ca-Si particles originating from degradation of wall plaster. In the SCVA and the KMSK, the works of art are not directly threatened by particulate pollution. Especially in the SCVA, they are adequately protected by display cases. Dry deposition values inside these cases were up to 100 times lower than out of the cases. In the KHM, soiling of the paintings could be induced by Ca-rich and Ca-Si particles originating from the construction works. Therefore, both in the CM and the KHM, the indoor air quality could be improved. However, it should be kept in mind that the construction works required to make the changes, could cause even more damage. This was clearly demonstrated by the results presented in Chapter 3. Leaving the situation as it is, even if not ideal, might cause less damage than the construction works needed to alter it.

Chapter 4 focuses on the analysis of aerosol, karstic water and soil samples from three European caves containing prehistoric rock art, namely the caves of Altamira and Tito Bustillo in Northern Spain and the Grotta dei Cervi in Southern Italy. In all three caves, the soil was found to the major factor determining the composition of both atmospheric and aquatic suspended particulate matter. In the Cave of Altamira, mainly calcite and aluminosilicate particles were identified. These particles were also present in the Tito Bustillo Cave, along with higher amounts of Si-rich particles. In this cave, the soil was also richer in Si than in Altamira. In the Grotta dei Cervi, significant amounts of P-containing biological particles were detected. These arise from the Guano (bat excrement) present in the cave in pure form, as well as mixed with the soil. In neither of the caves, any significant outdoor influence on the indoor aerosol was detected. Therefore, outdoor pollution does not constitute a serious threat for the paintings.

The composition of the karstic water was largely influenced by the sea. Relatively high Cl-concentrations were detected, which could have a negative influence on art conservation. At Tito Bustillo, this could even be enhanced by the high SO₄^2- concentrations arising from discharge into the San Miguel River, which flows through the cave. No pigment particles were encountered in the dripping water samples collected underneath a painting, so direct erosion by the percolating water does not occur.

In combination with the results of the other research partners, it can be concluded that at Altamira visitors constitute the most serious threat to the preservation of the prehistoric paintings, in spite of the already rather strict visits regulation. At Tito Bustillo, biological colonisation, induced by the electrical lighting, is very important. Purification of the wastewater discharged into the San Miguel River is also required to improve the situation. No serious problems were detected in the Grotta dei Cervi, because the cave is still closed to the public. However, visitors could significantly increase the CO₂ concentrations, and in combination with the high Cl-concentrations, this could constitute a problem when it is decided to turn the cave into a tourist attraction.

In Chapter 5, the results of a field evaluation of a wind tunnel-impactor system for sampling ambient aerosols are described. According to the theory of aerosol sampling, this system should be more efficient for the collection of coarse aerosol particles. On the average, when a Battelle impactor was operated inside a wind tunnel, in order to create isokinetic sampling conditions, higher apparent concentrations were found on the impactor stages with the largest cut-off values than on the corresponding stages of an identical impactor operating in the free wind. This is clearly indicated by the average elemental size distributions obtained by means of these two sampling techniques.

When individual samples were evaluated as a function of the weather conditions, the obtained results were mostly in agreement with theoretical considerations. In calm air conditions, higher concentrations were detected on the largest stages of the impactor operated in the wind tunnel, as well as for sampling periods when the wind came from the Southwest, aligned with the anisokinetic impactor. No significant differences between the two employed sampling methods could be detected when heavy rainfall occurred, possibly due to a more efficient removal of large particles by washout and rainout processes. When the wind directions were highly variable and when wind blasts occurred, inaccurate sampling also took place in the wind tunnel, probably because of the inefficiency of the wind tunnel system in these extreme conditions. For three out of 24 samples, no explanation for the observed results could be found based on the above-described considerations. Other effects, or a combination of these, like turbulence or secondary aspiration, could be responsible.

It can be concluded that operating a sampling device in a wind tunnel is useful when the weather conditions are not too extreme. In comparison with an impactor operated in the free wind, relative concentration differences of 20 to 150 % can occur. Generally, these differences are higher for the largest particles (> 4 µm dae), so especially when this particle size range is to be studied, the use of a wind tunnel can be recommended. When aerosol size distributions are assessed by means of cascade impactors, a significant underestimation of this size fraction can occur when the impactor is operated in the free wind.

Chemical characterisation of atmospheric aerosols: application in the context of air pollution and global climatic change. STEFAAN HOORNAERT, 2000

Aerosol science nowadays takes an important place in environmental analysis. This is due to the fact that aerosol particles play an essential role in air pollution, climate change and public health. Despite this, there still exists a lot of uncertainty in this field so that a further chemical and physical characterisation of atmospheric aerosols is needed. The work presented here summarises the results of several case studies, which were carried out in the framework of a number of national and international projects, in order to expand our knowledge about aerosols and to reduce the uncertainties.

In Chapter 1 of this work, an overview is given of the different characteristics of atmospheric aerosol particles. After defining atmospheric aerosols, the different ways of classifying them is discussed. Aerosols are produced as primary or secondary particles from both natural and anthropogenic sources. Once the particles have been emitted into or formed in the air, they can undergo a variety of transformation processes, including condensation and coagulation. The cycle ends by the processes by which they are being removed from the atmosphere. Dry removal processes include impaction, sedimentation and diffusion; wet removal processes can be divided into rainout and washout. These processes, together with physical properties, like particle size, determine the residence time of the particles in the air. The last section of Chapter 1 deals with the different effects of atmospheric aerosols. Their influence on public health, visibility, acid rain and a number of effects due to particle deposition is being summarised. More recently the possible climatic effects of aerosol particles have been recognised. Aerosol particles influence the global climate both by scattering radiation and by influencing cloud properties.

Chapters 2 and 3 summarise the results of two case studies on North Sea aerosols. Chapter 2 focuses on Ca-containing particles, especially CaSO₄ for five different (wind)sectors that could be identified. EPXMA data of about 50,000 individual aerosol particles were subjected to non-hierarchical cluster analysis, revealing that Ca-containing particles are present in all sectors mainly under the form of CaSO₄ and aluminosilicates, in variable amounts. Marine air masses are characterised by rather low fractions of Ca-enriched particles. On the other hand extremely high fractions are present for the continental air masses and more particularly those coming from the central and northern part of Germany. The big difference between the northern and southern continental air masses suggests some important anthropogenic sources in the former region. Similar trends are found for the CaSO₄ fractions. In most cases CaSO₄ particles are the main cause of the occurrence of Ca-containing particles. Marine and continental sources of gypsum are identified, the former being only of minor importance. The cluster analysis of the Ca-enriched particles yields several particle types. CaSO₄, with abundances of more than 30%, is by far the most important one. Other groups present are the aluminosilicates and CaCO₃. Minor particle types include Fe-Ca-rich and Ca-S-Si-rich particles. The combination of Ca-enriched particles with sea salt particles is probably due to a marine origin or coagulation of particles. The sources for each of the identified types are discussed in more detail.

In Chapter 3 literature data on Cd, Cu, Pb, Zn, Ni and Cr concentrations in aerosols above the North Sea have been gathered and intercompared to look for possible trends. A distinction is made between measurements in different regions of the North Sea. Also the sources and magnitude of the emissions are discussed.

The majority of the data on trace metals in the literature deals with the Southern Bight. So the trends observed for this part of the North Sea are the most reliable. In this part of the North Sea also the highest concentrations are found. More data are needed to confirm the concentration patterns found for the other regions.

