EVALUATION OF ATMOSPHERIC IMPACTS OF COATINGS VOCs · PDF fileW. P. L. Carter 04/11/2005 Atmospheric Impacts of Coatings VOCs 1 EVALUATION OF ATMOSPHERIC IMPACTS OF COATINGS VOCs William

W. P. L. Carter 04/11/2005 Atmospheric Impacts of Coatings VOCs 1

EVALUATION OF ATMOSPHERICIMPACTS OF COATINGS VOCs

William P. L. CarterCE-CERT, University of California, Riverside

April 13, 2005

Outline• Background and Objectives• Environmental Chamber Experiments – Ozone Impacts• Environmental Chamber Experiments – PM Impacts• Exploratory Glycol Availability Experiments• Estimation of Hydrocarbon Solvent Reactivities• Direct Reactivity Measurement• Summary and Conclusions


Recent UCR Coatings Reactivity ProjectsEvaluation of Atmospheric Impacts of Selected Coatings VOCs• CARB Contract 00-333

• Objective: Reduce uncertainties in estimations of ozone impactsof coatings VOCs

• Final report at http://www.cert.ucr.edu/~carter/coatings

Environmental Chamber Studies of VOC Species inArchitectural Coatings and Mobile Source Emissions• SCAQMD Contract No. 03468

• Relevant Objectives:• Evaluate O3 impacts of selected water-based coatings VOCs• Determine PM impacts for Coating VOCs studied for CARB• Evaluate use of chamber for availability studies

• Final report in preparation


Components of Coatings Projects• Environmental chamber studies

• Six complex hydrocarbon solvents and four water-basedcoatings VOC compounds chosen for study

• UCR EPA chamber employed• Chamber results used to evaluate ozone reactivity

predictions of the SAPRC-99 mechanism• PM measurement results used to derive qualitative

estimates of relative PM impacts of solvents studied• Exploratory studies of effects of aerosol and humidity on

glycol availability

• Development and evaluation of general procedures to estimatereactivities of complex hydrocarbon solvents

• Further development and evaluation of direct reactivitymeasurement methods


Measurement or Calculationof Ozone Reactivities of VOCs

• VOC Reactivities measured in chamber experiments are notexactly the same as VOC reactivities in the atmosphere.• Impractical to duplicate all relevant conditions• Chamber experiments have wall effects, static conditions,

higher levels of test VOCs, etc.

• Atmospheric Ozone impacts of VOCs must be calculated usingcomputer airshed models, given:• Models for airshed conditions• Chemical mechanism for VOC’s Atmospheric Reactions

• BUT mechanisms have uncertainties. reactivity calculations canbe no more reliable than the chemical mechanism used.

• Therefore, the purpose of chamber experiments is to testthe ability of the mechanisms to predict reactivity in models


Diagram of UCR EPA Chamber

20 ft. 20 ft.

20 ft.

This volume kept clear to maintain light

uniformity

Temperature controlled room flushed with purified air and with reflective

material on all inner surfaces

Dual Teflon Reactors

Two air Handlers are located in the corners on each side of the light

(not shown).

Gas sample lines to laboratory below Access

Door

200 KW Arc Light

2 Banks of Blacklights

SEMS (PM) Instrument

Floor Frame

Movable top frame allows reactors to collapse

under pressure control

Mixing SystemUnder floor of reactors


Photographs of Chamber and LightsLooking Towards Reactors (from light) Looking Towards Lights and Air Inlet

ArcLight

PartiallyFilledReactors

BlackLights

Reflectivewalls

AirIntake

(Alsoone on

otherside)

Reflectivewalls and

Ceiling


Incremental Reactivity Experiments

• Objective is to determine effects of VOC’s reactions in chemicalenvironments representing a range of atmospheric conditions.

• Approach is to conduct two simultaneous experiments in thedual chamber• Base Case Experiment: Irradiate surrogate ROG – NOx

mixture simulating an ambient chemical environment• Test Experiment: Same as base case experiment except

that a test compound or solvent added

• Effect of added VOC on O3, radicals, etc, provides a means totest model predictions reactivities of VOCs under similarconditions in the atmosphere

• Base case experiments should reflect range of atmosphericchemical conditions relevant to VOC reactivity.


