Rapid detection and respirometric profiling of aerobic bacteria on panels of selective media



Dmitri Papkovsky, Laboratory of Biophysics and Bioanalysis, Department of Biochemistry, University College Cork, Cavanagh Pharmacy Building, 1.28, College Road, Cork, Ireland. E-mail: d.papkovsky@ucc.ie



To evaluate high-throughput optical oxygen microrespirometry for selective detection and predictive identification of aerobic bacteria.

Methods and Results

Using GreenLight probe, standard 384-well plates and time-resolved fluorescence reader, a representative panel of 16 partially selective media and 9 aerobic bacteria (Escherichia coli, Bacillus cereus, Staphylococcus aureus, Campylobacter jejuni, Yersinia enterocolitica, Pseudomonas aeruginosa, Streptococcus pyogenes, Salmonella typhimurium and Listeria innocua) were analysed. For each medium, bacterial strain and dilution, growth profiles were recorded, from which calibrations, doubling/generation times and growth patterns in different media were determined. Analytical performance, selectivity and general usability of the method were assessed, and mixed bacterial cultures were analysed.


The microrespirometry platform facilitates simple, real-time detection and predictive identification of aerobic bacteria by looking at the patterns of their growth and respiration in several media and determining their growth and doubling times.

Significance and Impact of the Study

The new screening method can be used for routine microbiological analysis and testing of aerobic bacterial cultures as well as complex food, environmental and clinical samples.


There is an increasing need for simple, rapid, sensitive and selective methods that provide reliable detection, identification and enumeration of viable bacteria in complex samples. Such methods are particularly important for food safety, medicine and environmental monitoring (Kennedy et al. 2011; Maheux et al. 2011), for the detection of common and emerging foodborne pathogens, tracing outbreaks of microbial contamination, correct diagnosis of disease, microbial forensics, bioterrorism threats, etc. In such samples, microbial pathogens are often present at low numbers and together with other bacteria, which makes their specific detection difficult.

Polymerase chain reaction (PCR), antibody-based tests, flow cytometry and cultural methods are the main techniques currently used to detect microbial pathogens (Jasson et al. 2010). PCR allows for the detection of a single copy of target DNA/RNA sequence and a single bacterial cell in food sample (Batt 2007). Antibody-based methods (Iqbal et al. 2000) and flow cytometry (Flint et al. 2007) can also detect pathogens and their toxins with good selectivity. However, these methods often require complex and time-consuming sample preparation and enrichment to achieve the desired sensitivity and performance, are still prone to false-positive and false-negative results, have high cost and limited sample throughput and require skilled personnel.

Culture methods based on counting of viable cells and colonies grown on selective media still remain among the most reliable and commonly used techniques for detection of bacteria. Various selective media are used to culture and identify particular bacterial species, based on the metabolic features of each particular strain. Putative identification of a bacterial species on a particular selective media (usually solid agar) is usually coupled with a confirmatory test, for example a colour reaction. In recent years, a number of rapid (or alternative) microbiological testing methods and selective media for particular pathogens have been developed. Examples of such modified cultural methods for identification of bacteria include Petrifilm™ (3M), Compact Dry (Nissui), SimPlate (BioControl), Colilert® (IDEXX), Soleris™ (Jasson et al. 2010), and the automated systems TEMPO® (bioMerieux), BD Phoenix (Funke and Funke-Kissling 2004) which allows identification of Gram-positive bacteria from pure culture (Staphylococcus, Enterococcus genera and some other cocci and bacilli), the BACTEC (van Griethuysen et al. 1996), BacT/ALERT (Thorpe et al. 1990). Various biosensor-based systems are also coming to the fore, which enable the development of simple and automated analytical systems for pathogen detection (Jasson et al. 2010).