Despite some heavy fluctuations a strong decreasing trend is observed for the Pb and Zn concentrations above the North Sea. A remarkably consistent correlation is found between the concentration patterns of both elements in most parts of the North Sea. This similar behaviour between Pb and Zn is most clearly seen for the Southern Bight. Only in the coastal area of the central North Sea a different behaviour is found. Here Zn concentrations are increasing and Pb and Zn are anti-correlated.

For Cd, Cu, Ni and Cr, much less data are available in the literature. Despite this, also for Cd a decreasing trend is present, however not as strong as for Pb and Zn. The airborne Cu levels above the different regions of the North Sea do not always show a consistent behaviour. In most cases, however, the levels found in the late eighties and early nineties are lower than those observed earlier. Cr and Ni concentrations in most cases are fluctuating, without a certain pattern so with the small amount of data it is hard to recognise a particular trend.

In Chapter 4 a new approach is described to apply EPXMA for the estimation of the homogeneity of powder samples of candidate reference materials (RM), basing on the utilisation of the Kolmogorov-Smirnov statistics. The procedure was applied in the study of the “within-bottle” homogeneity of three candidate reference materials and the “bottle-to-bottle” homogeneity of a fourth reference material.

In the within-bottle homogeneity study, the distribution of the particle size and concentration of the main elements are the same for two of the three RMs so that they can be used as standards for microanalysis for amounts down to the nanogram. The third sample was found to be inhomogeneous at that level regarding its chemical composition.

In the bottle-to-bottle homogeneity study, all bottles of the candidate RM, except one, were found to be homogeneous at the level of ca. 300 ng. One out of the remaining five bottles becomes different from the other four samples, when the sample mass is being reduced down to 8 ng.

Even though the number of particles analysed using CC EPXMA is very small compared to that collected, the major elements for the studied candidate RM need to be analysed just for less than 200 particles, to assure 5% RSD in CC EPXMA measurements. For the minor elements, the required number of particles to be analysed, to assure 5% RSD, ranges from several thousands to tens of thousands. The number of particles required to be analysed, to insure a certain level of reproducibility, increases exponentially as the concentration of elements decreases.

Chapters 5 and 6 both focus on atmospheric aerosols, collected in the Former Soviet Union. In Chapter 5, the results are summarised of two sampling campaigns at a high mountain station in Kazakhstan. The background character of the site was confirmed by comparing the elemental concentrations, obtained by EDXRF analysis, with those reported for other remote stations. Based on the variability of the elemental concentrations, the elements could be classified into two groups: highly variable Si, Al, Fe, K, Ca, Ti, Sr and Mn concentrations and less variable S, Pb, Zn, Cl, Br, Cr, Ni and Cu concentrations. The latter indicates distant sources and mixing of the aerosol before reaching the sampling site. The concentrations of the first group of elements were found to be highly correlated, suggesting that they are originating from the same sources. These findings, complemented with enrichment factor calculations revealed the first group of elements to be originating mainly from more local sources (soil dust), and the second group to be related to long-range transport of anthropogenic aerosol.

Automated single particle analysis by SEM-EDX and subsequent cluster analysis resulted in the identification of 12 distinct aerosol particle types. Relative particle type abundances were converted to absolute abundances by estimating the number concentrations for the detected particles. A very pronounced seasonal variation of the particle number concentration is found, especially in springtime due to the melting of the snow cover.

The seasonal variation of the abundance of the different particle types was investigated. From the relative abundances only limited information could be extracted. Based on the absolute abundances, a strong seasonal pattern is found for Si-Al-Fe-rich, Si-rich, Ca-S-Si-rich, Ca-Si-rich, Fe-Si-rich and Ti-Si-rich particles, closely related to the particle number concentration. These particle types can mainly be attributed to local soil dust. The existence of some additional, but weaker sources is suggested by small differences in the summer patterns. For most of the remaining particle types a completely different seasonal trend is found. They are composed of elements like Pb, Zn and S, which can be related to anthropogenic emissions. Their abundance in winter can be completely attributed to long-range transport of anthropogenic aerosols. In summer, emissions from the Almaty region are superimposed on this base level. An annual cycle can be seen for biogenic aerosols, produced by natural sources, reaching their maximum activity in August.

In Chapter 6, the case of several anthropogenically influenced sites in the FSU is being investigated, with main focus on a sampling campaign near the city of Krasnoyarsk. Next to the bulk results by EDXRF and SEM-EDX results, a literature study is presented on the industrial activities in the different regions of the Siberian territory.

All bulk elemental concentrations at the Krasnoyarsk site were found to be higher than those at the Tien Shan background station. Studying the time evolution of the Krasnoyarsk bulk concentrations, air masses of different origin can be observed. Clean conditions are characterised by concentrations less than 2.6 times the Tien Shan background values, close to the global rural average. During polluted periods ratios up to 50 are observed, but still lower than the global urban average. Enrichment factor calculations indicated that almost all elements are enriched compared to the soil.

The application of non-hierarchical cluster analysis on the overall data set, obtained by SEM-EDX, resulted in the identification of 10 distinct aerosol particle types in the Krasnoyarsk winter aerosol. Relative particle abundances were converted into absolute abundances by estimating the particle number concentration, which was found to be about 20 times higher than the average winter background value at the Tien Shan site. The increase in the bulk concentrations during the sampling was confirmed by an increase in the particle number concentration.

For the source identification of the different particle types, the time evolution of their absolute abundances was investigated. Based on these results, four groups of particle types were found with similar time profiles. According to the Al profile, which can be directly linked to the aluminium industry in Krasnoyarsk, the second half of the sampling period was found to be strongly influenced by emissions from the city of Krasnoyarsk and its industrial activities. The Krasnoyarsk industrial area aerosol was also characterised by high amounts of Ca-S-rich, S-rich, low-Z and Fe-S-(Si)-rich particles, for which several sources could be identified in the city and its surroundings.

During the first part of the sampling campaign, the city’s influence was non-existent. On the base level, representing the Siberian background, two other pollution episodes are superimposed. A first episode, characterised by high abundances of Pb and Pb-Zn-rich particles is thought of originating from mining activities, north of the sampling site. Several coal combustion products could also be detected. Another pollution episode could be attributed to the emissions of wood burning for domestic heating in some smaller settlements free from industrial activities. For the conclusions regarding the other regions we would like to refer to Chapter 6.

In the framework of the Second Aerosol Characterisation Experiment, described in Chapter 7, size segregated samples have been collected for single particle analysis on the island of Tenerife, in both the marine boundary layer (MBL) and the free troposphere (FT). They were analysed by automated and manual SEM-EDX, followed by cluster analysis to identify the different particle types present in the aerosol and their abundance.

Basing on backtrajectory calculations and particle number and volume concentrations, different periods can be identified regarding the origin of the sampled air masses. In the FT, the air masses can be classified as clean Atlantic or influenced by dust from Africa or pollution from Europe. A specific situation occurred, which could only be identified by local observations when the cloud layer was just below the sampling site. In the MBL, air masses are classified as clean or polluted by European emissions, basing on the trajectories, and an additional class of perturbed samples (aged pollution), taking into account the aerosol volume concentration.

For the FT samples a very good agreement was found between the chemical composition of the samples and the air mass type, which was confirmed by applying principal component analysis. Variations of the chemical composition within a particular air mass type are only of minor importance compared to the differences between the air mass types. A fairly good agreement is observed also for the MBL samples but this needed some more consideration.

For both the FT and MBL samples, the biggest change in chemical composition is observed between the fine and coarse mode aerosol. The FT fine mode aerosol is dominated by S-poor aluminosilicates in the case of the dust samples or sulphates, carbonaceous particles and S-rich aluminosilicates in the polluted samples. The clean samples (and the ones taken in cloudy conditions) have an intermediate composition. When going to the larger size fractions, a strong decreasing trend is observed for the sulphates and carbonaceous particles. On the other hand, the S-poor aluminosilicates (and especially their ratio to the S-rich (or aged) aluminosilicates) and the Ca-rich particles are becoming relatively more important.

