Issues around field measurements with electronic noses : low concentration and specificity of odorous compounds Jacques NICOLAS Julien DELVA - Anne-Claude ROMAIN Department of Environmental Sciences and Management of University of Liège (previously FUL) Arlon (Belgium)
Main research topic : assessing the odour annoyance generated by different industrial or agricultural sources in the environment. Landfill sites Compost facilities Waste water treatment plants Settling ponds of sugar factories Piggeries, hen-houses Detection of moulds in buildings
Different tools : Lab analysis (GC-MS) Dynamic olfactometry Field inspection + atmospheric dispersion modelling
and electronic nose Very promising possibilities + Stress Monitoring odour emission and trying to predict and measure odour annoyance in the surroundings Controlling odour abatement system Using the odour as a process variable to detect failures
Preconcentration is necessary to improve the detection performance of e-nose and to broaden the range of possible applications of e-nose in the environment Only final users Our role in a possible project = expliciting the specifications of a "field preconcentrator" specifying the range of odour and VOC concentration observed in the field keeping in mind the relation between chemical composition and odour identifying key compounds particularly involved in odorous mixtures selecting material and condition of operation to avoid the risk of denaturing the sample choosing adequate sample time preconcentration to insure fast response for a dynamic odour detection testing prototypes in the field
1. Concentration comes down below the limit of detection of gas sensors At the emission level, or in the immediate surrounding, an electronic nose is able to detect and to recognise an odour, but not in the environment. Typical chemical composition? Composting facility esters 23% alcohols 27% & aldehydes ketones 17% aromatics 1% terpens 28% aliphatic hydrocarbons 3% sulfur compounds 1% Waste water treatment Nitrogen compounds Sulfur compounds Organic acids Aldehydes Ketones Ammonia Dimethylamine Methylamine Ethylamine Skatol Indole Cadaverine Dimethylsulfide Methanthiol Ethanethiol Hydrogen sulfide Acetic acid Butyric acid Valeric acid Methanal Ethanal Buthanal Acetone
Typical chemical concentrations at the emission? Composting facility Compound examples 3-methyl-butanal Butanoic acid, ethyl ester 2-butanol Phenol Ethyl-acetate 1-propanol 2-butanone Limonene Ethanol ppm(v) 0.022 0.019 0.038 0.044 0.065 0.114 0.116 3.340 1.155 Waste water treatment (sludge stabilization) Compound examples Ammonia Toluene Dimethyl sulfide ppm(v) 25 0.290 0.360 Rarely above 1 ppm(v) Acetone Slaughterhouse) Compound examples Dimethyl disulfide Dimethyl trisulfide Tetramethyl pyrazine But e-nose reacts on global volatile emission ppm(v) 0.007 0.005 0.001 0.001 Sum? not representative of the global e-nose response not representative of the odour concentration
To have an idea of the order of the "global chemical concentration" detected by the sensor array Finding an equivalence between the global "chemical concentration" of the odorous gas mixture and the concentration of ethanol vapour Common features = sensor signals Sensor resistance Ethanol Odorous gas Ethanol concentration Odorous gas dilution
E T H A N O L Ethanol Odorous gas (compost) 1 25 ppmv ethanol-equivalent
Limit of detection in "ethanol equivalent"? Limit of detection signal-to-noise ratio S/N = 3 Noise = standard deviation σ of the stabilised signal (e.g. in kω) For our configuration σ [ 0.07,1.8kΩ] according to the sensor type Corresponds to 0.04 1.03 ppm(v) in ethanol equivalent Or to a dilution of 40 100 for a typical compost sample (R0-R) sensor TGS2620 (kω) 50 45 40 35 5 30 4 25 3σ = 1.35 kω 3 20 2 15 Ethanol concentration = 0.20 ppmv 1 10 0 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 5 Ethanol concentration (ppmv) 0 0.00 5.00 10.00 15.00 20.00 25.00 30.00 (R0-R) sensor TGS2620 (kω) Ethanol concentration (ppmv)
and the limit of resolution may still be higher (e.g. recognition of compost odour among others) 2.0 1.5 1.0 Decreasing concentration 0.5 Factor 2 0.0-0.5-1.0-1.5 Decreasing concentration -2.0-1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 Factor 1 compost neutralizing product
Far away from the source the "global" concentration comes down below the limit of detection of gas sensors 10 ppm 1 ppm 100 ppb 10 ppb Improving the sample uptake is essential
Correspondance with "odour concentration" Measurement method : dynamic olfactometry Assessment of odour concentration expressed in ou/m 3 1 ou/m 3 = perception threshold
"Calibration curve" between odour concentration and sensor signals 3σ = 1.35 kω (R0-R) sensor TGS2620 (kω) 60 50 40 30 20 10 0 0 200 400 600 800 1000 1200 1400 1600 15 uo/m 3 ou/m3 For compost emission, the odour concentration corresponding to the limit of detection of the sensors is low and close to the odour perception threshold of human nose, defined as 1 uo/m3. For the different sensors : from 10 to 80 ou/m3.