Oxygen is a growth requirement and a key metabolite of aerobic cells, and O2 consumption can serve as a useful marker of cell proliferation metabolic status and responses to various stimuli. The possibility of contactless monitoring of dissolved O2 by means of phosphorescence quenching technique facilitated the use of this optical oxygen sensor technology in a number of applications, including microrespirometry (Papkovsky 2005). Optical microrespirometry, which relies on dedicated oxygen-sensing probes added to test samples, has been used for the analysis of mammalian cells (Hynes et al. 2006), isolated mitochondria (Will et al. 2006), and small organisms and embryos (O'Mahony et al. 2005). This methodology is also very efficient in the detection of aerobic bacteria. In particular, it has allowed simple and rapid determination of total viable counts (TVC) in complex food samples (crude homogenates of raw meat and fish), with the use of standard 96-well plates and commercial fluorescence reader (Hempel et al. 2011). The assay combines high degree of automation, instrumental readout, and minimal labour and space requirements. When compared with the standard agar plating method, the assay showed similar performance in regard to assessing TVC, but significantly outperformed the plating method when considering flexibility, throughput and speed (O'Mahony and Papkovsky 2006). A few other systems that use special dyes and multiwell plate format to monitor microbial growth and determine TVC have been described. One such method utilizes the reduction of a tetrazolium salt XTT {2,3-bis[2-methyloxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-car-boxanilide} by metabolically active cells to a coloured water-soluble formazan; the observed colour change is proportional to the metabolic activity representing the quantity of live bacteria present (Tunney et al. 2004). The use of this assay in highly coloured and/or turbid media is problematic. Another BacLight system (Invitrogen, Carlsbad, CA) employs two nucleic acid stains that differ in their ability to penetrate healthy bacterial cells. Green fluorescent SYTO9 can enter all cells and is used for assessing total cell counts, whereas red-fluorescing propidium iodide dye enters only cells with damaged cytoplasmic membranes (Berney et al. 2007). However, these dyes are genotoxic and can influence cell growth.

In this study, we describe the application of high-throughput O2 microrespirometry to the analysis of common foodborne pathogens on panels of culture media in 384-well plates. By monitoring probe fluorescent signal (preferably lifetime or intensity) over time, discrimination between different bacterial species via their characteristic respiration patterns in different media can be achieved, and the number of bacteria in the original sample and their doubling rates can be determined accurately.

Materials and methods


GreenLight™ O2-sensitive probe and mineral oil were from Luxcel Biosciences (Cork, Ireland). All culture media and supplements were from Oxoid/Thermo Fisher (Dublin, Ireland). Escherichia coli NCTC9001, Bacillus cereus NCTC11145, Staphylococcus aureus NCTC 6571, Campylobacter jejuni NCTC1162, Yersinia enterocolitica NCTC11599, Pseudomonas aeruginosa NCTC 13359, Streptococcus pyogenes NCTC 10876 and Salmonella typhimurium ATCC14028 were used. Listeria innocua was obtained from Food Technology Department, University College Cork. Mylar sealing film was from Sigma-Aldrich (Arklow, Ireland). All other chemicals and solvents were of analytical grade, and solutions were prepared using Milli-Q-grade water (Millipore, Carrigtwohill, Ireland). Sterile 384-well microtitre plates made of clear polypropylene were from Eppendorf (Stevenage, UK).

Selection of bacterial species and growth media

In this proof-of-concept study, representative panels of bacteria and selective media were selected based on a number of criteria. The bacteria have to be nonpathogenic strains of well-known pathogenic bacteria causing foodborne poisoning, which require standard biosafety level 1 and level 2 laboratories, are able to grow under aerobic conditions and have different affinities to O2. Several micro-organisms commonly found in food products belonging to a group of aerobes and facultative anaerobes and those requiring microaerophilic conditions (Campylobacter) were thus selected, while obligate anaerobes such as Clostridium were excluded. The media were selected based on the micro-organisms' panel: one or two media selectively promoting growth of each particular bacterium. Several nonselective media were also included. Other considerations were that the media have to be well known and available in broth form. The list of media used in this study is presented in Table 1.

