Organic aerosols are more and more in the focus of aerosol analysts. The simple reason for it is the inorganic part of the ambient aerosol can be measured easily with existing techniques (sampling & analysis). This is due to the thermo-dynamical stability of most inorganic compounds. Many measurement campaigns yielded a percentage between 50 – 70 % for Total Organic Carbon (TOC) without knowing in detail the individual composition. Within the last 20 years, also, the problem of artefact formation became evident; e.g., interaction of organic trace compounds like PAH with reactive trace gases (ozone, nitrogen dioxide) during the enrichment step. Many attempts are reported to avoid this by means like denuder tubes or application of protective group reactions. There is general agreement within the community that in situ and on-line analysis would be the optimum strategy for observing labile compounds.

A second obstacle with organic trace compounds is the enormous variety given. Not only do many isomers exist, even different modifications of the main element, carbon, causes enormous difficulties. Additionally, nature produces biogenic compounds of considerable complexity, like debris from living materials (organisms, plants etc.) or the microorganism itself. From the toxicological point of view it became also evident, that a non-negligible health risk is often connected with the presence of organic aerosols.

A good example is soot aerosol characterization in Europe. Not only the suspected health impact, but  also the importance of light-absorbing particles within the surrounding light-scattering tropospheric and stratospheric aerosol (related to the radiation budget) stimulates the need for reliable means to analyse its mass fraction within the ambient aerosol. Selected was a thermal technique (combustion at 650^oC under oxygen flow, with a previous extraction/desorption of adsorbed organic substances). This technique (VDI Guideline 2465, Part 1) was adapted from the occupational hygiene people, since the Air Quality Criteria at that time were already officially published without knowing the appropriate measurement technique. The Institute of Hydrochemistry at Technical University of Munich (Prof. Niessner) got the task from the German EPA to evaluate this so-called coulometric technique (the evolved CO₂ is measured coulometrically) for its intended use as soot measurement technique. Beside the coulometric approach, the applicability of the British Black Smoke technique (determination of light reflection), the US light transmission technique, and the aerosol photoemission became thoroughly tested. Different locations in Germany, partly heavily impacted by traffic, as well as rural sampling sites, were characterized for three years in parallel using the different techniques. The outcome was critically assessed by the VDI panel, published 1995 (A. Petzold and R. Niessner; Mikrochimica Acta 117, 215 – 237)

Two competing methods, light reflection and light transmission, exhibited serious drawbacks: a) site dependency of calibration (transmission); and, b) large data scattering at low concentrations (< 5 µg elemental carbon (EC)/m³ ), in case of the Black Smoke Method (reflection). Aerosol photoemission, originally developed for on-line PAH monitoring (e.g., R. Niessner; J. Aerosol Science 17 (1986) 705 – 714), showed a tremendous sensitivity to changing humidity and source distance. Since a mass–related signal was needed, the finding of water influencing the surface–related aerosol photoemission was not surprising. Aerosol particles, either covered by non-photoemissive water molecules, or possessing a varying PAH profile on the particle surface, can´t be expected to show a constant response. This was validated in independent studies (e.g. R. Niessner et al.; Analytical Chemistry 62 (1990) 2071 – 2074).

Within the following years the chosen thermal “VDI” technique came under heavy discussions. Many other groups, mainly from Austria, France and US, started Round Robin tests with aerosol samples from different locations. They compared the thermo-optical technique, a combination of transmission measurement and combustion, with combustion & CO₂  detection, reflectometry, and transmission measurement without combustion. Tested were several extraction procedures, too, before combustion. It became obvious, that non-elemental carbon (cell wall debris, pollen, proteins, cellulose, lignin etc.) contributes substantially to the EC determination, when samples from less polluted sites are analysed. An additional bottle-neck was the tedious extraction and combustion procedure limiting the daily throughput to about 20 filter samples per day. So the aerosol community started again searching for more reliable methods. From 1997 until now, various techniques show a renaissance : a) reflection under various observation angles and light wavelengths, b) photoacoustic spectroscopy; and c) Raman spectroscopy.