Choice of Base Case forIncremental Reactivity Experiments

• Major atmospheric chemical condition relevant to VOC reactivityis relative availability of NOx (relative ROG/NOx ratio). E.g.,

• Different aspects of the mechanism affect O3 impacts underdifferent NOx conditions• High NOx experiments: test effects of VOCs on radical levels• Low NOx experiments: test effects of VOCs on NOx removal

• Therefore, minimum of two base case experiments is needed

High

Moderate

Low

ROG/NOx

Most sensitive to NOxLowLow NOx

Optimum conditions for O3ModerateMOIR

Most sensitive to VOCsHighMIR

O3 sensitivityRelative NOxCondition


Base Case Experimentsused in Coatings Reactivity Studies

• NOx levels of 25-30 ppb, based on CARB recommendations ofrange of NOx that represents urban conditions in California

• 8- component ROG surrogate (used previously) employed torepresent major classes of VOCS present in ambient air• Formaldehyde removed in later experiments and other

VOCs increased by 10% to simplify experiments• Calculated to give essentially the same reactivity results as

base ROG mixture used to calculated reactivity scales.

½ MOIR NOx levels. NOx relatively low butnot so low that O3 insensitive to VOCs.

1.025MOIR/2

ROG levels calculated to yield MIRconditions

0.530MIR

DiscussionROG(ppmC)

NOx(ppb)Experiment


SAPRC-99 Model simulations of representative experiments

0.00

0.04

0.08

0.12

0.16

0 60 120 180 240 300 360

Time (minutes)

Ozo

ne (p

pm)

ModelEPA095BEPA123AEPA124BEPA126A

Experimental and Calculated O3 forRepresentative Base Case Experiments

0.00

0.04

0.08

0.12

0.16

0 60 120 180 240 300 360

Time (minutes)

Ozo

ne (p

pm)

Model

EPA110B

EPA114A

EPA127B

MOIR/2 BASE CASE(NOx=25 ppb, ROG=1 ppmC)

MIR BASE CASE(NOx=30 ppb, ROG=0.5 ppmC)


Dependence of SAPRC-99 UnderpredictionBias on Relative ROG/NOx Levels

-20%

0%

20%

40%

60%

80%

0.1 1 10

ROG / NOx Ratio Relative to ROG / NOx Giving Maximum O3

([O3 ]-

[NO

]) M

odel

und

erpr

edic

tion

erro

r(E

xper

imen

tal -

Mod

el) /

Exp

erim

enta

l

UCR EPA ChamberCSIRO Outdoor ChamberTVA Chamber

Maximum O3 (MOIR)Approximate MIR


Aromatic Mechanism Adjustmentsto Improve Base Case Simulations

• Biases in model simulations of base case experiment attributedto aromatics mechanism. So far, no aromatic mechanism foundthat gives satisfactory fits all the available chamber data.

• Base case biases may cause biases in simulations of reactivity.

• To investigate this, mechanisms for aromatics in the base ROGwas adjusted to remove biases in base case simulations.• Yields of aromatic fragmentation products AFG2 and AFG3

for toluene and m-xylene increased by factor of 1.75• Rate of reaction of AFG1 with O3 increased by factor of 10

• Comparing simulations with and without this adjustment showseffects base mechanism biases on reactivity predictions

• Not a “better” aromatics mechanism because this adjustmentmakes fits to single aromatic – NOx experiments worse.


Effects of Mechanism Adjustments onSimulations of Representative Experiments

∆([O

3]-[N

O])

(ppm

)

Surrogate - NOx Runs M-Xylene - NOx RunsEPA188A: Low ROG/NOx (MIR) CTC036: Low ROG/NOx

EPA124B: High ROG/NOx (MOIR/2) CTC029: High ROG/NOx

Time (minutes)

0.0

0.2

0.4

0.6

0 60 120 180 240 300 360

0.0

0.2

0.4

0.6

0 60 120 180 240 300 360

0.00

0.04

0.08

0.12

0.16

0 60 120 180 240 300 360

ExperimentalStandard ModelAdjusted Model

0.00

0.04

0.08

0.12

0.16

0 60 120 180 240 300 360


Effect of Aromatic Mechanism Adjustmenton Base Case Model Underprediction Bias

-30%

-20%

-10%

0%

10%

20%

30%

40%

0 10 20 30 40 50 60ROG/NOx

Mod

el u

nder

pred

ictio

n Bi

as

Standard Model Adjusted Aromatics Model

MIR MOIR/2

∆([O

3]-[N

O])


