Particle characterisation Methods applied for the characterisation of fine particulate emissions

Transcription

1 g a article characterisation ethods applied for the characterisation of fine particulate emissions homas runner nstitute for rocess ngineering Graz niversity of echnology.: +43 (316) ; FX: +43 (316) : thomas.brunner@tugraz.at HG: R G a g H R G H nstitute for rocess ngineering Graz niversity of echnology ontents ntroduction Relevance of particle characterisation Relevant characteristics of fine particulate emissions from biomass combustion article characterisation technologies applied onclusions and outlook 2

2 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology ntroduction () Relevance of the characterisation of fine particulate emissions from small-scale biomass combustion. Gain more detailed information about particle formation during small-scale biomass combustion and the mechanisms involved. rovide information to interlink the properties of emissions from small-scale biomass combustion plants with characteristics of ambient air. Future aim: provide information about characteristics of emissions from biomass combustion for a classification of their toxicity. 3 R G H nstitute for rocess ngineering Graz niversity of echnology ntroduction () Relevant characteristics of fine particulate emissions from biomass combustion article size distribution - mass-size distribution - number-size distribution hape and structure of the particles hemical composition - inorganic compounds - organic compounds and soot - Hs 4

3 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology article size distribution general issues Results from various research projects show that in small-scale biomass combustion plants almost no coarse fly ash particles are emitted, more than 90% of the total emissions are related to 10, there is an almost neglectable amount of particles in the size range between 1 and 10 µm and therefore, 1 almost equals 10. onsequently, to determine the particle size distribution of emissions from small-scale biomass combustion plants, mainly the size range <1 µm is of interest. 5 R G H nstitute for rocess ngineering Graz niversity of echnology Discontinuous determination of the particle size distribution impactors () scheme of a single separation stage of a erner-type low pressure impactor particle separation in an impactor gasket nozzle plate spacer ring sampling foil stagnation plate sampling foil with particles case 6

4 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology Discontinuous determination of the particle size distribution impactors () device 10 impactors ow pressure impactors remarks - usual cut diameters: 10 µm, 2.5 µm, 1 µm, final filter stage - aerodynamic diameters - cut sizes from some 0.1 µm to some µm - aerodynamic diameter advantages - gravimetric method - easy to apply - subsequent chemical analyses of particles sampled possible - no dilution needed - gravimetric method - various substrates can be applied - subsequent chemical analyses possible - no dilution needed disadvantages - low resolution regarding particle sizes - no information about D <1 µm - discontinuous measurement - discontinuous measurement - time consuming - short sampling time if no dilution is applied xplanations: diameters in µm aerodynamic diameter, D particle size distribution 7 R G H nstitute for rocess ngineering Graz niversity of echnology lectrical ow ressure mpactor n-line determination of the particle size distribution () canning obility article izer heath air in harged polydisperse aerosol Flow straightener Highvoltage rod source: xcess air out source: onodisperse aerosol out 8

5 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology n-line determination of the particle size distribution () device (electrical lowpressure impactor) (scanning mobility particle sizer) F (fast mobility particle sizer) remarks µm - aerodynamic diameter µm µm - electrical mobility diameter µm - electrical mobility diameter advantages - robust - indicates changes in process well - large size range - very small particles can be detected - high resolution -fast - indicates changes in process well disadvantages - overestimation of agglomerates and coarse particles - dilution needed - long scanning time (problem in varying combustion process) - dilution needed - more inaccurate than - high dilution ratios needed 9 R G H nstitute for rocess ngineering Graz niversity of echnology article size distribution aerodynamic and electric mobility diameter erodynamic diameter (ae.d.): Generally, are not of uniform shape and density. he aerodynamic diameter is used to make particles comparable. t is the diameter of a spherical particle with a density of 1 g/cm³ that has the same inertial properties in air as the particle of interest. lectric mobility diameter and F systems are based on Ds (differential mobility analysers), where charged particles are separated by an electrostatic classifier. herefore, they are classified according to their electric mobility diameter. 10