It was Petzold et al. (Atmos. Environ. 31 (1997) 661–672), who started the systematic search for using transmission and/or reflection measurement as a cheap and reliable technique for EC determination. It soon became clear, that the aerosol deposition within the filter matrix and angle and wavelength depending particle/light interaction have a tremendous effect on the signal strength. The current status is best described by the recent German Patent DE 102 402 04 B3 from the year 2004 (A. Petzold & M. Schoenlinner). They use an arrangement of photo-diodes operated at observation angles of 0^o, 120–140^o, and 165–180^o, cancelling out most of the size effects from particles and filter fibers. A commercialized system (Thermo Eberline ESM, Erlangen, Germany) is available in the meantime. Several comparison tests of the VDI method 2465 with the Eberline technique are reported as successful. The measurement uses a moving filter tape, recording and storing time-resolved aerosol deposits for further investigations.

Photoacoustic spectroscopy (PAS) applied to soot samples has its origin in the US and dates back to the 70ties. Several attempts with laser irradiation (at a wavelength of about 480 nm) were published very early. The main advantage of PAS is the linear relationship between light intensity and the observed pressure signal as a consequence of modulated light absorption by black carbon. Secondly, by definition and experiment, no contribution of light scattering to the signal formation will be observed. After some years experience with Ar-ion lasers as light source, PAS analysis of soot aerosol became no longer prosecuted. The reason was the strong interference by NO₂ and H₂O as trace gases in the soot aerosol. A lot of attempts were made by diffusion denuders or parallel arranged PAS cells with and without particles under illumination to avoid this interferences. A second difficulty was the shift of the resonance frequency observed with strong laser beams. Furthermore usage of chopper wheels inhibited any chance to correct this effect. With the advent of robust tuneable diode lasers in the mid of the 90ties the situation improved a lot. Petzold & Niessner (Applied Spectroscopy 66 (1995) 1285–1287) were the first who used diode lasers at a wavelength of around 800 nm. The influence of NO₂ and water vapour became negligible. Presently, this technique is successfully used for fast monitoring of Diesel soot emission. Detection limits are in the lower µg-range for black carbon (H. Beck et al.; Analytical and Bioanalytical Chemistry 375 (2003) 1136 – 1143) at a time resolution of 1 sec. A similar arrangement became successfully demonstrated for ambient air monitoring (L. Krämer et al.; Analytical Sciences S 175 (2001) s563 – s566). Time resolution was set to 5 min, and a detection limit of 0.5 µg Black Carbon per m³ was achieved.

So far, determination of soot aerosol is of some arbitrariness. If only “black” properties are requested, the optothermal approach (e.g. PAS) seems to offer highest performance. On the other hand, it is quite clear, that any particulate material offering a measurable light absorption at the selected wavelengths, e.g. dark minerals or wood residues, will contribute to some extend to the detection signal.

This limitation has led the community very early to the use of Raman spectroscopy for carbon species identification. The Raman effect gives access to the chemical and physical structure of carbon particles. T. Novakov et al. (1976) were the first reporting about the use of Raman spectroscopy for carbon identification. Unfortunately the Raman absorption cross section is rather low. Strong laser sources are therefore a must. Strong PAH fluorescence and light scattering in general avoided a breakthrough of the technique up to now. The new notch filter technology, paired with strong laser sources in the NIR opened within the last years first promising opportunities. By means of Raman spectroscopy a clear identification of graphitic or distorted carbon within the particle carbon lattice is possible. Twinned mineral phases within a particle conglomerate can be identified and quantified. First characterization experiments with NIST Diesel soot standard 1650 shows the direction where to go to (A. Sadezky et al. (2004) Raman Spectra of Soot : Spectral Analysis and Structural Information. Submitted to PCCP). Knowing the intrinsic structure of soot (degree of graphitization) gives also new possibilities for source apportionment. Latest results from Diesel engine developments in Europe indicate a change of the crystalline carbon structure along the course to initiate fast oxidation within exhaust after-treatment means. Distorted graphite structures show an oxidation under air at temperatures below 200^oC (A. Messerer et al. (2004) Topics in Catalysis 30/31, 247 – 250; D. Su et al. (2004) Microstructure and Oxidation Behaviour of EURO IV Diesel Engine Soot : A Comparative Study with Synthetic Model Soot Substances. Catalysis Today, in press). This explains at least in part the observed failure of thermal and/or thermo-optical techniques for soot determination.