Table 1. Summary of selective media used in the study and their features
MediumSelective componentDifferentiating componentUsed for selective growth of (oxygen affinity)
High-salt-nutrient broth (hsNB)High salt (osmotic pressure) Staphylococcus aureus (obligate aerobe)
Mannitol–egg yolk–polymyxin (MYP)Polymyxin BMannitol, Phenol RedBacillus cereus (facultative aerobe)
Bacillus cereus selective medium (PEMBA)Polymyxin BMannitolB. cereus (facultative aerobe)
Preston Polymyxin B, Rifampicin, Trimethoprim, Cycloheximide Campylobacter jejuni (micro aerophilic)
Bolton Sodium Metabisulfite and sodium pyruvate Camp. jejuni (microaerophilic)
Tetrathionate (TTH)Thiosulfate, Tetrathionate Salmonella typhimurium (aerobic)
Selenite–Cysteine (SC)Sodium Selenite, l-Cysteine Salm. typhimurium (aerobic)
MacConkey (MAC)Bile saltsLactoseGram-negative
Cetrimide (CETR)Cetrimide Pseudomonas aeruginosa (aerobic)
Yersinia selective medium (CIN)Sodium deoxycholate, cefsulodin, Irgasan Yersinia enterocolitica (aerobic)
Listeria enrichment broth (LEB)Nalidixic acid, Cycloheximide  Listeria innocua (facultative anaerobic)
ECBile saltsLactoseFaecal coliforms
M 17pHDisodium GlycerophosphateStreptococcus pyogenes (aerobic)
Luria–Bertani broth (LB)Rich medium  
Buffered peptone water (BPW)Rich medium  
Nutrient broth (NB)Rich medium  

Preparation of microbial cultures and samples

A set of selective media was prepared in 500-ml flasks and sterilized by autoclaving. Colonies of bacterial strains plated on selective media plates were collected and suspended in 100 ml of LB broth in a 500-ml flask and grown at 30°C on a rotary shaker at 250 rev min−1 until OD600 of c. 0·8 was reached (typically overnight). Campylobacter jejuni was cultured in NB broth according to Mohammed et al. (2005). The cells in bacterial cultures (stocks) were enumerated using Alphaphot-2 YS2 microscope (Nikon Instruments, Amsterdam, Netherlands) and improved Neubauer hemocytometer (Assistant, Sondheim, Germany) and used immediately to prepare the desired dilutions (107–101 cells ml−1) in corresponding media.

GreenLight probe (1× package) was reconstituted in 100 μl of PBS to give 10 μmol l−1 stock, which was stored in the dark at 4°C until further use. Aliquots of different media (3–10 ml) were taken in sterile 15-ml plastic vial (Sarstedt), and GreenLight probe was added to it to produce a final concentration of 100 nmol l−1 (1 : 100 dilution). These solutions were dispensed into the wells of a 384-well plate (90 μl in each well). Each bacterial dilution was then added to assay wells (10 μl per each well). Typically, 3–6 replicates were prepared on the plate for each condition, and pure medium was added to the wells served as negative controls. The samples were sealed by applying heavy mineral oil (40 μl per each well), and then, the plate was measured kinetically on a fluorescent reader. To reduce variability of results, plate preparation time was standardized and kept to minimum (15–20 min per plate).

Optical measurements and data analysis

Measurements were taken on a Victor V multilabel reader (Perkin-Elmer Life Sciences, Waltham, MA) using time-resolved fluorescence (TF-R) mode, standard 340-nm excitation and 642-nm emission filters. Two TR-F intensity signals F1 and F2 are measured at different delay times t1 = 30 and t2 = 70 μs using gate time 100 μs, and lifetime is calculated according to the following formula: τ = (t− t1)/ln(F1/F2) (Waters and Burns 1993). The plate was monitored kinetically over 12–24 h at 30°C, reading each well on the plate every 5–15 min. The resulting time profiles of the phosphorescence lifetime (τ) for each sample/condition were analysed, and threshold times (TT) required to reach threshold lifetime signal of the probe (set at 35 μs) were determined and used as readout parameter to generate calibration curves (seeding density vs TT) and determine doubling/generation times for different bacteria and media. When testing probe performance in different media, TR-F intensity (F1) and lifetime signals were analysed in both air-saturated (pure medium) and deoxygenated condition (medium with an excess of microbial growth).


Assessment of probe performance in different media

The effects of different media on the performance of the new GreenLight probe in respirometric measurements are shown in Table 2, which show the ability of this probe, which represents an improved, shielded modification of the first-generation probe MitoXpress™ (O'Mahony and Papkovsky 2006), to produce high-phosphorescent-intensity signals in different media at air-saturating conditions (basal intensity signals and blanks), and robust response to oxygen depletion via changes in intensity and lifetime signals. For most of the media, basal intensity signals were in excess of 150 000 cps (counts per second), for Preston and Bolton media (both containing leaked horse blood, giving them deep purple colour), they were reduced sevenfold to 10-fold, but at the same time, signal-to-blank ratio for all the media still remained high (>70). Therefore, the probe can be used for monitoring (de)oxygenation of all the media. At the same time, the use of intensity readout for quantitative assessment (TT determination) is cumbersome. The large variation of basal and maximal intensity necessitates the use of different threshold signals for each medium (determined experimentally). In contrast, phosphorescence lifetime signals of the probe in different media were stable, under both oxygenated and deoxygenated conditions. Stable lifetime readings also prove that none of the selected media have quenching interferences or cross-sensitivity on GreenLight probe. Therefore, this parameter is more convenient and reliable for quantitative assessment: one threshold value can be used for all the media for the determination of TT values. In our case, it was set at 35 μs and used in further studies.