The MBL fine mode aerosol is completely dominated by sulphates, and to a lesser extent, carbonaceous particles and aged sea salt, in the case of polluted and perturbed air masses. No fresh sea salt particles are found in the fine aerosol for these air masses. The clean air masses, in their fine mode, are made up mainly of sea salt particles with varying degrees of ageing. A drastic change in chemical composition is observed between the fine and coarse mode aerosol. When going to the larger size fractions, the latter is more and more dominated by fresh sea salt particles. Also the difference between the air mass types is largely reduced.

Morphological, microstructural and chemical characterisation of individual soot aggregates from different sources by electron microscopical techniques. AN SMEKENS, 2000

The increase of fine aerosol in the environment has been a hot topic in the last years. Particles with small diameters can penetrate deeply into the airways and thus cause a lot of harmful health effects. Soot, a major component of fine atmospherical dust, is one of the least understood atmospheric aerosols. Incomplete combustion forms particles, which condense into a black aerosol when cooling down. Due to their small dimensions (10 nm to 2 µm), their structure, and the ability to adsorb thousands of different species, these particles can cause severe health effects to men. Because soot scatters and absorbs light strongly and because it enhances cloud formation, it plays an important role in global change. These effects are enforced by the long residence times in the atmosphere and by the transport over long distances.

Because of the effects soot causes, the morphology, microstructure and chemical composition are studied to understand how these effects are produced and how they could be avoided. Since these characteristics of the soot depend on the formation conditions and on the combustion process, some are different for soot of different sources. Thus, it might be possible to designate the origin of environmental soot particles.

Environmental soot principally originates from the combustion of fossil fuel. The main source of soot in the atmosphere is diesel traffic. Other important sources are heating, air traffic and industry. Therefore, we focus in this study on soot emitted by a car, running on diesel and biodiesel, both at idle run and at stationary regime, soot from a oil-fired domestic heating installation and soot from an aircraft exhaust. Carbon black, a product with similar features as soot, used commercially, is used as a kind of reference material.

By performing individual particle analysis, both the morphology and the composition are measured on single particles. The main advantage over bulk analysis is that we gain more detailed information on the nature of the particles, their origin, their formation processes and their potential transformation in the environment. In order to link these parameters up, the particles are in this work characterised as a function of their aerodynamic diameter. The aerodynamic diameter is an important physical dimension in describing aerosols when considering diffusional or inertial deposition into the human respiratory system and when considering particle transport and deposition in the atmosphere. A methodology was established to characterise the morphology of aggregates with automated image analysis on transmission electron microscopical (TEM) photographs. The newly defined parameters that describe the morphology indicate differences for soot of different origin.

Studying the electronic bonding state of carbon, the main constituent of soot, we see that the soot from domestic heating, diesel soot and carbon black all have a significant amount of double bonding character. The analysis of the elemental composition, by energy dispersive X-ray analysis (TEM-EDX), shows that carbon, oxygen and sulphur, probably in the form of adsorbed sulphate, are present in all soot types. The chemical composition differs for different sources.

The size distributions of the soot aggregates, studied by image analysis of field emission scanning electron microscopical (FE-SEM) images, show that the majority occurs in the smallest fraction of the studied range (0.15-0.45 µm). This trend is most pronounced for aircraft exhaust soot, with almost 80 % of the particles having a size of 0.15 µm.

Next to these results, also the visual interpretation of high resolution TEM (HRTEM) images enables to differentiate between soot emitted by different sources, which can make it possible to determine the origin of soot in the environment.

The marine environment and weathering of historic stone monuments. FILIP DELALIEUX, 2001

Sculptures and monuments are often made of porous materials that are subject to the singular or combined action of different weathering phenomena: e.g. dissolution (intensified by air pollutants), freeze/thaw and wet/dry cycles, salt crystallisation and biological attack. Salt weathering of building stones is recognised as one of the primary agents in the loss of historic architecture. The growth of salt crystals within the stone pores can generate stresses that are sufficient to overcome the stone's tensile strength and disintegrate the stone into a powder or flakes. Salt accumulations found in historic buildings originate from a variety of sources, e.g. air pollution, rising groundwater, marine spray, unsuitable cleaning material, incompatible building materials, microbiological activity and de-icing salts. In the past most research has been focused on the salt input caused by air pollution, mainly through deposition of SO₂ and sulphuric acid and the subsequent enrichment with sulphates. In recent decades the atmospheric SO₂-concentrations have declined markedly in developed countries and the lack of knowledge on natural salt input sources such as microbiological colonisation and sea-salt have become more apparent.

The research presented here is concerned with the latter source of soluble salts and the effects of marine salts on various types of natural building stones (e.g. granites, limestone and sandstone). In the context of a project funded by the EC, stone monuments located in four different coastal environments were studied. In addition to the in-situ exposed stone blocks, analyses were also performed on small stone prisms, which were freshly quarried and submitted to artificially accelerated weathering (allowing the simulation of stone deterioration in a coastal environment on a practicable time scale and under standardised conditions). Soluble salt concentrations were determined with Ion Chromatography (IC) and Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES). Micrometer-scale salt distributions inside the stone samples were investigated with a Scanning Electron Microscope equipped with an Energy Dispersive X-ray detection system (SEM-EDX). To get a detailed view on the range of marine salts over land, atmospheric deposition was collected as a function of distance from the coastline and altitude, and analysed with IC and ICP-AES. To get a broader view on the distribution of the deposition of marine salts over land, rain chemistry data from European monitoring stations were compiled and Cl^- and Na⁺ - concentrations were mapped using kriging interpolation. The described research has shown roughly how much and at what depth NaCl is expected to crystallise in different stone types. It was also shown that the NaCl-content is not directly proportional to the susceptibility of the stone type to sea-salt decay. This susceptibility is mainly determined by the pore network characteristics.

The amounts of NaCl deposited from the atmosphere and accumulated in monument stones decrease logarithmically and are linearly interrelated. The influence of microclimatic conditions at the stone surface (e.g. airflow and sun exposure) on the NaCl-concentrations in the stone were found to be up to the same order of magnitude as the influence of the distance to the coastline.

The new insights and quantitative results obtained in this study improve the assessment of the susceptibility of natural stone to sea-salt decay. They help to design more efficient conservation strategies, especially for the numerous monuments located in coastal areas. More generally, this work provides the possibility to predict the order of magnitude of the amount of sea-salt to which an outdoor construction is exposed. Studying the problem of sea-salt decay has not only solved answers but has raised additional questions by exposing shortcomings in the theories and instrumental methods used in this field. We were able to clearly pinpoint these shortcomings and have outlined new pathways for future research.

Atmospheric deposition of nutrients to aquatic ecosystems. KURT EYCKMANS, 2003

Man’s activities, which introduce an excess of nutrients, along with other pollutants, into lakes, streams, estuaries and seas, are causing significant changes in aquatic environments. The pollution problem is critical, because of increased population, industrial growth, intensification of agricultural production, river-basin development, recreational use of waters, and domestic and industrial exploitation of shore properties. Accelerated eutrophication causes changes in plant and animal life, changes that often interfere with the use of water and that reduce property values.

