fatcat: a new characterization method for particulate ...fatcat: a new characterization method for...

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


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.


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


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)


• 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


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)


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)


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)


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)


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

)


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


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.


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.


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!


Backup Slides


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)


200

400

600

800

1000

Tem

pera

ture

(°C

)


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)


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)


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%


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, …)

fatcat: a new characterization method for particulate ...fatcat: a new characterization method for...

Documents