VOCs Identified in the 2001 CARB Surveyof Water-Based Architectural Coatings

O

O OH

OO

OH

Mixtures of C8-C12 alkanes andaromatics4%Various Hydrocarbon

Solvents

3%Butyl Carbitol;2-(2-Butoxyethoxy)-Ethanol

(2)

3%Benzyl Alcohol

21%Ethylene Glycol

25%Propylene Glycol

(1)31%

Texanol®;Isobutyrate esters of 2,2,4-Trimethylpentyl-1,3-diol

StructuresMass %Compound

OHOH

OHOH

OH O

O

OH


Results of a Texanol® Injection Test

0.0

0.2

0.4

0.6

0.8

1.0

0 5 10 15 20 25Hours after injection

ppm

C T

exan

ol o

r TH

C

FID Total Hydrocarbon Analyzer DataCalculated amount injected, corrected for dilution using CO dataGC Analysis, Based on liquid calibration (sum of two GC peaks)


0.0

0.2

0.4

0.6

0.8

1.0

0.0 0.2 0.4 0.6 0.8Ln ([m-Xylene]0 / [m-Xylene]t)

ln([T

est V

OC

]0 /

[[Tes

t VO

C]t )

Texanol 1

Texanol 2 (offset)

Butyl Carbitol / 2 (offset)

Best fit to data

SAPRC-99 Estimate

OH Radical Rate ConstantsDerived from Chamber Data

4.29 x 10-11Butyl Carbitol

1.62 x 10-11Texanol 2

1.29 x 10-11Texanol 1

kOH (cm3

molec-1 s-1)Compound


Reactivity Data for Texanol®∆[(O3]-[NO]) (ppm) ∆[(O3]-[NO]) Change (ppm) IntOH (ppt-min)

Time (minutes)

EP

A22

9A (M

IR)

EP

A23

2A (M

OIR

/2)

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.03

-0.02

-0.01

0.00

0.01

0 60 120 180 240 300 360

0

10

20

30

40

0 60 120 180 240 300 360

Test Experiment Standard Model Standard Base ModelBase Experiment Adjusted Aromatic Model

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.02

-0.01

0.00

0.01

0 60 120 180 240 300 360

0

20

40

60

0 60 120 180 240 300 360


Comparison of Chamber and AmbientReactivity Calculation For Texanol®

"MIR" Incremental Reactivity Chamber experiment

"Averaged Conditions" MIRBox Model Airshed Scenario

Both Cases: Moles Texanol added = 5% of moles Carbon in Base Case ROGsBoth simulations predict measurable effect of Texanol on OH radical levels.

0.00

0.05

0.10

0.15

0 2 4 6Irradiation time (hour)

Ozo

ne (p

pm)

Base Case

Added Texanol

0.00

0.05

0.10

0.15

0.20

0.25

8 10 12 14 16 18Simulated Local Time (hour)

Ozo

ne (p

pm)

Base Case

Added Texanol

Two curvesalmost on topof each other


Reactivity Data for Butyl Carbitol∆[(O3]-[NO]) (ppm) ∆[(O3]-[NO]) Change (ppm) IntOH (ppt-min)

Time (minutes)

EP

A35

2B (M

IR)

EP

A35

3B (M

OIR

/2)


0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.02

-0.01

0.00

0.01

0.02

0 60 120 180 240 300 360

0

20

40

60

0 60 120 180 240 300 360

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.02

-0.01

0.00

0.01

0 60 120 180 240 300 360

0

10

20

30

40

0 60 120 180 240 300 360


Reactivity Data for Propylene Glycol∆[(O3]-[NO]) (ppm) ∆[(O3]-[NO]) Change (ppm) IntOH (ppt-min)

Time (minutes)

EP

A27

7A (M

IR)

EP

A25

2A (M

OIR

/2)

Test Experiment Standard Model Standard Base ModelBase Experiment Adjusted Aromatics Model

