Antonella Macagnano

An electronic nose is now becoming available as a commercial product. Nevertheless its performances are not fully understood and interpreted. Also the differences between electronic noses and the human olfaction have not yet been sufficiently studied. This is an important issue in many industrial sectors, such as food analysis. In this paper a comparison between the performances of an electronic

Soil is a complex ecosystem comprised of several and mutually interacting components, both abiotic (organo-mineral associations) and biotic (microbial and pedofaunal populations and plants), where a single parameter depends on other factors and affects the same and other factors, so that a network of influences among organisms coexists with the reciprocal actions between organisms and their environment. Therefore, it is difficult to undoubtedly determine what is the cause and what the effect within relationships between factors and processes. Soil is commonly studied through the evaluation and measurement of single parameters (e.g. the content of soil organic matter (SOM), microbial biomass, enzyme activities, pH, etc.), events (e.g. soil erosion, compaction, etc.) and processes (e.g. soil respiration, carbon fluxes, nitrification/denitrification, etc.), often carried out in laboratory conditions in order to limit the number of factors acting within the ecosystem under study, but missing the information about the global soil environment that way. In the last decade, several scientists have proposed and suggested the need for a holistic approach to soil ecosystems in different contexts. Recently, we have applied a sensing system developed in the last decades and capable of analysing complex mixtures of gases and volatiles (odours or aromas) in atmospheres, namely called electronic nose (EN). Typically, ENs are devices consisting of an array of differentially and partially specific, despite selective, sensors upon diverse coatings of sensitive films, i.e. interacting with single analytes of the same chemical class, despite not highly specific for a single substance, only, but showing also lower extent of cross-selectivity towards compounds of other chemical classes. ENs can be used in the classifications of odours by processing the collected responses of all sensors in the array through pattern recognition analyses, in order to obtain a chemical fingerprint (olfactory fingerprint) typical of the analysed air sample. Due to these features, we decided to apply such a sensing technology to the analyses of soil atmospheres, because several processes in soil, both abiotic and biotic, result in gas and/or volatile production and the dynamics of such releases may also be affected by several additional environmental factors, such as soil moisture, temperature, gas exchange rates with outer atmosphere, adsorption/desorption processes, etc. Then, the analysis of soil atmosphere may provide information about global soil conditions (e.g. soil quality and health), according to a holistic approach, where several factors are contemporarily taken into account. At the same time, the use of such a technology, if adequately trained on purpose, can supply information about a single or a pool of processes sharing similar features, which occur in soil over a certain period of time and mostly affecting soil atmosphere. According to these premises and hypotheses, we demonstrated that EN is an useful technology to measure soil microbial activity, through its correlation to specific metabolic activities occurring in soil (i.e. global and specific respiration and some enzyme activities), but also soil microbial biomass. On the basis of such evidences, we also were able to use this technology to assess the quality and health conditions of soil ecosystems in terms of metabolic indices previously identified, according to some metabolic parameters and biomass quantification of microbial populations. In other studies, we also applied EN technology, despite using a different set of sensors in the array, to analyse the atmosphere of soil ecosystems in order to assess their environmental conditions after contamination with polycyclic aromatic hydrocarbons (PAHs) (i.e. semivolatile - SVOCs - organic pollutants). In this case, EN technology resulted capable of distinguishing between contaminated and uncontaminated soils, according to the differences in a list of substances, occurring in the atmospheres of differently treated soils, which were identified through SPME-GC/MS analyses and then suggested to be responsible for the different classification. Analysing the EN responses, it was also possible to follow the degradation process of pollutants by resident microbial populations over time, on the basis of the contemporary decrease of contaminant and the increased release of CO2. Then, we suggest that EN technology may be usefully employed in the analyses of soil ecosystems in order to both supply information about global soil environment, according to the holistic approach, and about specific processes occurring therein. Furthermore, since EN technology resulted to be effective and successful in detecting processes in soil, in both natural and perturbed conditions, involving microbial populations, which are commonly considered as the most sensitive and responsive to soil environmental modifications, we suggest it might be reasonably employed in analyses concerning the assessment of soil quality and health. Consequently, such a technology may also be used to study several processes involving soil ecosystems, such as soil management practices, soil restoration, soil contamination and remediation, soil fertility, etc.

To study and monitor environments, a plethora of sensors in last decades have been proposed and claimed to be as the most specific, sensitive, reliable, durable, affordable or whatever. However, they rarely take into account probable interactions of compounds of interest with other substances (i.e. molecules, matrices, surfaces, etc.) occurring in the environments where the analytes are present (although some corrections due to a few interfering compounds have been sometimes carried out), then, generating misinterpretations of results (e.g. overestimation or underestimation) or incorrect evaluation of effects (e.g. about toxicity and disease diagnoses). Another quite rare evaluation in the detection of analytes in environments concerns the partition of substances of interest into different phases, as well as adsorption/desorption and absorption/release events, thus often leading to misinterpretations of results. That issue is of outmost importance in complex multiphasic environments, such as soil, where these phenomena commonly occur. An improvement in sensor applications to environmental monitoring, as concerns the competition and interference of other compounds in measurements, has been the development of electronic noses. The electronic nose (E-nose) is a sensing technology, where the presence of arrays of several suitable but unspecific sensors for volatiles and gases can deal with this problem, since the different features of sensors, despite overlapping responses to different compounds, are then evaluated in post-measurement data processing analyses (namely multivariate analyses) and integrated into a chemical image reproducing the fingerprint of the sample headspace or atmosphere (i.e. the odour), such as occurs in the olfactory system of mammalians. E-noses in the last decade have been extensively used to monitor volatile and gaseous analytes and odours in several contexts and environments. In the last 5 years, a very few groups have applied this technology in studies on soil science to detect and monitor the presence of natural or anthropogenic compounds (i.e. nutrients and pollutants), soil processes (i.e. pollutant degradation and metabolic activity) and soil quality assessment (through the use of suitable metabolic indices), on the basis of volatiles and gases detection, and some devices have been set up on purpose. The E-nose approach, can be remarkably useful in the study of soil environments, because it may supply a holistic image of the entire soil ecosystem under study, with respect to very specific, accurate and sensitive but partial measurements that other sophisticated techniques commonly provide, taking into account both biotic and abiotic processes, as well as interactions between different populations and communities of living organisms. Very recently, a different approach has been developed in pollutant monitoring, in order to relate the quantification and behaviour of contaminants in soil (e.g. solubility, volatility, phase partitioning, adsorption and desorption, etc.) to the relative environmental conditions, by measuring chemical (pH) and physical (temperature and moisture) parameters, which can affect such processes, while contemporarily detecting pollutants in soil atmosphere. According to this approach, an E-nose-like multi-parametric hybrid probe has been set up, where several types of sensors have been included, some of them in an array for pollutant detection and some others peculiar to different features concerning the site-specific environmental conditions where those compounds are present. Such a kind of device can permit, through the additional support of suitable algorithms and statistical analyses, the quantification and monitoring of toxic substances taking into account their behaviour in their specific environmental conditions, then providing more reliable quantification of substances and supplying a more sound interpretation of results, which can be useful in decision making processes.