Table 2. Performance of GreenLight probe in the presence of different culture media. TR-F signals: intensity and lifetimea
MediaIntensity (cps × 1000)Lifetime (μs)
  1. a

    Signals measured using 100 μl of medium and 100 nmol l−1 probe concentration.


Microbial growth and respiration in different broths

The ability of the selected panel of bacteria to grow in different media was assessed, by monitoring their growth in test wells of a 384-well plate at several serial dilutions (final concentrations ranging from 107 to 10CFU ml−1) at 30°C. Typical profiles of probe intensity of lifetime signal over a time scale for respiring bacteria are shown in Fig. 1. Sigmoidal shape of probe signal reflects the process of sample deoxygenation, which is determined by the initial number of bacteria and their proliferation rate. After the initial temperature equilibration, probe signal stays constant (flat part of the profile) and corresponds to oxygenated (air-saturated) conditions. When bacteria reach a certain concentration (c. 10CFU ml−1) as a result of their exponential growth in the medium, a sharp signal increase is produced, which reaches its highest value and levels off (see e.g. profile in LB). In some cases, full deoxygenation of sample is not observed (see e.g. profiles of Ps. aeruginosa in Cetrimide, Staph. aureus in LB), and this suggests that these bacteria are dying or, in case of facultative aerobes, switching respiration to anaerobic metabolism when oxygen concentration becomes too low. Again, respiration profiles in phosphorescence lifetime scale are more smooth, reproducible and easy to interpret, while intensity profiles showed larger variation and susceptibility to optical interference (changes in turbidity or colour of the medium occurring during the measurement). Overall, respiration profiles can be regarded as an important characteristic of individual bacterial species and medium. Furthermore, respiration profiles in different media, that is, sets of curves or processed patterns of growth, can serve as metabolic fingerprints or signatures of different bacteria, which can be used for their selection and identification.

Figure 1.

(a) Example profiles of oxygen depletion over time recorded for Pseudomonas aeruginosa in selected broths. (image_n/jam12049-gra-0001.png) EC; (image_n/jam12049-gra-0002.png) LB; (image_n/jam12049-gra-0003.png) SC; (image_n/jam12049-gra-0004.png) LEB; (image_n/jam12049-gra-0005.png) CETR and (image_n/jam12049-gra-0006.png) TTH. (b) Pseudomonas aeruginosa, Staphylococcus aureus, Bacillus cereus and Salmonella typhimurium in LB broth. For both A and B, profiles are presented in intensity (left, cps) and lifetime (right, μs) mode. (image_n/jam12049-gra-0001.png) Psaeruginosa; (image_n/jam12049-gra-0002.png) Staphaureus; (image_n/jam12049-gra-0003.png) B. cereus and (image_n/jam12049-gra-0004.png) Salmtyphimurium.

In this work, we operated mainly with TT values that provide accurate, quantitative and automated assessment of bacterial growth. For each dilution, the TT value (h) was determined, and the relationship between seeding density (cells ml−1) and TT was plotted, from which corresponding doubling times (DT) were calculated from linear regression fits. The results are summarized in Table 3. Standard 384-well plates and simple, automated measurement format provide high sample throughput and rapid generation of microbiologically relevant and quantitative data.

Table 3. Results of bacterial growth in different selective media. The numbers represent doubling time values (min), calculated from the linearized calibration curves
Medium Bacillus cereus Escherichia coli Listeria innocua Salmonella typhimurium Staphylococcus aureus Pseudomonas aeruginosa Yersinia enterocolitica Campylobacter jejuni Streptococcus pyogenes
  1. ∞, No growth during 24 h of monitoring.


