EVALUATION OF A NEAR REAL TIME METALS MONITOR FOR … · 2012. 4. 9. · Performance Evaluation 1....

EVALUATION OF A NEAR REAL TIME EVALUATION OF A NEAR REAL TIME METALS MONITOR FOR MEASURING METALS MONITOR FOR MEASURING METALS MONITOR FOR MEASURING METALS MONITOR FOR MEASURING

FUGITIVE METAL EMISSIONSFUGITIVE METAL EMISSIONS

KragKrag PettersonPetterson, John Cooper, John CooperCooper Environmental ServicesCooper Environmental Services

DD Bi iBi iDan Dan BivinsBivinsUSEPA Research Triangle ParkUSEPA Research Triangle Park

Jay TurnerJay TurneryyWashington UniversityWashington University

Will Will WetherellWetherellMissouri Department of NaturalMissouri Department of NaturalMissouri Department of Natural Missouri Department of Natural ResourcesResources

Cooper Environmental Services 1

A&WMA Annual Conference and ExhibitionCalgary, AlbertaJune 25th 2010

Why Measure Metals?Why Measure Metals?Significant adverse human health effects D l i f & hild i ll iblDeveloping fetus & children especially susceptibleMetals represent 8 of US EPA’s top 33 pollutants of highest concernhighest concernArsenic is one of US EPA’s top three priority pollutants of concernsGlobal concern for metals – As, Cd, Cr, Pb and Hg -RoHS, WEEE and ELV

•US EPA. Framework for Inorganic Metals Risk Assessment. EPA 120/R-07/001. March 2007.•US EPA. Health Effects Notebook for Hazardous Air Pollutants. EPA Technology Transfer Network Air Toxics Website. Accessed November 2007. •Agency for Toxic Substances and Disease Registry. Toxic Substances Portal. Accessed November 2007.


g y g y

Why Measure Metals (cont.)?Why Measure Metals (cont.)?Metals Represent Significant Health ConcernsMetals Represent Significant Health Concerns

fNo apparent threshold for blood PbPersistent – will not biodegradeSource Apportionment


Why Continuous Real Time Metals Why Continuous Real Time Metals M it ?M it ?Monitors?Monitors?

Improved exposure estimates – time resolution of concentration dataImproved source apportionment – Correlation with wind speed and wind direction

Pl t f db k t li i t blPlant feedback to eliminate problemIdentification of hot spots


Xact 620 Ambient Metals MonitorXact 620 Ambient Metals Monitor

PM size-selective inlet

Heater

Sampling & Analysis Module

Operator interface/control panelOperator interface/control panel

Flow control module


X t 620 O ti i d f Xact 620 Optimized for Remote Measurement of

As Se Hg & Pb As, Se, Hg & Pb

Cooper Environmental Services 6 6

XACT SAMPLING AND ANALYSISXACT SAMPLING AND ANALYSIS

XX--Ray TubeRay TubeAerosol DepositAerosol DepositSample FlowSample Flow

Analysis AreaAnalysis AreaAnalysis AreaAnalysis Area

Filter TapeFilter Tape Filter TapeFilter Tape

Previous sample is analyzed while t l i ll t dcurrent sample is collected


Performance EvaluationPerformance Evaluation

1. Minimum Detection Limit Determination2. Precision – Comparison of Two Xacts Side by Side3. Comparison with Federal Reference Method (FRM)

Sampling followed by XRF Analysis (IO 3.3)a

a. EPA Compendium of Methods for the Determination of Inorganic Compounds in Ambient Air. June,1999. Compendium Method IO-3.3 Determination of Metals in Ambient Particulate Matter Using X-Ray Fluorescence (XRF) Spectroscopy.


Minimum Detection Limit Minimum Detection Limit DeterminationDetermination

Laboratory EvaluationCalculated from Spectrum of Blank Sample (95 %Calculated from Spectrum of Blank Sample (95 % Confidence Interference Free)

– Standard method used to calculate MDL for XRF d l l i li i i i O 3 3– Used to calculate Detection limits given in IO 3.3

Empirically Determined - 66 Blank SamplesReplaced PM inlet ith Hi ol filter to remo e PM– Replaced PM inlet with Hi-vol filter to remove PM

– Operated the unit in 1 hour mode– 95% confidence MDL – means 95% of samples fell below stated

i l lconcentration level


XactXact 620 Minimum Detection Limits620 Minimum Detection Limits

General Agreement between calculated and empirical detection limitsdetection limits

– Fe and Ca are exceptions

Xact 1 hour detection limits areXact 1 hour detection limits are generally lower than 24-hour FRM/IO 3.3 Detection Limits

All these MDL’s are Interference Free


Field Deployment Field Deployment -- SummarySummary

Two Xact 620sTwo Reference Method Samplers (R & P 2025)Near Fence Line of a Primary Lead Smelter (Herculaneum, Missouri)25 d f R ti25 days of Run time605 hourly data points on each Xact23 elements reported per Xact23 elements reported per XactOver 27,000 individual hourly concentration data points


Field Deployment Field Deployment -- LocationLocation

LeadLead Smelter

Sampling Location


Field Deployment Field Deployment -- SamplersSamplers

Reference Method Inlets

CES X t I l tCES Xact Inlet

M DNR X tMoDNR Xact Inlet


XactXact PrecisionPrecision

Percent Difference Used as an Indicator– Non-detects and concentrations less than the limit of Quantitation

( i t l th ti th i i d t ti li it) d(approximately three times the minimum detection limit) removed from calculation

Least squares linear regression also used as an indicator of PrecisionPrecision also calculated for FRM IO 3.3