The influence of the atmosphere as a transport route of nutrients to aquatic ecosystems such as the North Sea and lake Balaton (Hungary) has been studied in detail. The atmospheric deposition depends totally on the emission sources and various removal mechanisms. The atmospheric deposition of nutrients is, however, not easily measured. For the wet deposition, practical implications, such as sea or lake water spray, makes the amount of precipitation hard to determine. Dry deposition is even more difficult to measure and it is not possible to do so routinely. Because of these reasons, it is not easy to estimate the nutrient input to the water surface from measurements and, therefore, atmospheric transport models are frequently used. In these models, all the processes influencing the concentrations of all the involved components are taken into account from the place where the sample is originally taken to the place where the deposition finally occurs.

For the North Sea (see Chapter III), nutrients are defined by the OSPAR convention (the Convention for the Protection of the Marine Environment of the North-East Atlantic) as one of the most important human pressures. The excessive anthropogenic input of nutrients from land and changed nutrient ratios primarily affect the coastal zone. Most of the nutrient related problems occur in particular in estuaries and fjords, or in parts of the North Sea where the riverine input is poorly mixed with fresh marine water. These negative impacts include periodic disturbances of the ecosystem, as well as changes in the abundance and diversity of different faunal and floral communities. Data from a sampling campaign of nearly one year at the Belgian coast were used to optimise the model’s parameters in order to obtain a more accurate calculation of the deposition of the various nutrients above the North Sea.

For Lake Balaton (chapter IV), much research has been carried out concerning the eutrophic episodes. However, limited research has been done about the atmospheric input of various nutrients. This less trivial nutrient input source could lead to an underestimation of the total nutrient budget and should not be neglected. Especially in regions far downstream the river Zala such as the Siofok basin and during periods when there is a strong nitrogen limitation or depletion, the atmosphere could be the most important source of nutrient compounds to the water. Many scientists already stated that the atmosphere is a very important source of nitrogen compounds, with special attention to ammonia. Previous studies managed to make some accurate flux calculations of ammonia to the lake, taking into account the specific lake conditions. However, these studies were carried out in the beginning of the 1980s and since then, the lake’s water composition as well as the atmospheric composition drastically changed. In this study, the contribution of the atmosphere as a source of gaseous and particulate nutrients was investigated and the importance of this contribution evaluated for the whole lake.

Furthermore, the complementary information in combining two analytical techniques, bulk Ion Chromatography and Thin Window Electron Probe Micro Analysis, is demonstrated (Chapter V). Combining the results of both techniques, we have been able to make a more detailed interpretation of the aerosols in the air mass.

Finally, in Chapter VI, the different outputs of various ammonia samplers, both active as passive, have been compared. The systems were evaluated concerning their precision, reliability and labour intensity. This study resulted in a profound validation of the working of the lab’s ammonia sampler and possible explanations for deviating results were formulated.

Optimisation and application of different Electron Probe X-ray Microanalysis approaches for the chemical characterisation of individual environmental particles. ANNA DE MAEYER-WOROBIEC, 2004

The study of atmospheric aerosols has recently become very important for two major reasons: the small particle size fraction has a very important influence on human health, and it affects the global climate. Characterization of individual microscopic particles generally permits more information to be obtained than is feasible with bulk analysis. As a result recognition of pollution sources and processes becomes simplified and facilitated. The characterisation of single environmental particles using micro-analytical techniques plays an important role in the environmental science.

In the framework of this Ph.D., particular attention was paid to the application of EPMA to the analysis of single environmental particles. Apart from detailed information on chemical composition and size distribution, another important advantage is the possibility of automation that allows large numbers of particles to be analysed for a better statistical result. On the other hand, the development of thin-window energy-dispersive detectors was of major importance for electron probe microanalysis (EPMA), in order to detect light elements (6<Z). Using the thin-window EPMA, the diversity and the complex heterogeneity of the chemical composition of atmospheric particles can be investigated. In the framework of this investigation a short survey of the possibilities as well as the applications of quantitative EPMA is given. Unlike conventional EPMA, the recent thin-window technique combined with the quantification based on the Monte Carlo simulations allows the concentration of low-Z elements such as carbon, nitrogen and oxygen to be determined.

Apart from the thin-window EPMA approaches, there are also other alternative EPMA methods. For samples containing large numbers of biogenic and organic particles impregnation with heavy metals can be used. The “spot-test“ or microchemical transformation is another analytical procedure based on micro-chemical reactions. Their use in optical microscopy is described, while their application to common atmospheric aerosol anions such as sulphate, chloride or nitrate is also addressed.

Using thin-window EPMA, the low–Z element information is obtained to better account for the overall chemical composition. In this way the classification of individual particles can be at least semi-quantitatively assessed. However, one of the most important problems of EPMA remains beam damage, which causes geometrical and morphological structures to be changed during analysis. These artefacts cannot be accounted for in quantitative concentration calculations. It should also be noted that knowledge about the particle concentrations plays a very significant role in the study of environmental pollution. The use of a liquid-nitrogen-cooled sample stage during the automated analysis of the atmospheric aerosol particles results in a spectacular protection against the beam damage effect. It seems that the beam damage depends not only on the composition of the particles, but also the diameter of the particles and their shape. Moreover, the collection substrate for atmospheric particles plays an important role during the beam damage process.

In the framework of these Ph.D. investigations, some examples of the application of quantitative thin-window EPMA in environmental research are presented. The results from the analysis of the individual atmospheric particles collected in the Amazon Basin, South Africa and Balaton Lake in Hungary are discussed in detail. Since the development and the optimisation of this new method are still in progress, the interpretation of the results and quantification should be done with caution. The comparison of the quantitative results from thin-window EPMA with those obtained from the conventional EPMA is also presented in detail.

Optimization and environmental application of TW-EPMA for single particle analysis. JOHAN DE HOOG, 2005

The thesis describes the different steps that were taken for developing a new method for the semi-quantitative analysis and characterization of (environmental) particles by TW-EPMA.

First, we evaluated the need for such a new method by showing the different technological evolutions in electron microscopy. We saw that with the arrival of a new generation of energy-dispersive X-ray detectors, it became possible to determine light elements. Their ultra-thin polymer windows allow for a better transmission of low-energy X-rays, which made sure that at least qualitative analysis of elements with an atomic number Z > 5 was possible. Since these light-element compounds are very abundant in environmental particles, the breakthrough in detector technology was also a big step forward for enhancing our knowledge about their composition. The use of a so-called cold stage for cooling down samples with liquid nitrogen to –193 °C also allowed us to analyze volatile species, for example ammonium and nitrate compounds, since we could considerably reduce the beam damage effects.

Soon the idea rose to investigate the possibility to develop a method for the quantification of both light and heavy elements contained in environmental particles. A recently developed Monte Carlo model for the simulation of electron-solid interactions at low energies was adapted for particulate samples by implementing the necessary features for taking into account particle geometry effects. The Monte Carlo model was integrated into an algorithm for the iterative quantification of particle compositions based on reverse simulations. While testing the quantification method using particle standards made of inorganic and organic salts, several directly or indirectly critical aspects were investigated in order to adapt the involved analytical procedures like the sample preparation (the quality of substrates and the use of coatings), the beam damage effects, the spectrum evaluation (peak identification and corrections for peak overlap) and the physical parameters used in the model. A comparison with other quantification methods showed that our method produced quite satisfying results, also considering the uncertainties and assumptions we had to deal with.