0.00

0.05

0.10

0.15

0.20

0.25

0 60 120 180 240 300 3600.00

0.02

0.04

0.06

0 60 120 180 240 300 3600

20

40

60

80

0 60 120 180 240 300 360

0.00

0.05

0.10

0.15

0.20

0.25

0 60 120 180 240 300 3600.00

0.02

0.04

0.06

0 60 120 180 240 300 3600

10

20

30

40

0 60 120 180 240 300 360


Reactivity Data for Ethylene Glycol∆[(O3]-[NO]) (ppm) ∆[(O3]-[NO]) Change (ppm) IntOH (ppt-min)

Time (minutes)

EP

A27

8B (M

IR)

EP

A25

0B (M

OIR

/2)

Test Experiment Standard Model Standard Base ModelBase Experiment Adjusted Aromatics Model

0.0

0.1

0.2

0.3

0 60 120 180 240 300 3600.00

0.02

0.04

0.06

0.08

0 60 120 180 240 300 3600

20

40

60

0 60 120 180 240 300 360

0.0

0.1

0.1

0.2

0.2

0.3

0 60 120 180 240 300 3600.00

0.01

0.02

0.03

0.04

0.05

0 60 120 180 240 300 3600

10

20

30

40

0 60 120 180 240 300 360


Glycol Decay Rates in Reactivity Runs:Comparison with Literature k(OH) Values

Propylene Glycol vs m-Xylene Ethylene Glycol vs. m-Xylene

ln([m-Xylene]0/[m-Xylene]t)

ln([G

lyco

l] 0/[G

lyco

l] t)

-0.1

0.1

0.3

0.5

0.7

0.9

1.1

1.3

0.0 0.2 0.4 0.6 0.8 1.0 1.2

Single Run Single Run Single RunSingle Run Single RunNot used for kOH Fit Best kOH fit to Data Literature kOH Line

0.0

0.2

0.4

0.6

0.8

0.0 0.3 0.6 0.9 1.2


Glycol Reactivity Data with aNon-Aromatic Surrogate

∆[(O3]-[NO]) (ppm) ∆[(O3]-[NO]) Change (ppm) IntOH (ppt-min)

Time (minutes)

Pro

pyle

ne G

lyco

lE

thyl

ene

Gly

col

Test Experiment Standard ModelBase Experiment Standard Base Model

0.00

0.04

0.08

0.12

0.16

0 60 120 180 240 300 3600.00

0.02

0.04

0.06

0.08

0 60 120 180 240 300 3600

5

10

15

20

25

0 60 120 180 240 300 360

0.00

0.04

0.08

0.12

0.16

0 60 120 180 240 300 3600.00

0.02

0.04

0.06

0 60 120 180 240 300 3600

10

20

30

0 60 120 180 240 300 360


Development of aBenzyl Alcohol Mechanism

• Reaction with OH radicals assumed to dominate.

• Single measurement of k(OH) given by Atkinson (1989) used

• Reaction at -CH2OH, forming Benzaldehyde + HO2 assumed tooccur 30% of the time, to fit benzaldehyde data in experiments

• Mechanism of OH addition to ring based on that used fortoluene

• Overall nitrate yield adjusted to be 5% to give best fits to data

• Benzaldehyde – NOx experiments also carried out to provideadditional basis for developing adjusted mechanism.


Reactivity Data for Benzyl Alcohol∆[(O3]-[NO]) (ppm) ∆[(O3]-[NO]) Change (ppm) IntOH (ppt-min)

Time (minutes)

EP

A32

3B (M

IR)

EP

A30

2 (M

OIR

/2)


0.00

0.05

0.10

0.15

0 60 120 180 240 300 360 -0.02

0.00

0.02

0.04

0.06

0 60 120 180 240 300 360 0

20

40

60

0 60 120 180 240 300 360

0.00

0.04

0.08

0.12

0.16

0 60 120 180 -0.005

0.000

0.005

0.010

0.015

0.020

0.025

0 60 120 1800

10

20

30

0 60 120 180


Simulations of RepresentativeBenzaldehyde – NOx Experiments

∆([O3]-[NO]) (ppm) Benzyl Alcohol (ppm) Benzaldehyde (ppm)EPA322A (0.41 ppm Benzyl alcohol, 26 ppb NOx)

EPA325A (0.27 ppm Benzyl alcohol, 55 ppb NOx)

Irradiation time (minutes)

0.000

0.005

0.010

0.015

0.020

0.025

0 120 240 3600.0

0.1

0.2

0.3

0 120 240 360

ExperimentalModel

0.0

0.1

0.2

0.3

0.4

0 120 240 360

0.00

0.05

0.10

0.15

0.20

0.25

0 120 240 360

0.0

0.1

0.2

0.3

0 120 240 3600.00

0.01

0.02

0.03

0 120 240 360


Summary of Mechanism Evaluation Resultsfor Water-Based Coatings VOCs

2.9Previous mechanism simulated datasatisfactorily. Not changed.