1. Concentration comes down below the limit of detection of gas sensors 2. Risk of denaturing the sample Very important for odour measurement (olfactometry) to avoid interaction between sampling material and sampled air The bag must be odour free without reaction or adsorption with the sample impervious, to prevent significant losses before measurement Pumping system cannot interact with the sample Tedlar bags (PVF) Sealed-barrel maintained under negative pressure Air drawn into the bag by the pressure difference
ideally, same precautions for a preconcentrator at the e-nose inlet Purpose of e-nose : recognising an odour (e.g. compost emission) and monitoring it e.g. : detecting that the odour level is over a "warning threshold" 5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 exhaust gaz of a truck 700 uoe/m³ : odour concentration warning threshold 9:40 9:46 9:51 9:56 10:01 10:07 10:12 10:17 10:23 10:28 10:33 10:39 10:44 10:49 10:54 11:00 11:05 11:10 11:16 11:21 11:26 11:32 11:37 Time
But : gas sensors respond to both odorous and odourless compounds The global e-nose signal "odour" if "chemical" concentration is correlated to odour concentration 1801 1601 1401 1201 1001 801 601 401 201 1 TGS 822 Concentration = 1+9352 *(G-G 0 ) with G 0 = 0.0219 r² = 0.88 0 0.05 0.1 0.15 0.2 0.25 Conductance G(kS) OK for compost emission Odour concentration 2000 1800 1600 1400 1200 1000 800 600 400 200 0 TGS 800 TGS 822 0 0.05 0.1 0.15 0.2 0.25 Conductance (G, ks) NOT for printing house emission
But : gas sensors respond to both odorous and odourless compounds The global e-nose signal "odour" if "chemical" concentration is correlated to odour concentration 1801 1601 1401 1201 1001 801 601 401 201 1 TGS 822 Concentration = 1+9352 *(G-G 0 ) with G 0 = 0.0219 r² = 0.88 0 0.05 0.1 0.15 0.2 0.25 Conductance G(kS) Odour concentration Keeping that relationship after preconcentration OK for compost emission
Sometimes : no specific compound but characteristic "signature" Example (indoor air pollution) : detection of mould contamination of wood material Typical fungal signature ("musty smell") does not necessary involve new specific compounds with respect to original material α-pinene β-pinene α-pinene β-pinene
In indoor air, simple sampling is not sensitive enough for the VOC detection Requires pre-concentration Best = active sampling process But suitable analyte recovery is essential In some cases, selective pre-concentration could also be interesting If the recovered chemicals are precisely those which are typical of the odour or pollution
1. Concentration comes down below the limit of detection of gas sensors 2. Risk of denaturing the sample 3. Preserving the dynamics of the detection Continuous monitoring 0 c o m p o s t p i l e m o v i n g c o m p o s t a e r a t i o n 1 0 Mahalanobis distance (D²) 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 w o r k i n g d a y s o f f w e e k - e n d 1 0 0 1 8 : 5 7 6 : 5 7 1 8 : 5 7 6 : 5 7 1 8 : 5 7 6 : 5 7 1 8 : 5 7 6 : 5 7 1 8 : 5 7 6 : 5 7 1 8 : 5 7 6 : 5 7
Ideally Pre-concentrator should work continuously with improved cycling time to allow a greater number of samples to be analyzed in a given time period. 9:40 9:45 9:50 9:55 9:59 10:04 10:09 10:14 10:19 10:23 10:28 10:33 10:38 10:43 10:47 10:52 10:57 11:02 11:07 11:11 11:16 11:21 11:26 11:31 11:35 Time
Time 9:40 9:45 9:50 9:55 9:59 10:04 10:09 10:14 10:19 10:23 10:28 10:33 10:38 10:43 10:47 10:52 10:57 11:02 11:07 11:11 11:16 11:21 11:26 11:31 11:35 OK
Time 9:40 9:45 9:50 9:55 9:59 10:04 10:09 10:14 10:19 10:23 10:28 10:33 10:38 10:43 10:47 10:52 10:57 11:02 11:07 11:11 11:16 11:21 11:26 11:31 11:35 Too much "filtering"
Conclusions Pre-concentration is needed for many applications of ambient gas monitoring in the environment (odour, indoor air pollution, ) with electronic nose gas concentration is often below the limit of detection of sensors Final users may contribute to the development writing the "specification sheet" typical gas composition and concentration of ambiences relation chemistry/odour dynamics of gas monitoring in the field testing prototypes in the lab in the field