Then, we tried to go one step further by exploring several possibilities to get the most out of the obtained concentrations for light and heavier elements. For the characterization of environmental samples, huge datasets have to be dealt with in a fast, efficient and intelligent way to get a correct profile of the analyzed particles. Experiments with artificial particles showed that the calculation of light element concentrations by our iterative Monte Carlo quantification procedure indeed produced an added value for differentiating between different particle types when applying data reduction tools. These mathematical methods are able to reduce large datasets to smaller sets of specific characteristic parameters. The purely mathematical data treatment was then extended with an expert system for the elucidation of particle composition. The mental algorithms used to identify particle types were translated into an automatic procedure that could do the same job of the analyst in an unsupervised and computerized way. This method starts on the individual particle level and ends with the classification of similar particles into typical groups.

Two methods were developed and tested for the additional characterization of the internal particle composition. Grazing-Exit EPMA (GE-EPMA) could offer specific information about core-shell structured particles, but some disadvantages due to beam damage and the instrumental setup strongly limited the possibilities and made data interpretation quite difficult.  Multiple-Voltage EPMA (MV-EPMA) appeared to have more potential, since particles are analyzed at different voltages to vary the information depth, so that layered particles could be qualitatively differentiated from homogeneous particles. The technique was also used to determine the surface layer thickness of artificial, heterogeneous particles, but beam damage again played a disadvantageous role. It is clear that the application of GE-EPMA or MV-EPMA to real atmospheric particles will probably be limited to qualitative analysis due to the complexity involved in the analysis, since both their complex chemical composition and their structural heterogeneity cannot be known a priori. It was however shown that both techniques, in combination with light-element analysis, are able to extract information that could not be obtained before.

The developed and optimized method was also put into use by applying it in projects that required the characterization of different environmental samples. The semi-quantitative determination of light elements by TW-EPMA was applied to the analysis of North Sea aerosol particles, and it showed to be an efficient tool for the investigation of a huge number of individual microparticles. The method offered good statistics to study their chemical properties and their behavior in the atmosphere. The results of the collected aerosol samples showed that the light-element analysis of single particles gave additional useful chemical information to marine aerosol studies, compared to the conventional EPMA methods. The detection of characteristic X-ray radiation of low-Z elements, which was not possible before, offered the possibility to identify more compounds (e.g. NH₄NO₃, organic particles), or it enabled the differentiation between particle types, like e.g. (NH₄)₂SO₄ and sulphur-containing organic particles which previously only showed a sulphur peak in their X-ray spectra. The extended, semi-quantitative knowledge of low-Z element concentrations led to a more relevant particle classification; therefore, the possible sources and interactions occurring in the troposphere can be traced in much more detail by the analysis of large numbers of individual particles. Through the combination of chemical data with meteorological data, the environmental link was found between the chemical composition of the different samples and the environmental conditions during sampling.

TW-EPMA has already been applied in many projects for the characterization of different sample types from around the world, like atmospheric aerosols (Amazon region, Brazil; Lake Balaton, Hungary; North Sea; sugar cane plantation, Brazil; indoor and outdoor environments in residential areas, Flanders; indoor and outdoor environments near historical monuments all over Europe) and sediment particles (North Sea; Tisza river, Hungary). For this thesis, the method was compared to and combined with different complementary bulk techniques. As well in the abovementioned North Sea project, as in another project for the characterization of PM_2.5 in Flanders, it was shown that TW-EPMA could offer new insights in the composition and (trans)formation of particulate matter in the atmosphere. The combination with ion chromatography showed that the matching of bulk and microanalytical results could lead to mutual benefits in aerosol characterization. TW-EPMA proved to be a valuable element in the multi-analytical approach of the PM_2.5 study in Flanders.

Should the method from now on be used in every project about the characterization of environmental particles?  No, because for some problems a qualitative approach is already enough or the knowledge about light elements is not necessary. For example, studies about the distribution of heavy metals in river sediments will probably not always require an exhaustive and time-consuming quantitative characterization of the light-element composition. The analyst will have to evaluate what is the scope of the projects and what are the capabilities of the available methods in order to choose the right level of characterization. TW-EPMA combined with our quantification method has proven to be a valuable tool for single particle analysis, with still some considerable potential for improvements in the future … when better technology will again lead to new insights.

Characterisation of health related particulate and gas-phase compounds in multiple indoor and outdoor sites in Flanders. MARIANNE STRANGER, 2005

The quality of the air we breathe has been studied and evaluated extensively worldwide, not only because of its known harmful effects on the climate, vegetation and our cultural heritage, but also because of its recognized potential to cause adverse health effects on human kind. This work aimed to assess the results of a pilot study on the characterisation of gas-phase and particulate matter air pollutant levels in Flanders in different environmental settings, with special reference to its applicability as indicators for the personal exposure of the population.

The majority of the available open literature concerning the causal relationship between air pollution and the morbidity and mortality rates of a population, focussed on the hypothesis that ambient levels of air pollution, measured at one or a few fixed ambient air monitoring sites, could be related to the health implications of a broader population. Furthermore, these studies hypothesised that each individual of a population has a similar personal exposure to air pollution, which is directly related to the ambient pollutant levels measured at one ambient monitoring site. Besides climatological and vegetative arguments, these studies have lead to the establishment of a legislation which is fixed to the control and -if necessary- the reduction of outdoor air pollution. To test the hypothesis that the air quality information obtained from a fixed regional sampling station is an estimation of the quality of air the larger population is exposed to, the obtained data were evaluated against those of a fixed monitor.

In order to assess the situation of the outdoor air pollution in Flanders, this work presented four case-studies which aimed to characterise the outdoor air quality at specific critical sites. Two programs focused on the gas-phase pollutants and two others on the particulate matter phase. For the gaseous pollutants, the attention was paid to two areas in Flanders: in the first case study the ambient levels of the precursor gases NO₂, SO₂ and O₃ were assessed in the city centre of Antwerp and the southern suburbs Wilrijk and Hoboken. On a larger scale, the second case study focused on the ambient NO₂ levels in West-Flanders. Not only was the seasonal variability of these pollutants studied, but also the concentration differences due to local (traffic-)emissions. The traffic related pollutant NO₂ occurred at considerably higher concentrations in the urban centres compared to the more rural suburbs, whereas the ambient O₃ levels occurred inversely correlated to ambient NO₂. So-called hotspot sites in both areas were characterised by the most elevated NO₂ levels and the corresponding lowest O₃ levels. The results pointed out that the existing interpolation charts for NO₂ based on fixed sampling site measurements underestimated the ambient levels at certain sites in West-Flanders. In general, the ambient NO₂ levels increased at increasing distance from the North Sea. In Antwerp, SO₂, which is independent of traffic emissions, occurred at higher levels in the southern suburbs; the nonferrous plant Umicore was indicated as a local source of SO₂.

For particulate matter (PM) the distribution of ionic compounds in different PM fractions was assessed, indicating that the water-soluble ions NH₄⁺ and NO₃^-, and Cl^- and Na⁺ occurred mainly in particles with an aerodynamic diameter smaller than 1 µm. The most abundant water-soluble ions were NO₃^- and SO₄^2-. Furthermore, in a case study in Menen, nearby the Belgian frontier of France, the bulk elemental and the water-soluble ion concentrations in PM_2.5 were assessed. It was concluded that the aerosol composition in Menen was a function of local, meteorological and seasonal circumstances; however an important influence occurred through cross-boundery air pollution, originating mainly from Germany and the U.K., which mainly affected the anthropogenic elements in PM_2.5.