Butyl Carbitol

4.9No previous mechanism. Parameter-ized mechanism adjusted to fit data.

Benzyl Alcohol

≥ 3.4Mechanism may underpredict ozoneimpact, but uncertain whether changeis appropriate. Not changed.

Ethylene Glycol

≥ 2.7Mechanism may underpredict ozoneimpact, but uncertain whether changeis appropriate. Not changed.

Propylene Glycol

0.88Previous mechanisms simulated datasatisfactorily. Not changed.

Texanol® isomers

MIR *(mass basis)

Mechanism Performance andModificationsCompound

* MIR of base ROG (Ambient Mixture) = 3.7 gm O3 / gm VOC


Representative Hydrocarbon MixturesChosen For Reactivity Experiments

11-9-12Dearomatized MixedAlkanes

ASTM-1C

60.2%8-9VMP NaphthaVMP-NAPH

AROM-100

ASTM-1A

ASTM-1B

ASTM-3C1

Designation

12-Mostly 11Synthetic isoparrafinicalkane mixture

22

15

14

CARBBin No.

AromaticContent

Carbon No.Range

Description

100%Mostly 9Aromatic 100

19%9-12Regular Mineral Spirits

6%9-12Reduced AromaticsMineral Spirits


Chemical Type Distributionsfor Hydrocarbon Mixtures Studied

n-Alkane Br-Alkane Cyc-Alkane Aromatic

Aromatic 100ASTM-1AASTM-1B

ASTM-3C1ASTM-1CVMP Naphtha


Reactivity Data for Dearomatized MixedAlkanes (ASTM-1C)


Time (minutes)

EP

A16

8A (M

IR)

EP

A15

2B (M

OIR

/2)


0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.06

-0.04

-0.02

0.000 60 120 180 240 300 360

0

20

40

60

0 60 120 180 240 300 360

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.05

-0.04

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360

0

10

20

30

40

0 60 120 180 240 300 360


Comparison of Chamber and AmbientReactivity Calculation for ASTM-1C

"MIR" Incremental Reactivity Chamber experiment

"Averaged Conditions" MIRBox Model Airshed Scenario

Moles Carbon Mixture added = 25% of Moles Carbon in Base Case ROGsBoth simulations predict measurable effect of Mixture on OH radical levels.

0.00

0.05

0.10

0.15

0 2 4 6Irradiation time (hour)

Ozo

ne (p

pm)

Base Case

Added ASTM1C

0.00

0.05

0.10

0.15

0.20

8 10 12 14 16 18Simulated Local Time (hour)

Ozo

ne (p

pm)

Base Case

Added ASTM1C


Reactivity Data for Aromatic 100∆[(O3]-[NO]) (ppm) ∆[(O3]-[NO]) Change (ppm) IntOH (ppt-min)

Time (minutes)

EP

A24

4B (M

IR)

EP

A23

9A (M

OIR

/2)


0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 3600.00

0.02

0.04

0.06

0 60 120 180 240 300 3600

20

40

60

80

0 60 120 180 240 300 360

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.04

-0.02

0.00

0.02

0.04

0 60 120 180 240 300 360

0

10

20

30

40

0 60 120 180 240 300 360


Results for Other Petroleum DistillatesChange in ∆([O3]-[NO]) (ppm)

VMP Naphtha ASTM-1B ASTM-1A

Time (minutes)

MIR

MO

IR/2


EPA167A

-0.02

-0.01

0.00

0.01

0.02

0 60 120 180 240 300 360

EPA153A

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360

EPA151A

-0.04

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360

EPA242B

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360

EPA238A

-0.05

-0.04

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360

EPA243B

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360


Reactivity data for SyntheticHydrocarbon Mixture (ASTM-3C1)


Time (minutes)

EP

A16

3A (M

IR)

EP

A23

7B (M

OIR

/2)

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.05

-0.04

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360

0

10

20

30

40

0 60 120 180 240 300 360


0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360 -0.05

-0.04

-0.03

-0.02

-0.01

0.000 60 120 180 240 300 360

0

20

40

60

0 60 120 180 240 300 360


Assessment of Model Performance andMIRs for the Hydrocarbon Solvents Studied

7.51

1.82

1.26

0.81

0.91

1.41

BinMIR

Consistent with data forMIR. Underpredicts O3inhibition at low NOx

Reasonably consistentwith data.