Nearly all the bacterial species showed O2 depletion in the media considered as growth promoting (LB, NB and BPW), except for Camp. jejuni and Strep. pyogenes, which did not show O2 depletion in NB medium. The other media showed partial selectivity with respect to different bacteria. Streptococcus pyogenes, Camp. jejuni and Bacillus subtilis were most susceptible to the components of the media, showing growth in two, five and seven types of media, respectively. Analysis of the data in Table 3 shows that SC broth was the most selective, only allowing the growth of Salm. typhimurium. SC is enrichment medium for isolation of Salmonellae from faeces and food products. Cetrimide broth, which is used for the selective isolation and identification of Ps. aeruginosa, also showed good selectivity. It also allowed growth of salmonella, but at significantly slower rates with a DT of >2 h. The least selective media were CIN and MAC, which allowed growth of almost all micro-organisms, and in several cases, DT did not show significant difference. Pseudomonas aeruginosa was the most resistant bacterium growing in all tested media, except for SC.

When analysing the DTs over the whole panel of broths, each bacterium showed a unique pattern, which can be used for identification. Inverted DT can be used to create such patterns (Fig. 2). These growth patterns were successfully used for the identification of bacteria in blind tests with pure cultures. Similar patterns can be generated for each medium. One useful application of this method is the monitoring of culture purity.

Figure 2.

Patterns of growth of different bacteria (1/DT(h)) on the media tested.

To further assess the selectivity of this method, respiration of mixtures of bacteria was analysed. Two bacterial species were mixed at different concentrations in a medium, and respiration profiles of these samples were measured and analysed. The medium was usually selected such that it promotes growth of either only one micro-organism or both.

The results obtained with E. coli and L. innocua mixtures in EC broth are shown in Fig. 3a. EC is a selective medium for differentiation of faecal coliforms and confirmatory test for E. coli from food and environmental samples. The presence of L. innocua cells in E. coli culture at concentrations up to 10CFU ml−1 did not affect TT values. Escherichia coli could be detected selectively and quantified at different concentrations down to 10 CFU ml−1, while growth of L. innocua was completely inhibited. Figure 3b relates to a mixed culture of E. coli and L. innocua in MAC broth, which is commonly used for primary isolation of coliform bacteria and which promotes growth of both bacterial species. The presence of L. innocua is seen to alter the TT values for each dilution of E. coli, making it shorter at high concentrations of the former. As the DT for Listeria is 2·5 times longer than for E. coli, cross-sensitivity was only seen at very high concentrations of Listeria (>10CFU ml−1), when its contribution to O2 depletion is significant, while at lower concentrations, it did not interfere. Figure 3c demonstrates the case of two bacterial species with similar growth rates: B. subtilis and Salm. typhimurium in CIN medium. The contribution of both bacteria to consumption of oxygen was equal, and cross-sensitivity is significant. These results show that identification of the bacteria in complex samples depends on the selective properties of the media used and that the TT readout can be influenced by the presence of other bacteria that show growth in the same medium as target bacterium.

Figure 3.

Respiration analysis of bimixtures of bacteria. (a) Effect of Listeria innocua on the threshold times (TT) of Escherichia coli in EC broth. (b) Effect of L. innocua on the TT of E. coli in MacConkey broth. (c) Effect of Salmonella typhimurium on the TT of Bacillus subtilis in CIN broth. (♦) 107; (○) 106; (▲) 105; (□) 104; (■) 103; (●) 102 and (Δ) 10CFU ml−1

Finally, the respirometric data generated in this study also demonstrate that successful identification of different bacterial species can be achieved using rather small panels of media and small number of measurements. Thus, Fig. 4 shows that six of nine bacteria from the panel can be identified positively based on their growth/respiration patterns on just four selective media. The remaining three (B. cereus, Camp. jejuni and Strep. pyogenes) do not grow on these media, so they do not interfere. Measurement of 2–3 serial dilutions of test samples (e.g. 1 : 10, 1 : 30 and 1 : 90) in each medium is deemed sufficient for reliable and accurate quantification of corresponding DT values. Significant differences in 1/DT values for different bacteria (e.g. >2 vs <1) is a measure of selectivity for a given media, and the values themselves can be used for putative identification of the species.

Figure 4.