In spite of the attention which is generally drawn to outdoor air pollution, several international studies have emphasised the importance of the indoor air quality, since increased indoor levels of certain contaminants have been reported in comparison to the ambient levels. Moreover, considering the fact that people spend on average 80% of their time indoors, the indoor environment has been indicated to be a more suitable estimator of personal exposures. In a limited number of countries, indoor air quality guidelines have been formulated as it is, since very recently, the case for Belgium. Two case studies were organised in which a set of particulate and gaseous air pollutants was monitored inside houses and schools (classroom and on the playground) in Antwerp and its southern suburbs. Parallel outdoor measurements were performed to assess the relation between indoor and outdoor air quality. The first case-study was performed in 19 houses in Antwerp, aiming to determine the relation between the indoor and outdoor air quality, and to identify indoor sources typical to residential buildings. Additional attention was paid to the personal exposure to gas-phase pollutants of one of the inhabitants of each house. The second study assessed the indoor air quality of 27 primary schools in the city centre of Antwerp and its southern suburbs, with the specific aim to investigate the real-time daily exposure of 7 to 8 year old school children. Attention was paid to the concentrations of air pollutants with recognized potentials to cause adverse health effects, such as the gas-phase compounds NO₂, SO₂, O₃, HAc, HFor and BTEX and respirable particulate matter (PM_2.5). Furthermore the elemental composition and elemental C content of PM_2.5 were determined.

In residential air, tobacco smoke was identified as being a major source of indoor PM mass concentrations and was most enriched in the smallest particle size fractions (PM₁). According to this case study the indoor occurrence of the elements in houses (metals, Cl and S) mainly originated from outdoor sources. A study of the elemental indoor/outdoor ratios in the houses indicated a unique micro-environment for each residence. The major indoor source of NO₂ formation was gas cooking. Indoor gases without major indoor sources were strongly correlated to ambient air. Nevertheless, the personal exposure to gaseous compounds showed a highly significant linear relationship with the indoor air and only poor or no correlations were identified with ambient air. These findings confirmed the previously formulated hypothesis, which stated that each house is characterized by its own micro-environment. Consequently, the indoor air quality of residences cannot be estimated by the use of a fixed ambient air monitoring site.

In the second case study, which was organised in 27 primary schools, much attention was paid to correlations between indoor and ambient air, in order to test the hypothesis that a strong correlation between indoor, ambient and consequently fixed monitor air pollutant levels would exist, which could then be used to support the use of a fixed monitoring site as a good estimator for personal exposure. It was found that only in the cases of BS, NO₂, SO₂ and O₃, such correlations could be established. However, due to their different deposition velocities and decomposition rates in indoor air, the respective I/O ratios of the gases varied considerably, decreasing in the sequence NO₂, SO₂ and O₃. Concerning PM_2.5 mass concentrations, BTEX, HAc and HFor, no correlation was established between the classroom and the outdoor levels, except for benzene. In spite of the fact that also for these pollutants no indoor sources were expected to be present inside the classrooms, they occurred at I/O ratios exceeding unity. For BTEX the most elevated I/O ratios were identified for toluene, followed by ethylbenzene and the xylene isomers. The highest I/O values were identified for HAc and HFor. Classroom PM_2.5 occurred at different mass concentrations and with a different elemental composition than ambient PM_2.5. In fact, the indoor mass concentrations were considerably influenced by re-suspension as a consequence of the room occupation, leading to an increased dust contribution to indoor PM_2.5. The presence of carpets inside of the classrooms was identified as having a significant influence on the I/O ratios.

In general, the use of ambient air pollutant concentrations as estimators of the indoor air quality, and consequently of personal exposures inside classrooms in the city of Antwerp and its southern suburbs, would lead to an overestimation of the exposures to NO₂, SO₂ and O₃ and an underestimation of the exposure to PM_2.5, BTEX, HAc and HFor. Thus, neither the ambient air concentrations, nor the fixed site measurements provide an accurate estimation of the children’s personal exposures to the identified air pollutants inside the classrooms. Of course one has to notice that the extent, to which the indoor air quality of a child’s classroom can be used as an indicator of its personal exposure, is strongly dependent of the child’s indoor residential air quality. As concluded in the residential air case study, e.g. parental smoking would cause highly elevated PM_2.5 levels inside the house, which would then diminish the effect of the contribution of the classroom air quality to the child’s total personal exposure. In the ideal case, personal monitoring combined with indoor and outdoor monitoring in the child’s home and classroom would give a decisive answer to this question. Unfortunately, the large dimensions of the sampling program in such a case would not allow its practical realisation.

The results of this study highlight the importance of the assessment of the indoor air quality in different environmental settings. Moreover it confirms the large variations between various indoor micro-environments and their differences from the ambient air quality. This work emphasises the importance and the urgency of a profound understanding of the impact of air pollution on human health. As for PM_2.5, scientists did not conclude yet whether the adverse health effects are caused by its composition or by its mass concentration. Furthermore, this study also stresses the value of a detailed indoor air quality legislation, which assigns limit values at reasonable, practical measurable and meaningful averaging times. Only then the indoor air quality can be evaluated and, if necessary, actions can be undertaken to reduce the indoor pollutant levels.

Polycyclic Aromatic Hydrocarbons in the atmosphere: Fast determination, concentrations, sources and health risks. KHAIWAL RAVINDRA, 2006

The Polycyclic Aromatic Hydrocarbons (PAHs) have gained attention due to their ubiquitous presence in different environments and being the first atmospheric pollutants to have been identified as suspected carcinogen. Considering this the health risk associated with the various size fractions of suspended particles with special reference to PAHs was reviewed. Some exposure studies to PAHs were appraised and the prevalence of different PAHs at selected global sites were also compared. The various known sources of PAHs and their formation were also studied to identify and characterize the emission sources of PAHs.

Environmental analysis requires PAHs levels to be determined faster than the conventional methods and at less cost. The approach of pressured liquid extraction (PLE) was evaluated for fast extraction of PAHs and to reduce the amount of toxic solvents. Further, a method was developed using low pressure gas chromatography – ion trap mass spectrometry (LPGC-ITMS) for the fast analysis of PAHs in environmental samples. The method reduces the analysis time by three factors but results in limited loss of separation power compared to that of conventional GC-MS due the occurrence of three critical pairs for high-molecular weight PAHs.

The limitation of LPGC-ITMS motivated the use of high performance liquid chromatography (HPLC) for the analysis of PAHs samples collected from various cities of Flanders. The site-specific variation in vapor and aerosol PAHs was assessed and the application of diagnostic ratio and principal component analysis (PCA) was applied to identify and characterize their emission sources. The influences of meteorological parameters on PAHs levels as well as the potential toxic fraction of PAHs in ambient air of Flanders were also calculated. The impacts of regional and global activities on PAHs levels provide useful information for the regulation and control of PAHs in Europe.

The water soluble fraction (WSF) mostly consists of ionic compounds and can comprise up to 60-70 % of total particulate mass. The WSF components may increase the solubility of toxic organic compounds, such as n-alkanes and PAHs by acting as surface active agents and hence the WSF of PM_2.5 was identified and the relation to gaseous pollutants was studied. Finally, a case study of Delhi (India), where to improve the air quality the entire public transport was amended to use compressed natural gas (CNG), was assessed. The CNG is considered to be clean fuel and hence emits less pollutants.

Characterization of indoor gaseous and particulate pollutants for conservation in museums and churches. VELICHKA KONTOZOVA-DEUTSCH, 2007

This study is focused on the environmental analysis of gaseous and particulate air pollutants in the context of the conservation of our Cultural Heritage. A variety of analytical techniques has been applied to the different types of samples. The first main topic of this work deals with the investigation of the concentration of pollutant gases and the chemical composition of fine particulate matter for the conservation of ancient stained glass windows in several famous European cathedrals and churches. The second main topic is related to the determination of outdoor and indoor generated gaseous and particulate air pollutants for the preservation of museum objects in galleries and showcases of eight museums.