6 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology article size distribution general remark concerning measurement results From low-pressure impactors information about the mass-size distribution of particles based on aerodynamic diameters is gained. he provides information about the number-size distribution of particulate matter based on aerodynamic diameters. he determination of the mass-size distribution from the number size distribution is problematic since several particle properties have to be known (shape, density, etc.). and F results are given as number-size distributions and are based on electric mobility diameters. hese differences have to be considered when comparing measurement results gained from these devices. 11 R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation arameters of interest ajor ash and aerosol forming elements: e.g.:, a,, l, a, i, g, n Heavy metals: e.g.: Zn, u, d, b arbonaceous compounds organic carbon elemental carbon (soot) inorganic carbon (carbonates) H 16 Hs according to 12

7 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation major and minor ash forming elements Wet chemical analyses different sampling substrates applicable (metal foils, quartz filters) low detection limits can be achieved only bulk-analyses possible lectron microscopy different sampling substrates applicable structural analyses of single particles are possible if appropriate sampling methods are applied (single particles and no bulk samples are needed) compared to wet chemical analyses less sensitive 13 R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation wet chemical analyses () Digestion l: eluation (24h) with bi-distilled water all others: multi-step pressurised digestion with H 3 /HF/H 3 4 Detection l: ion chromatography () all others: adsorption spectrometry () or plasma emission spectrometry (-, -) 14

8 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation wet chemical analyses () Restrictions concerning wet chemical analyses of impactor samples ample mass needed: >50 µg With increasing sample mass the accuracy of the analyses increases. However, usually only 2 to 3 impactor stages with sufficient particle mass can be provided. herefore, a micro balance with high resolution must be used in order to avoid weighing errors which later can cause incorrect analyses results. o increase the sample mass and therefore also the accuracy, a mixed sample of more than one impactor stage can be made. However, thereby information about the differences of the chemical compositions of different particle size classes is lost. specially during field tests assembling and disassembling of the sampling device usually take place under conditions different to a clean room. t has to be taken care that the samples are not contaminated during sample handling, storage and transport. 15 R G H nstitute for rocess ngineering Graz niversity of echnology echnologies applied canning electron microscopy () ransmission electron microscopy () nergy dispersive x-ray spectrometry (DX) hemical characterisation electron microscopy Restrictions for f the surface of the sample is not perfectly plane, the electron beam is scattered and signals from the sample surface next to the particle are detected together with signals from the particle. he penetration depth of the electron beam can be greater than the particle size also signals from the substrate are detected. he sample usually has to be coated to make it electroconductive. he element chosen for coating (in many times but also u and others are applied) can thus not be detected correctly. Detection limit: usually ~1 wt% calibration with standard materials is recommended 16

9 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology ampling with a polycarbonate filter filter holder duct hemical characterisation particle sampling for electron microscopy (examples) pump gas flow counter leakage air casing grid filter sealing casing tubing ssembled filter holder with tubing and nozzle nozzle samples taken downstream the boiler during a combustion test with bark picture widths: 5.6 µm 17 R G H nstitute for rocess ngineering Graz niversity of echnology ndividual agglomerates can be collected on grids and then be analysed hemical characterisation particle sampling for electron microscopy (examples) ource: niversity of uopio Fine article and erosol echnology aboratory Dept. of nvironmental cience 18

10 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation electron microscopy - examples 50nm /DX micrograph of an aerosol particle sampled during beech combustion (good burnout conditions) sampling with a u-grid dark zone: mainly and with traces of l bright zone: mainly and l with traces of image of a soot agglomerate sampled during poor burnout conditions (source: niversity of uopio Fine article and erosol echnology aboratory, Dept. of nvironmental cience 19 R G H nstitute for rocess ngineering Graz niversity of echnology /DX analyses hemical characterisation comparison of results from wet chemical and /DX analyses wet chemical analyses wt.% wt.% wt.% l wt.% wt.% wt.% wt.% l wt.% dp µm ae.d dp µm ae.d. xplanation: sample from wood chip combustion; sampling with a low-pressure impactor on l sampling foils; the sampling foils were cut into two pieces of which one was forwarded to /DX and the other one to wet chemical analyses; ae.d. aerodynamic diameter 20

