Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Alejandro Keller, Heinz Burtscher University of Applied Sciences and Arts Northwestern Switzerland
FATCAT: a new characterization method for particulate emissions from wood burning appliances
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 2
Toxicity of Particle Emissions (not are particles are created equal) In Vitro exposure of lungcells (chin. hamster)
Source: N. Klippel & T. Nussbaumer, 9. Holzenergie-Symposium 2006
0 1 2 3 4 5 6 7 8 9 100
10
20
30
40
50
60
70
80
90
100
Diesel Soot Automatic Installations Good wood combustion (SCI) Bad wood combustion
Cel
l sur
viva
l (%
)
Dose in cell nutrient solution (mg/ml)
Soot and organic
molecules
Additionally: Water soluble substances are cleared faster from the lungs than (the non-soluble) carbonaceous fraction*.
*More information regarding health impact: Sigsgaard, 2015. doi:10.1183/13993003.01865-2014
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 3
Emissions Composition (Small Combustion Installations)
Pelletboiler
Modern logwood stove
Masonry heater
Masonry heater
Masonry heater
Sauna stove
0 100 200 300 400 500Particulate Matter Emissions (mg/m3
STP 13% O2)
CO3
Organic Carbon Elemental Carbon Non Carbonaceous
Source: Lamberg et al. / Atm. Env. 45 (2011) 7635-7643
Comply with the European ecodesign requirements (uncertainty subtracted as required by regulation)
High C-content, High toxicity
Low C-content, Low toxicity
Current standards based on total PM are far from ideal: 1. large particles (dp>1µm) are the most
relevant for this metric although they are not relevant for health or climate effects.
2. Toxicity is very relevant. A standard should give more importance to carbonaceous material than the non-carbonaceous fraction.
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 4
What does aging do?
• Aging produces secondary organic aerosol (SOA).
• SOA amounts ~2/3 of the organic aerosol mass in the atmosphere**.
• SOA can duplicate organic carbon emissions from wood burning*.
• Aging reduces volatility and makes sampling robust against differences in temperature or dilution level.
• Gravimetric sampling at 160°C (current standard) misses an important fraction of the organic carbon (including some primary organic carbon).
OC1 OC2 OC3 OC40
20
40
60
80
100
Low
volatility
Con
tribu
tion
to to
tal o
rgan
ic c
arbo
n (O
C, %
)
Highvolatility
aging
Logwood Stove example (9 kW, warm start cycle)*
Before aging After aging
* Keller and Burtscher, J. Aerosol Sci., under review. ** e.g., Lanz et al. ACP 7, p. 1503, 2007 and references therein
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 5
The Metric: Why total carbon (TC)? 1. TC is directly related to the combustion
quality. This is not the case for PME (TC fraction of PME is between 5% and 90% for “certifiable” appliances)
2. Total carbon is a better metric for health and climate impact than PME
3. Single value and less complicated than, e.g., OC/EC
4. Method already available as commercial devices and is also offered by independent laboratories
5. Potential for automatic semi-online analysis system
Total Carbon (TC) is the mass of all carbon atoms in a sample. This includes the mass of soot (elemental carbon, EC) as well as the mass of carbon atoms in organic molecules (organic carbon, OC). Usually measured by means of CO2 determination. Carbonates are also detected by this technique. (graphic: TU-Wien)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 6
• On a first step, [particle bound] organic carbon (OC) and elemental carbon (EC) measurements by means of thermal-optical analysis (EUSAAR 2 protocol). Total Carbon, TC = OC + EC
• Analysis may be substituted by a relatively simpler thermal analysis for TC.
Heated line (200°C)
UV Reactor
Dilution Air (up to 1:14) at 200°C
Filters(1nlpm)
to FID
to MRU
Activated CDenuder
Co2sensor
Dilution air
Rotating-DiscDiluter
to miniDiSC & CO2 Sensor(final dilution ~1:400)
Heated block
Setup for filter analysis experiments
Semi-online Analysis System:
FATCAT
simulates atmospheric aging
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 7
FAst Thermal CArbon Totalizator (FATCAT)
• Sampling: 30 minutes @ 1lpm (dilution ~1:5).
• Target mass: ~4-40µg.
• Analysis time: ~3.5 min (50s heating).
• Cool down time (to 30°C): 25 minutes (including analysis time). Reusable sampling-filter
& flash heating (>800°C)
Pt-Catalyst (Diesel)
CO sensor2
flow:1nlpm
SampleSynth.
air
1000.75
0.80
0.85
0.90
0.95
1.00
Filtr
atio
n E
ffici
ency
Diameter Midpoint (nm)
105
106
107 Before filter After filter
dN/d
logD
p (#
/cc)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 8
Sampling filter
• Flash heating is possible due to direct heating of the filter (as opposed to using a conventional furnace).
• Design restrictions require a custom-made solution.
• Filtration efficiency: >95% for particle diameter dp>100nm, but as low as 85% for certain sizes.
• Filter is reusable.