In Chapter I of this dissertation atmospheric gaseous and particulate air pollutants are discussed in detail. Special attention is given to air pollutants that are potentially harmful for Cultural Heritage objects. Among the pollutants with mainly outdoor sources, NO_x, SO₂, O₃ as well as primary and secondary particulate matter is described. The main indoor generated air pollutants of interest in this work are formic and acetic acid. The most important sources and characteristics are mentioned. The chemical composition of atmospheric particulate matter is also elucidated.

Subsequently, special attention is given to Indoor Air Quality (IAQ) and factors influencing IAQ. Damage to works of art can raise from a variety of processes, including soiling by particles and chemical attack by aggressive air pollutants, but also e.g. by hygrometric stress or by biodegradation. A short overview on these processes is given. Among others e.g. the influence of the air exchange rate between indoor and outdoor air and the deposition of particulate matter to museum object surfaces is discussed. Finally the basic principles of the preservation of works of art are explained, including the assessment of the risk of deterioration of objects and possible control strategies for objects displayed in galleries and in showcases.

Chapter II gives an overview of the analytical techniques used in the framework of this dissertation. The sampling equipment comprised passive diffusion tubes from different manufacturers for the sampling of the gaseous pollutants NO₂, SO₂, O₃, formic acid and acetic acid as well as instrumentation for the sampling of atmospheric particulate matter. Bulk particles were sampled on membrane filters whereas different kinds of cascade impactors were used for the size-fractionated sampling of particles.

Ion Chromatography (IC) was applied for the determination of the compounds sampled by means of the diffusive samplers and for the investigation of water soluble compounds in aerosol particles and efflorescence crystals collected from museum objects. Bulk elemental concentrations in aerosol particles were determined by energy dispersive X-ray fluorescence (EDXRF) analysis. A large number of samples was investigated by automated single particle analysis by means of electron probe microanalysis (EPMA). Finally, samples were also analyzed with Micro Raman Spectrometry. The main advantages and disadvantages of these techniques with respect to the samples analyzed in this work are discussed.

Chapter III of this work deals with the sampling and analysis of gaseous and particulate air pollutants in three European cathedrals with the aim to assess the environmental conditions of ancient stained glass windows with and without protective glazing. The Basilica Saint-Urbain in Troyes ( France ), the Sainte Chapelle in Paris ( France ) and the Cathedral of Cologne ( Germany ) were selected for a continuous field survey with a duration of one year at each of the three monuments. The interspace between the ancient stained glass window and the protective glazing is flushed in all three monuments by indoor air, hence isolating the historic glass from the outdoor air and exposing it to indoor air on both sides of the glass panels.

Before discussing the obtained results, the composition and structure of stained glass and the glass paint (grisaille) is explained. Special attention is given to the mechanisms of glass weathering, especially under the influence of aggressive gases and aerosol particles. Additionally, the different types of (protective) double glazing systems along with their characteristics are presented.

Sampling at the three cathedrals was always carried out at several locations in the outdoor air as well as in the indoor air and inside the interspace between the historic stained glass and the protective glazing. The results from the Basilica Saint-Urbain in Troyes showed that both SO₂ and O₃ had much lower concentrations inside the Basilica than outside during the whole year, whereas NO₂ showed significantly higher indoor concentrations, which were with high probability due to burning candles in the church. CO₂-levels inside the church clearly reflected the amount of people present inside - the highest concentrations were detected on Christmas Eve and during All Saints.

The results for the particulate air pollutants show that the average indoor and outdoor concentrations of all investigated elements for all sampling campaigns were rather similar. However, different Indoor/Outdoor-ratios were found for the different sampling campaigns. The single particle analysis showed the highest abundances for soil dust particles, organic particles, CaCO₃-particles and NH₄NO₃-particles. Soil dust particles are not aggressive for the glass, but they cause soiling of the windows. Organic particles instead have to be considered as potentially dangerous because they can enhance the adsorption of aggressive compounds to the glass. Ammonium nitrate particles are also dangerous for the glass because of their hygroscopic nature. They can provoke condensation of water vapour on surfaces where these particles are deposited. Subsequently, aggressive ions are released into the solution and gases and other particles are deposited much faster on the wet glass surface leading to an enhanced attack of the glass. In general, higher abundances of organic particles and nitrates were observed indoors in comparison to outdoors. Hence, it can be concluded that the protective glazing in the Basilica Saint-Urbain is predominantly advantageous (except for NO₂) from the gases’ point of view, but it is only fairly advantageous from the particles’ point of view.

The investigations at the Sainte Chapelle, which is included into UNESCO’s list of World Cultural Heritage, showed that the concentrations of NO₂, SO₂ and O₃ were always higher in the outdoor air than in the indoor air. However, in comparison to the Basilica Saint-Urbain in Troyes , significantly higher Indoor/Outdoor-ratios were found for SO₂ and O₃, which is due to the extensive amount of ventilation in the Sainte Chapelle, especially during summer months. This led to indoor concentrations of O₃ and NO₂ of around 20 and 30 µg/m³, respectively, which can be considered as much too high for a Cultural Heritage environment. NO₂ had, in contrast to the Basilica Saint-Urbain, no important indoor source.

As far as the particulate matter concentrations are concerned, significantly higher indoor concentrations were found for all investigated elements in comparison to the outdoor concentrations. The main reason for this observation has to be seen in the high number of visitors in the Sainte Chapelle. The visitors enter the Sainte Chapelle and introduce with them rather large amounts of particles from outside attached to their shoes and clothes. These particles are released inside the building when the visitors walk around and are resuspended by them. Single particle analysis revealed again soil dust particles, organic particles, CaCO₃-particles and CaSO₄-particles. Interestingly the latter type of particles was only found inside the Sainte Chapelle which could be due to their formation in the indoor environment from CaCO₃ and SO₂ which was present in rather high concentrations. The chemical composition of the particles collected from the interspace was close to the one of the particles collected indoors. To conclude, the protective glazing can be considered as advantageous from the gases’ point of view, whereas it can not be considered as advantageous from the particles’ point of view with respect to the high indoor concentrations of particulate matter.

The results from the investigations in the Cathedral of Cologne were generally rather similar to the ones from the Basilica Saint-Urbain in Troyes . SO₂ and O₃ again had very low indoor concentrations, whereas NO₂ had with the burning candles again an indoor source, leading to NO₂-concentrations being higher indoors than outdoors. The volume of air inside the cathedral is very large and the amount of ventilation is only limited. As most of the stained glass windows in the Cathedral of Cologne are situated at least at a height of 20 m above the ground, sampling of pollutants was also carried out at this height. Elemental concentrations in bulk particulate matter was found to be significantly lower inside the Cathedral in comparison to the outside air, but this fact is probably (in part) connected to the high sampling position in Cologne. However, this result is advantageous for the stained glass windows.

Single particle analysis of the samples from Cologne showed also soil dust and organic particles as well as sulphates and nitrates, from which the latter two compounds are dangerous for the stained glass windows. Generally, no large differences in particle types were found between samples from outdoors, indoors and from the interspace. Hence, the protective glazing in the Cathedral of Cologne can be considered as predominantly advantageous from both the gases’ point of view (except for NO₂) and from the particles’ point of view.

Chapter IV focuses on the characterization of indoor and outdoor generated air pollutants in eight museums. As an introduction to airborne pollutants in museums, an overview is given on the most important pollutants in museums along with their sources and their effects on works of art. The different methods for the quantification of the exposure-effect relationship for Cultural Heritage objects are shown as well. Using the “no observed adverse effect level” (NOAEL) and the “lowest observed adverse effect dose” (LOAED), maximum permitted pollutant concentrations were calculated for different kinds of object materials and for certain preservation targets, i.e. periods of time during which no adverse effect may occur. Another point of interest is the damage of the surface of museum objects due to the formation of efflorescence.