Reasonably consistentwith data

Model underestimatesO3 impact by 40-80%



O3 Model Performance

7.70

1.97

1.21

Approx.1.1 – 1.5

0.96

1.35

Best Est.MIR

C9-C12 AlkanesASTM-1C

C8-C9 AlkanesVMPNaphtha

ComponentsDesignation

Methyl ethyl andtrimethylBenzenes

Aromatic100

C9-C12 Alkanes,~20% AromaticsASTM-1A

C9-C12 Alkanes,~6% AromaticsASTM-1B

C11 BranchedAlkanesASTM-3C1


PM Measurements

• Number densities of particles in 71 size ranges (28 - 730 nm)measured using a a Scanning Electrical Mobility Spectrometer

• Data used to compute total particle number and volume(measured as mass assuming density of H2O) per unit volume

• PM alternately sampled from each of the two reactors, switchingevery 10 minutes (15.3 data points/hour/reactor)

• PM measurements made during most incremental reactivityexperiments for the coatings projects

• Background PM measurements made in experiments where PMprecursors not expected

• Seed aerosol not used in most experiments


PM Data in Base Case Experiment

EPA233: MOIR/2 Surrogate (Side Equivalency Test)PM Number (cm-3) PM Volume (µg/m3)

Irradiation time (hours)

0

2000

4000

6000

8000

10000

12000

14000

0 1 2 3 4 5 60.0

0.5

1.0

1.5

2.0

0 1 2 3 4 5 6

Side A

Side B


PM Volume in Background Experiments 5

Hou

r PM

Vol

ume

(µg/

m3 )

Side A Side B

EPA Run Number

0.0

0.5

1.0

1.5

2.0

150 250 350

Pure Air Propene - NOx CO - Air CO - NOx

0.0

0.5

1.0

1.5

2.0

150 250 350


Effects of Texanol® and the Glycolson 5-Hour PM Volume

5 H

our P

M V

olum

e (µ

g/m

3 )

Side A Side B

EPA Run Number

0.0

0.3

0.6

0.9

1.2

150 200 250 300 3500

1

2

3

4

125 175 225 275 325

Base Experiment Base Fit or Avg. Base SDev.

Texanol Propylene Glycol Ethylene Glycol


Effects of Hydrocarbon Solvents on5 Hour PM Volume

5 H

our P

M V

olum

e (µ

g/m

3 )

Side A Side B

EPA Run Number

0

2

4

6

8

125 175 225 275 325

Base Experiment Base Fit or Avg. Base SDev.VMP Naphtha ASTM-1C ASTM-3C1ASTM-1B ASTM-1A Aromatics 100

0

1

2

3

4

5

125 175 225 275 325


Effects of Benzyl Alcohol and Butyl Carbitolon 5-Hour PM Volume

5 H

our P

M V

olum

e (µ

g/m

3 )

Side A Side B

EPA Run Number

0

10

20

30

40

150 200 250 300 350

Base Experiments Base Fit or AverageBenzyl Alcohol Butyl Carbitol

4-hour

0

5

10

15

20

25

30

150 250 350


Summary of PM Volume Reactivity Results

0 2 4 6 8 10 12 14 16 18 20

Base Case

Propylene Glycol

Ethylene Glycol

Texanol®

ASTM-3C1

ASTM-1C

VMP Naphtha

ASTM-1B

ASTM-1A

Aromatic 100

Butyl Carbitol *

Benzyl Alcohol /2

Average 5 Hour PM Volume µg/m3

Side B Side A

Butyl Carbitol Side A is for 4 hours.

Error Bars are 1-σ standard deviations.No error bar means only one experiment.


Summary of PM Measurement Results

• Background PM formation in chamber is up to ~1 µg/m3,depending on reactor employed• Probably due to contaminant reacting with OH, forming SOA• Reason for higher background in “A” reactor unknown

• Secondary PM formation from ethylene and propylene glycol,Texanol®, and the ASTM-3C1 synthetic mixture negligible.