Filter with better efficiency (+ improved
hardware) will be available in few weeks
0 10 20 30 40 50 60 70 80 900
1
2
3
4
5
Tota
l Car
bon
(µg/
seco
nd)
Analisys Time (Seconds)
CAST, "Blacksoot" Base line 5µg 11µg 20µg 43µg 76µg
0
200
400
600
800
1000 T
empe
ratu
re (°
C)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 9
Soot sample measurements
Heating Stopped at t=50s
CAST: Combustion Aerosol Standard, Jing AG, is a propane-flame particle generator
0 20 40 60 80 100
0
20
40
60
80
100
Background Signal
-0.4
-0.2
0.0
0.2
0.4
C/O=0.26 (EC>>OC) C/O=0.41 (EC>OC) C/O=0.49 (EC<<OC)
Tota
l Car
bon
(µg-
C)
Collected Aerosol Mass, TEOM (µg)
a) b)
TC R
ange
(µg-
C)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 10
CAST soot samples
CAST: Combustion Aerosol Standard, Jing AG, is a propane-flame particle generator TEOM: tapered element oscillating microbalance, Thermo Scientific
Slopes: C/O=0.26 -> m=0.88 (Adj. R2=0.999) C/O=0.41 -> m=0.82 (Adj. R2=0.999) C/O=0.49 -> m=0.82 (Adj. R2=0.998)
0 100 200 300 400 500 6000
100
200
300
400
500
600Cooking Stove ( cold, warm)Old Stove ( cold, warm)
TC, F
ATC
AT
(µg)
TC, EUSAAR2 (µg)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 11
Wood burning example (test bench measurements)
Slopes: m=0.88 (Adj. R2=0.995)
0 15 30 45 60 75 90
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09 CAST Black (C/O=0.26) CAST Brown (C/O=0.41) CAST Brown (C/O=0.49)
Tota
l Car
bon
(1/s
econ
d)
Analysis time (seconds)
200
400
600
800
1000
Tem
pera
ture
(°C
)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 12
Thermograms?
• Can we extract some information about the OC/EC split or the volatility of the sample from the individual peaks?
• Pyrolysis: promoted or reduced by flash heating?
CAST (C/O=0.26) CAST (C/O=0.41) CAST (C/O=0.49)-20
-10
0
10
20
30
40
50
Sam
pled
-C (b
lank
sub
stra
cted
) [µg
] OC EC
EUSAAR2 Thermal-Optical Analysis
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 13
To do….
• Improve filter efficiency (on its way)
• Transmission efficiency for complete setup (e.g., due to the use of a denuder)
• Further testing using wood burning appliances like, e.g., clean automatic boilers.
• Questions about real time data (thermograms).
• We have plans to adapt our device for ambient monitoring of total carbon as a stand-alone, continuous-operation instrument.
We are looking for partners for characterization, further development and/or commercialization. Both for emission measurements as well as for ambient monitoring.
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 14
Summary
• Total carbon (TC) analysis (combined with simulated atmospheric aging) should be considered as a candidate metric for biomass burning emissions. This is only possible if simple reliable instruments are available.
• Our solution, FATCAT, shows very good linearity during tests with standard combustion aerosol, as well as for tests using biomass burning appliances.
• The limit of detection is 3σ=0.4µg (noise-to-background). This is 10 to 100 times lower to the expected TC filter-load from a type approval tests performed on clean, certificated, biomass burning appliances.
• Filtration efficiency is as low as 85% for some sizes. A new improved prototype will be available in few weeks time.
• Thermograms show well separated peaks (fingerprints for different sources?). There may be more information in this real time data.
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 15
Thanks to:
Markus Zürcher Hug Engineering
Andreas Mayer TTM
Daniel Egli and Peter Steigmeier University of Applied Sciences Northwestern Switzerland
Thank you for your attention!
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 16
Backup Slides
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 17
Field measurement example (wood charcoal grill)
Image: lotusgrill.de
Total Carbon
Total PM (TEOM)
mg/m3stp @13% O2
Start (first 10 minutes) 16.4 (0.5)
Full power (last 20 min.) 8.5 (0.4)
Full power (full cycle) 3.3 (0.1) 21.0
Half power (first 40 min.) 6.1 (0.2)
0 15 30 45 60 75 90
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
Tota
l Car
bon
(µg/
seco
nd)
time (seconds)
Heating Stop
0 15 30 45 60 75 90
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09 CAST Black (C/O=0.26) CAST Brown (C/O=0.41) CAST Brown (C/O=0.49)
Tota
l Car
bon
(1/s
econ
d)
Analysis time (seconds)
200
400
600
800
1000
Tem
pera
ture
(°C
)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 18
Thermograms? (2)
0 15 30 45 60 75 90
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09 tetracontane "smoke-free" grill
Tota
l Car
bon
(1/s
econ
d)
Analysis time (seconds)
CAST EC CAST OC
• It may prove difficult to set a proper split point between OC and EC
• Accumulation of non-carbonaceous material may have a catalytic effect on EC (i.e. filter history may cause EC combustion at lower temperature)
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 19
EUSAAR2 Measurements*
1 2 3 4 5 6 7100
101
102
103
EC OC (aging)
Logwood Stove, 9 kWNon-carbonaceous ~30mg/m3
a)
TC (m
g/m
3 STP
@ 1
3% O
2)
Sample #1 2 3 4 5
EC OC (aging) OC (no aging)
Sample #
b)
6 7 8
c)
9 10
Woodchip Boiler, 40 kWNon-carbonaceous ~10mg/m3
d)
* Keller and Burtscher, J. Aerosol Sci., under review.
Average OC increase due to aging: 34%
22.06.2017 A. Keller, ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland 20
How does a better standard should look like
Ecodesign MSC + TC
Avoid large particles - +
Correlate with toxicity - +*
Reflect combustion quality o +
Online or semi-online -(+) +
Suitable for test bench as well as field measurements o(+) +
Include primary as well as secondary emissions - +
Other requirements: low cost, low detection limit, reproducibility, ISO traceable, etc. • Current method: Isokinetic sampling (160°C) and gravimetric analysis • TC: total carbon analysis after conditioning in the micro smog chamber • (+) some online mass measurements systems already available (Wöhler, Testo, …)