The museums investigated in this work are situated in Brussels ( Musical Instrument Museum ), Antwerp ( Plantin-Moretus Museum (which is included into UNESCO’s List of World Cultural Heritage as well) and Vleeshuis Museum ), Kortrijk ( Broel Museum ) and Morlanweltz (Royal Museum of Mariemont). In addition to these five museums in Belgium , two museums in The Hague in the Netherlands (Prison Gate and Gallery Prince Willem V) and finally also the Metropolitan Museum of Art in New York City , which belongs to the most famous museums in the world, were investigated.

Sampling was carried out at all locations following the same principles as described before for the sampling in the cathedrals. Samples of gaseous and particulate air pollutants were always taken in the outdoor air as well as indoors. Special attention was given to the aggressive pollutants acetic acid and formic acid.

In all investigated museums, the concentrations of the mainly outdoor generated pollutant gases NO₂, SO₂ and O₃ were found to be reduced in the indoor environment, as there were no significant indoor sources of NO₂ present. For SO₂ and O₃, the most significant reduction in concentrations was found in the galleries in comparison to the outdoor values, as both gases have rather high deposition velocities to a large range of indoor surface materials. For NO₂ instead the concentrations were mainly reduced inside museum showcases in comparison to the galleries, as NO₂ does not deposit as fast as SO₂ and O₃. The concentration ratios between the outdoor air and the air in the galleries depended mainly on the extent of ventilation of the galleries. The two museums with a heating, ventilation and air conditioning system (HVAC) being operative throughout all the galleries had by far the highest I/O-ratios for NO₂ of all investigated museums.

Generally, inside showcases and other more or less sealed enclosures, the concentrations of SO₂ and O₃ were very low. For NO₂, the ratio between the concentration inside a showcase and in the gallery depended mainly on the extent of ventilation of the showcase. Apart from the ventilation rates of the showcases also the construction materials used for the showcases play an important role for the showcase/gallery concentration ratio. Showcases made entirely from metal and glass (as some of the showcases investigated in the Metropolitan Museum of Art) showed the highest showcase/gallery concentration ratio as the pollutants deposit only very slowly onto these materials.

In contrast, formic acid and acetic acid were generally detected in much higher concentrations in the indoor air in comparison to outdoor concentrations. By far the highest organic acid concentrations were found inside showcases made from wood or composite products containing wood. Also showcases with wooden objects of art showed an enrichment with organic acids in the inside air. These materials emit rather large amounts of formic and especially acetic acid. However, huge differences were observed between different types of showcases according to their ventilation and the construction materials used. An accumulation of acetic acid to a concentration of more than 1 mg/m³ was observed exclusively due to the emissions of wooden objects of art themselves. In showcases constructed from wood composite products, acetic acid concentrations of even up to 6 mg/m³ were observed.

When comparing the observed concentrations of NO₂, SO₂, O₃ and acetic acid from this work with the maximum permitted pollutant concentrations calculated from the NOAEL- and LOAED-values of the different pollutant/material-pairs, it can be concluded that actual concentrations of air pollutants inside museums often exceed the maximum tolerable concentrations by far. For NO₂, even a preservation target of only 10 years is currently not reached for a lot of objects displayed in the galleries. However, it is reached in part of the investigated display cases. For O₃, a preservation target of 10 years is reached in part of the galleries, but a target of 100 years is reached almost never. Inside showcases, the situation is again better. SO₂ profits from the serious decrease in ambient concentrations during the last decades and a preservation target of 10 years is reached for a lot of galleries and most of the showcases. Instead, acetic acid forms a serious problem in all investigated museums except for the Metropolitan Museum of Art. Inside most of the investigated showcases, acetic acid concentrations were high enough in order damage to occur within a time period of less than ten years. Several objects with a visible degradation due to the formation of efflorescence on the object’s surface were found and the chemical composition of the crystals was investigated. At least at one location the presence of an acetate salt on the material’s surface could be confirmed, proving the important role of acetic acid in the deterioration of museum objects.

The obtained results show that museum showcases protect works of art from aggressive outdoor generated air pollutants, but on the other hand they can accumulate high concentrations of dangerous formic and acetic acid. Depending on the kind of pollutant being especially dangerous for a certain object of art, the display in a showcase can improve, but also deteriorate the conditions for its conservation. Very different situations can be encountered between locations and no universal advice can be given to conservators. Nevertheless some basic principles and often reoccurring problems in Cultural Heritage environments could be identified in this work.

Application of high-energy polarized-beam energy-dispersive X-ray fluorescence for industrial and environmental purposes. KATLEEN VAN MEEL, 2009

Throughout this work, EDXRF methods for various applications have been developed. The secondary target set-up not only proved to be very useful to obtain a low background due to polarization, but it is a real asset to the versatility of the spectrometer. Not only does it allow improved sensitivity, but it also enables selective excitation. Furthermore, a combination with high excitation energy (up to 100 kV) and a suitable detector performance in the high energy region, extents the application range for high-Z elements compared to conventional EDXRF.

High-Z elements such as precious metals or heavy metals are often difficult to detect with conventional EDXRF. Due to lack of excitation energy, anode material and decreasing detection capacities in the high-energy range, conventional EDXRF is usually limited to L-lines for the determination of elements with Z starting from around 40. L-line spectra are more complicated than K-lines and there is an increased risk for overlapping, e.g. the L-lines of Cd, In and Sb may overlap with the K-lines of Ca as well as with each other. In environmental soil samples it is easy to imagine that this may cause problems when e.g. dealing with Ca-rich soils. However, with the high-energy secondary-target EDXRF spectrometer, K-lines can be used for elements up to tungsten (Z = 74) and this problem is easily overcome.

The performance of the spectrometer is in some cases comparable or even better than classical reference techniques. The precious metals used in ACCs (Pt, Pd and Rh) can be determined in this matrix with equal precision and accuracy as the reference ICP-OES method. On top of this, the EDXRF procedure has the benefit of being cheaper and faster. Detection limits of around 5 ppm for this application (for a 500 s irradiation) demonstrate the improved sensitivity due to proper secondary target choice. In heavier matrices such as NiS-beads the technique is still applicable, but the performance is less spectacular. The method is faster and cheaper than an ICP-OES method, which requires an additional digestion and co-precipitation step, but the detection limits are not competitive in all materials due to large dilution (30 times or more). In very complex matrices, on the other hand, not even the secondary targets can provide an easy one-preparation-only solution for the determination of all elements.

Also in environmental applications, the improved spectrometer can provide elegant solutions for delicate analysis. Thin film analyses are very straightforward on EDXRF devices. This sample preparation can be used for e.g. air and water analysis. Aerosol filters already have low backgrounds for conventional EDXRF, but the polarizing geometry of the spectrometer provides a practically horizontal background for most secondary targets, allowing even lower detection limits than before. The higher energy and the extended applicability of the K-lines also enable better detection of harmful heavy metals and detection limits of a few ng/m³ are achieved. Determination of Cd in saline solutions is often difficult and preconcentration on a thin film sample (membrane) can offer an easy and cheap solution. The obtained detection limit, 700 ppb, is competitive with other reference methods such as ETAAS and ICP-AES. In the sediment fingerprinting application field, heavy and rare earth elements (such as e.g. Ce) can be interesting tracers, making it desirable to be able to detect them at low levels. A robust method has been optimized for the determination of a large variety of elements in Belgian soils and sediments, using a combination of SRMs and secondary standards for the calibration.