Precision Precision –– Percent DifferencePercent Difference

Precision of hourly Xact measurement similar to that of 24Precision of hourly Xact measurement similar to that of 24 hour FRM sampling with XRF analysisPrecision of Xact daily average is generally better than that for FRM li i h XRF l iFRM sampling with XRF analysisXact Precision results include two samplers and two analyzersFRM Precision results include two samplers and one analyzer


FRM Precision results include two samplers and one analyzer

Precision Precision –– Regression AnalysisRegression Analysis

y = 0.992x + 9.8452R² 0 9991

40,000

50,000

3)

Lead

y = 0.9787x + 4.7124R² = 0.9942

5,000

6,000

3)

Calcium

R² = 0.9991

20,000

30,000

uri X

act Co

nc (ng/m

3 R 0.9942

3,000

4,000

uri X

act Co

nc (ng/m

30

10,000

Missou

0

1,000

2,000

Missou

0 10000 20000 30000 40000 50000

CES Xact Conc (ng/m3)

0 1000 2000 3000 4000 5000 6000

CES Xact Conc (ng/m3)


Comparison With FRM/IO 3.3Comparison With FRM/IO 3.3

FRM /IO 3.3 – 24 hour sample on Teflon Filter followed by XRF AnalysisXRF Analysis Performed by CESQA of XRF Analysis Performed by USEPA ORDX t 24 h l l t d b i 24 1 hXact 24 hour average calculated by averaging 24, 1 hour samplesDaily average Xact (average of two units) plotted against y g ( g ) p gdaily average FRM/IO 3.3 (average of two samplers when both where available)All d t i l d dAll data included


XRF Analysis “Round Robin” XRF Analysis “Round Robin” Comparison (CES and EPA)Comparison (CES and EPA)p ( )p ( )

Element N Average Percent DifferenceDifference

Calcium (Ca) 23 - 0.41Iron (Fe) 23 0.92

( )Copper (Cu) 8 8.66Zinc (Zn) 23 - 1.48Lead (Pb) 23 -0.42

Excellent Agreement Between CES’ and EPA’s XRF Analysis


XactXact 620 and FRM/IO 3.3 Comparison620 and FRM/IO 3.3 Comparison

Calcium Iron

y = 1.1085x + 9.5378R² = 0.99441500

2000

m3 )

y = 1.1904x - 0.8478R² = 0.9928

300

350

400

450

m3 )

500

1000

ct C

alci

um (n

g/m

100

150

200

250

Xac

t Iro

n (n

g/m

0

0 500 1000 1500 2000

Xac

FRM Calcium (ng/m3)

0

50

100

0 50 100 150 200 250 300 350 400

FRM Iron (ng/m3)FRM Calcium (ng/m ) FRM Iron (ng/m )



Copper Zinc

y = 1.2193x + 0.5128R² = 0.991230

35

40

y = 1.0626x + 5.0407R² = 0 9989

800

1000

R 0.9912

15

20

25

30

oppe

r (ng

/m3 )

R² = 0.9989

400

600

Zinc

(ng/

m3 )

0

5

10

0 5 10 15 20 25 30 35 40

Xac

t Co

0

200

0 200 400 600 800 1000X

act

0 5 10 15 20 25 30 35 40

FRM Copper (ng/m3)

0 200 400 600 800 1000

FRM Zinc (ng/m3)



Lead

y = 1.0651x + 15.451R² = 0.9976

2000

2500

3000

m3 )

1000

1500

Xac

t Lea

d (n

g/m

0

500

0 500 1000 1500 2000 2500 3000

X

FRM Lead (ng/m3)FRM Lead (ng/m3)


Post Field Test Work Post Field Test Work –– Deposition Size Deposition Size MeasurementsMeasurements

Xact Tape DepositsXact Tape Deposits

Deposit Area Used = 0.747 cm2

Corrected Deposit Area = 0.707 cm2 (6%)Determined with actual Herculaneum deposits


ConclusionsConclusions

1 hour Minimum Detection Limits Are as good or better than those achieved using 24 hour FRM S li d L b t XRF A l iSampling and Laboratory XRF AnalysisPrecision – Xact 620 Exhibits Excellent Precision – 1 hour precision is as good as 24 hour sampling and analysis1 hour precision is as good as 24 hour sampling and analysis– Precision of 24 hour average is better than FRM/XRF

Comparability – Agrees well with FRM/XRF analysis i l l f h l iparticularly for those metals present at concentrations

well above MDL


WHAT YOU MISS WITH 24 HOUR WHAT YOU MISS WITH 24 HOUR AVERAGESAVERAGES

50000

Hourly Lead Concentrations 8/27/2010

Is this degree of variability included in exposure

35000

40000

45000

)

1 hour = 11% of total monthly exposure to Pb

Is this degree of variability included in exposure risk assessments?

20000

25000

30000

Pb Con

c (ng/m

3

0

5000

10000

15000

Pb Daily Ave = 2700 ng/m3

0

Hour


Further WorkFurther Work

Further Data Analysis – Jay TurnerAnalysis of Filter samples by ICP-MS


AcknowledgementsAcknowledgements

• Jim Brunnert, Celeste Koon – Missouri Department of Natural Resources

• Terri Conner, Bob Willis – USEPA Office of Research and Development

• US EPA Office of Air Quality Planning and Standards• US EPA Office of Air Quality Planning and Standards• MACTEC Federal Programs


QUESTIONSQUESTIONS??

For further information:www.cooperenvironmental.comorKrag [email protected]


503 670 8335

EVALUATION OF A NEAR REAL TIME METALS MONITOR FOR … · 2012. 4. 9. · Performance Evaluation 1....

Documents

Transcript of EVALUATION OF A NEAR REAL TIME METALS MONITOR FOR … · 2012. 4. 9. · Performance Evaluation 1....