Spectroscopical techniques, especially when laser light is used,  possess an intrinsic high potential to serve as in situ and on-line analysis technique : they operate with (hopefully non-reactive) light, the light beam can be directed to the location of the problem, and the appropriate spectroscopical method has many degrees of freedom, which means a certain selectivity can be expected. Up to now, some laser-based techniques have been applied to organic aerosols : fluorescence for PAH detection, photoacoustic spectroscopy for soot, laser photofragmentation for N-, P- or S-species, and aerosol photoemission for PAH detection, too. In all cases, interferences by interaction of the analyte molecule with the particle core is seen. Only in rare cases, e.g. when desorption of the analytes is possible without destruction, the direct and undisturbed observation within the gaseous state is possible, e.g. PAHs (U. Panne et al. (2000) Fresenius´ Journal of Analytical Chemistry 366, 408). In general, the solid state yields less distinct, broad and often featureless spectra. Raman spectroscopy, so far applied on suspended single particles, is not sensitive enough. Fortification of the signal by the stimulated Raman effect is only possible after contacting an appropriate substrate. Photoacoustic spectroscopy offers a strong potential, but its sensitivity is limited by using favourable absorption bands in the optical spectrum. Similarly suited is Cavity Ring-down Spectroscopy (J. Thompson et al. (2003) Analytical Chemistry 74, 1962 – 1967) offering detection limits (expressed as extinction coefficient) down to 10^-9 cm^-1.

A different issue is the determination of  analytes with a molecular weight  > 1 kDalton. Presently only wet chemical group analyses are applicable. To determine biogenic organic compounds in the ng/m³ range, large volume sampling to get some milligrams aerosol mass is needed. Classical digestion steps, followed by amino acid or sugar analysis are applied. Typical separation tools, e.g. polyacrylamid gel electrophoresis, lack sensitivity and sometimes selectivity. Quantification is often impossible due to transfer problems of  the separated proteins. Mass spectrometry needs the analyte in the gaseous state, and the change-of-state (solid to gas phase) makes enormous problems, even with ESI-TOF-MS or MALDI techniques.

A different approach is the usage of high-affinity targets, like antibodies of molecular imprints. With the need to get rapid information on bioaerosol concentrations in air, the combination of  quasi-continuous sampling (e.g. wetted wall cyclone) connected with rapid bioanalytical identification of the sampled material was developed. Not only detection of cell wall proteins by antibodies is common in the meantime, also the identification and quantification of  difficult (in terms of stability and low concentrations) large organic molecules became feasible. Good examples are the determination of pollen proteins by immunoassays (T. Franze et al. (2003) Analyst 128, 824 – 831). Within 30 min the complete analysis in the pg/m³ range is done. Prerequisite is a clear strategy about what has to be analysed. Also the respective antibody must be available. Once this is fulfilled, very cost-efficient, fast and reliable analysis is possible. Upcoming chip technologies with multiple recognition targets in a high-parallel arrangement, combined with a fast read-out technique, allow rapid determination of many analytes within several minutes (e.g. A. Knecht et al. (2004) Analytical Chemistry 76, 646 – 654). Very interesting applications are recently presented by D. Blake et al. (2003) Biochemistry 42, 497. They developed antibodies directed towards organometal species (e.g. Cd – complexes). Currently the analytical community is developing artificial “antibodies”, so-called molecular imprints, for enhancing the performance of protein separation. The aerosol scientists will certainly make use of these developments within some years.

Conclusions:

From my point of view I see the following trends and needs:

• Analytes of molecular weight > 1 kDalton will become more important, to fill the gap within aerosol mass balance.

• Biogenic material is extremely complex. Detection and quantification is possible with typical bioanalytical techniques (immunoassay, polymerase chain reaction etc.). Combination with continuous sampling (wetted wall cyclone) is advantageous.

• Laser spectroscopy seem suited well, but compromises on selectivity have to be made as long as in situ observation is demanded. Also lack of strong photon-absorbing chromophor groups prevents successful use of spectroscopy.

• Polymorphism (e.g. elemental carbon in its different modifications) is a problem. Discrimination of internally mixed substances and the three-dimensional analysis of individual particles remains a strong challenge for future.

---