• Small but measurable PM from petroleum distillate solvents. Notsimply related to aromatic content.

• Highest secondary PM from butyl carbitol and (especially)benzyl alcohol

• Chamber effects PM model needed before PM data can beused for quantitative mechanism evaluation


Glycol Availability Screening ExperimentsObjective

• Determine if added aerosols affect gas-phase consumptionrates and reactivities of glycols

Experiments Carried Out

• Dark decay experiments with ethylene and propylene glycol with10 µg/m3 (NH4)2SO4 seed aerosol at 35% RH.

• ROG – NOx ambient surrogate irradiation with added propyleneglycol with 9 µg/m3 (NH4)2SO4 seed aerosol at 25% RH.

• ROG – NOx ambient surrogate irradiation with added ethyleneglycol with 7 µg/m3 NH4HSO4 seed aerosol at 30% RH.

Note: Aerosol and humidity added to only one reactor becauseof limited aerosol generation and humidification capacity.


Results of Glycol Dark Decay Experiment

Time after first sample (hours)

Con

cent

ratio

n (p

pm)

35% RH10 µg/m3

(NH4)2SO4Seed Aerosol

0.0

0.1

0.2

0 2 4 6

Dry,No Aerosol

0.0

0.1

0.2

0.3

0 2 4 6

Ethylene Glycol

Propylene Glycol


Propylene Glycol Ethylene GlycolDry, no Aerosol

25% RH, 9 µg/m3 (NH4)2SO4 30% RH, 7 µg/m3 NH4HSO4

Irradiation time (minutes) `

0.00

0.05

0.10

0.15

0.20

0.25

0 60 120 180 240 300

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360

Test Data

Base Data

Test Model

Base Model

0.00

0.05

0.10

0.15

0.20

0.25

0 60 120 180 240 300 360

0.00

0.05

0.10

0.15

0.20

0 60 120 180 240 300 360

Effects of Aerosol and Humidity on GlycolReactivity Experiments

Model calculationsused adjustedaromatics basemechanism for bestfits.

No base data inadded aerosol runsbecause of limitedhumidification andaerosol generationcapacity.

∆([O

3]-[N

O])

(ppm

)


Glycol Decay Rates in Availability Runs:Comparison with Literature k(OH) Values

Propylene Glycol vs m-Xylene Ethylene Glycol vs. m-Xylene

ln([m-Xylene]0/[m-Xylene]t)

ln([G

lyco

l] 0/[G

lyco

l] t)

0.0

0.2

0.4

0.6

0.8

0.0 0.2 0.4 0.6 0.8

Humidified Seed Aerosol Run Literature kOH Line

0.0

0.2

0.4

0.6

0.8

0.0 0.3 0.6 0.9 1.2


Glycol Availability Experiments:Preliminary Conclusions

• No clear effect on glycol consumption rate or ozone reactivity forhumidity up to 35% and (NH4)2SO4 or NH4HSO4 seed aerosol upto 10 µg/m3.

• But there still may be a measurable effect at higher humidity oraerosol concentration, with a different type of aerosol

• Upgrades are being made to the chamber facility to facilitateexperiments at higher RH, aerosol levels.

• But experiments that measure increases in aerosol mass whenexposed to gas-phase VOCs may give a more sensitivemeasure of VOC uptake on aerosols


Evaluation of Methods to EstimateComplex Hydrocarbon Solvent Reactivities

• MIRs for complex hydrocarbon solvents currently estimatedusing a “binning” procedure based on correlations betweencarbon numbers, type distributions, and MIRs

• Bin MIRs evaluated by comparison with MIRs calculated usingdetailed compositional data for a wide variety of solvents• Agree within ±25% except for bins with light cycloalkanes

• An alternative “spreadsheet” method developed for derivingestimated compositions for solvents with limited data• Separates compositional and reactivity estimates. Permits

derivations in reactivities for other scales besides MIR.• Agrees with calculations using detailed compositional data

within ±10% in most cases• Can be used as a basis for updating hydrocarbon bin

reactivities when reactivity scale is changed or updated.