Growth patterns of different bacteria on four selective media. (image_n/jam12049-gra-0007.png) Bacillus; (image_n/jam12049-gra-0008.png) Escherichia coli; (image_n/jam12049-gra-0009.png) Listeria; (image_n/jam12049-gra-0010.png) Salmonella; (image_n/jam12049-gra-0011.png) Staphylococcus aureus; (image_n/jam12049-gra-0012.png) Pseudomonas aeruginosa; (image_n/jam12049-gra-0013.png) Yersinia entero; (image_n/jam12049-gra-0014.png) Campylobacter jejuni and (image_n/jam12049-gra-0015.png) Streptococcus pyogenes.


The microrespirometry platform, which uses GreenLight probe, standard 384-well plates and detection on commercial multilabel reader with TR-F capabilities, provides facile, contactless, real-time monitoring of growth of aerobic micro-organisms with quantitative readout. It allows parallel measurement of large number of samples containing different bacterial strains, conditions (media, additives) and samples.

In this study, we have demonstrated that this method allows differentiation and putative identification of particular bacterial species, based on their characteristic doubling/generation times and patterns of respiration produced on a panel of partially selective media. This method was demonstrated with pure bacterial cultures and then with binary mixtures of bacteria, where it has been shown to provide robustness and high level of selectivity. It is advantageous over the conventional microbiological methods of selective determination of bacteria, which usually rely on solid media (just one or a few) and end-point qualitative or semiquantitative readout.

Respirometric assays are based on the principle of dynamic (reversible) quenching of phosphorescence of a probe by O2, such that depletion of O2 in the sample causes an increase in phosphorescent signal intensity and lifetime. The GreenLight probe was shown to work reliably in various media, even in very complex ones. Selective media used in this study vary significantly in their composition (salts, pigments, antibiotics, other additives) and optical properties (some are deeply coloured – see Supporting Information, Fig. S1). This can potentially affect probe signal and response to O2 (cross-sensitivity, inner filter effect, etc.). Even though time-resolved fluorescence detection provides high signal-to-blank ratio and stability to optical interferences (O'Mahony and Papkovsky 2006), the GreenLight probe represents an improved, shielded modification of the first-generation probe MitoXpress, providing high signal-to-blank ratios, stable lifetime readings, reliable detection of growth of different aerobic bacteria via their respiration and determination of TT. The probe accurate enumeration of bacteria (CFU ml−1) based on predetermined calibrations and determination of their growth parameters (DT).

For all the media and bacteria used, the respirometric assays were capable of providing sensitivity from a single-cell level up to about 107 cells ml−1. Even the slow-growing bacteria present in the original sample at very low concentrations (10 CFU ml−1) were detectable in less than 24 h. Thus, single-cell detection of E. coli in LB and PBW required c. 7 h and c. 22 h using TTH medium, whereas high and medium levels of contamination can be detected and flagged in just a few hours of the measurement. Like other microbiological methods, this assay only detects viable cells present in the original sample, and it is not prone to false-positives which can be effectively filtered out.

One should also keep in mind that the respirometric approach has a number of inherent limitations, many of which emanate from its microbiological nature. For example, respirometric profiling of very complex samples containing multiple bacterial species can be problematic due to their high cross-sensitivity on different media. The presence of metabolic toxins or drugs (e.g. antibiotics, substrates) in the original sample or metabolic stress imposed on bacteria may also have an effect on both respiration profiles and DT. Unlike the molecular techniques (PCR and antibody-based), this test cannot provide absolute selectivity and unambiguous identification of particular bacterial strains and species. The latter task will require additional confirmatory tests; however, these tests can be conducted on positive samples identified in the respirometric assay, which are already enriched (>107 CFU g−1) and can be processed straightaway. These issues have to be considered carefully before applying this technique to each particular analytical task or sample type.

On the other hand, the respirometric methodology can also benefit from the ongoing development of more efficient and selective media and microbiological procedures for particular bacterial species, especially for emerging food pathogens. Such new media and method modifications are very easy to integrate in the assay. Selectivity of the respirometric assay can be further enhanced by combining it with the monitoring of colour changes in chromogenic growth media. As GreenLight probe produces no colour, colour changes associated with microbial growth can be used as an additional parameter in profiling and identification of bacteria present in samples being tested. Analysis of full respiration profiles in addition to DT determination can also be applied to improve method selectivity.


Financial support to this work by the European Commission National Development Plan and Enterprise Ireland (CFTD 05/112, CFTD 07/124, PC 08/0184) is gratefully acknowledged.