11 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation carbonaceous compounds () General remark When sampling particles on a filter, besides the particles also absorbed gaseous compounds are collected. his has to be considered for substrate treatment, sampling and analyses. ompounds of interest organic carbon elemental carbon inorganic carbon (carbonate) Detection of inorganic carbon and sum of and with a -analyser he sample is heated under oxidising conditions up to 1,000. u-catalyst transforms -species to 2 which is detected by D-R. 2 detected below 550 is allocated to reactions of elemental and organic carbon non carbonate carbon ( + ). ignals detected at temperatures >550 are allocated to carbonates. 21 R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation carbonaceous compounds () Detection of and (thermo-optical methods) he sample is heated according to a defined temperature program. arbonaceous species are released from the sample and are converted in catalysts to 2. he 2 is either - detected by a D-R analyser or - reduced to H 4 which is detected with a FD. he filter transmittance is permanently monitored with a laser to correct for charring effects. Due to the fact that a laser is used to correct for charring effects, the sampling media applied must allow transmission (no analyses of metal impactor foils possible). 22

12 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation carbonaceous compounds () ethod according to H 5040 is determined under He-atmosphere at 300, 470, 610 and 865 is determined under an atmosphere containing 2% 2 and 98% He at 550, 620, 700, 780, 850 and 865 Detection by using a FD ethod according to uxbaum et al. ample is heated continuously (20 /min) to 800 in 2 atmosphere. 2 detection by a DR analyser. he first peak in 2 that evolves after the transmission has reached again the initial value is assigned to (the corresponding temperature is around ). ll previous peaks are assigned to. causes another distinct peak which evolves after the peak at about R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation H concentrations () General approach ampling of gaseous and particulate phase Hs on adsorbents/filters. Detection of Hs by G/ from the extracts of adsorbents and filters. Denuder ethod pplied when sampling from diluted exhaust gas. Denuders, filter and polyurethane foam (F) are extracted separately. gas phase particle phase denuder 1 denuder 2 filter F sample from diluted flue gas to pump and gas meter 24

13 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology hemical characterisation H concentrations () Filter-cooler method pplied when sampling from undiluted (hot) exhaust gas. article phase and gas phase are sampled separately. t has to be considered, that during the cooling of the flue gas between the stove/boiler and the chimney outlet Hs condense and therefore the distribution between gas and particle phase is changed. particle phase gas phase sampling filter from duct adsorbent (polyurethane foam, XD resin) cooling (water circulation) to pump and gas meter condensate bottle 25 R G H nstitute for rocess ngineering Graz niversity of echnology onclusions and outlook () For the characterisation of fine particulate emissions from biomass combustion a broad range of different technologies is applicable. article size distribution he information gained from the different instruments applied significantly differs in: - resolution concerning particle sizes - time resolution - mass or number-size distribution - aerodynamic and electric mobility diameter hese differences have to be considered regarding the selection of a certain measurement method as well as concerning the evaluation of the results. 26

14 g g a a R G H nstitute for rocess ngineering Graz niversity of echnology onclusions and outlook () For structural and chemical characterisation of particulate emissions proven methods are available. he following issues have to be considered: he sampling substrate and its pre-treatment has to be tailored to the demands of the analyses method applied. pecial care has to be taken during substrate preparation, sampling, sample storage and sample transport in order to avoid contaminations. resently, technologies for on-line analyses of are gaining rising interest (e.g.: aerosol mass spectrometers). However, these devices are, compared with off-line methods expensive and only limited experiences (especially concerning emission measurements) are available so far. 27 R G H nstitute for rocess ngineering Graz niversity of echnology cknowledgement R- ioenergy project iomass- lean biomass combustion in residential heating particulate measurements, sampling, and physicochemical and toxicological characterisation eine nitiative des undesministeriums für Verkehr, nnovation und echnologie (V) 28

15 g a R G H nstitute for rocess ngineering Graz niversity of echnology hank you for your attention Dipl.-ng. Dr. homas runner nffeldgasse 21b, Graz, ustria.: +43 (316) ; FX: +43 (316) mail: thomas.brunner@tugraz.at HG: 29