Comparison of CARB Bin MIRs withMIRs Calculated Using Compositional Data

-75%

-50%

-25%

0%

25%

50%

75%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Bin Number

Bin

Ass

ignm

ent

Rel

aitv

e to

Det

aile

d C

alcu

latio

n

Analyzed Solvent Bin Average

Bins 1, 3-5 are mixtureswith cycloalkanes in thelowest boiling point range

No data for bins 18-20


Comparison ofSpreadsheet Estimated MIRs with

MIRs Calculated Using Compositional Data

-20%-15%-10%-5%0%5%

10%15%20%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Bin Number

Spr

eads

heet

Cal

cula

tion

Rel

aitv

e to

Det

aile

d C

alcu

latio

n

Analyzed Solvent Average

Note that previous plothad ± 75% range


Further Development of aDirect Reactivity Measurement Method

• VOCs affect O3 directly through their own reactions or indirectlythrough the effects of their reactions on radicals and NOx.

• A measurement of direct reactivity would reduce uncertainties inmechanism evaluations and provide a reactivity screening tool

• A direct reactivity measurement method was developed in aprevious CARB project but was not suitable for coatings VOCs• Required GC analysis, so not suitable for complex mixtures

or low volatility compounds

• For this project, a total carbon measurement was interfaced tothe system to eliminate the need for GC analysis.

• Problems encountered. Absolute direct results not consistentwith model predictions; but better agreement with relative results

• Resources for this task exhausted before it could be completed.


Direct Reactivity MeasurementsRelative to n-Octane

0.0 0.5 1.0 1.5

Mineral Spirits

n-C14

n-C12

n-Octane

Propane (x 3)

Direct Reactivity Relative to n-Octane

Model Experimental Error bars show range of variability of experiments


Overall Conclusions of Chamber Studies• Chamber data for Texanol®, butyl carbitol, and primarily alkane

petroleum distillates are consistent with SAPRC-99 predictions.

• Chamber data for Aromatics-100 consistent with SAPRC-99 forMIR conditions, but O3 inhibition at low NOx underpredicted.

• Reactivities of at least some synthetic hydrocarbon mixturesmay be underpredicted by up to a factor of 2.

• Glycol reactivities underpredicted by ~30% in some experiments,but unclear whether adjustments are appropriate.

• New mechanism developed for benzyl alcohol that simulateschamber data about as well as mechanisms for other aromatics

• Relative secondary PM impacts: benzyl alcohol >> butyl carbitol> petroleum distillates. No measurable PM impacts for others.

• No evidence that humidity and aerosol affects glycol availabilityat the relatively low aerosol loadings and humidities examined


Recommendations• Aromatics mechanisms need to be improved to further reduce

uncertainties in reactivity assessments (e.g., glycols)

• Extrapolation of current mechanisms to higher aromatics, suchas Aromatics 200, still highly uncertain

• Direct reactivity measurements needed to reduce uncertaintiesfor some VOCs, particularly mixtures of branched alkanes.

• A modified base case experiment that gives better correlationsbetween chamber and atmospheric reactivity would be useful

• No compelling need to change current bin assignments, exceptperhaps for those with light cycloalkanes and synthetic mixtures.But new procedure will be needed when reactivity scale updated

• Well-characterized environmental chamber data needed todevelop predictive secondary PM models. Work needed onbackground PM characterization in chambers


Acknowledgements• Funded by CARB and SCAQMD

• CARB, SCAQMD staff and RRAC members: Helpful discussions

• Andrew Jaques of the ACC, Bob Hinrics of Citco, and ArleanMedeiros of ExxonMobil: Helpful discussions and providinghydrocarbon mixture compositional data

• David Morgott and Rodney Boatman of Eastman Kodak: Helpfuldiscussions regarding Texanol®.

• Albert Censullo of California Polytechnic State University:Provided compositional data prior to completion of his report

• Irina Malkina, Kurt Bumiller, Chen Song, Bethany Warren,Dennis Fitz, Charles Bufalino, and John Pisano of CE-CERT:Carried out or assisted with experiments

• David Cocker and Chen Song of CE-CERT: Assistance with PMdata and analysis

Documents

EVALUATION OF ATMOSPHERIC IMPACTS OF COATINGS VOCs · PDF fileW. P. L. Carter 04/11/2005 Atmospheric Impacts of Coatings VOCs 1 EVALUATION OF ATMOSPHERIC IMPACTS OF COATINGS VOCs William