Kevin Goohs

Thermo Electron, Inc.

Jay Turner

Washington University in St. Louis

Thermo Electron Model 5030 SHARP for

Measurement of Ambient PM2.5 Mass Concentration

National Air Monitoring Conference

Las Vegas, NV

November 6-9, 2006

Hybrid Concepts

The continuous use of two (2) compatible technologies on a

combined platform to enhance the overall performance while

maintaining it’s original function.

Hybrid Circuit

Hybrid Bicycle

Hybrid Vehicle

Hybrid Spudgun

Hybrid Concepts (continued)

Thermo SHARP

• Nephelometry

• Beta attenuation

This Presentation

• de-Gooh’zd

• Turner-ized

Synchronized Hybrid Ambient Real-Time

Particulate Monitor (SHARP)

• Accuracy of a beta gauge with high time resolution of a nephelometer

– Heater on the inlet line for moisture control

– Nephelometer and beta gauge in series

– Dynamic digital filtering of both data streams to continuously

adjust the nephelometer response using the beta gauge response

• No HVAC control needed –

– Can be sited outdoors in an

environmental shelter, better chance to

capture FRM-like artifacts

– Can be sited indoors

• Digital filtering stabilizes the typically

noisy beta gauge response

– can lead to complicated signal dynamics

• High time resolution

– 1 minute concentration output

Hybrid Dynamic Digital Filtering

General Equation

SHARPn = Rn * (Cββββf_ττττv / Rf_ττττv)nRn = Nephelometer 1 minute running average

Cβf_τv = Dynamically filtered Beta-derived conc.

Rf_τv = Dynamically filtered Neph-derived conc.

Dynamic Digital Filtering

ττττv = ττττo * A’

A’ = dynamic time constant factor

τv = dynamic digital filtering time constant

Conceptualize as a mass scattering efficiency that is dynamically adjusted

using the relationship between aerosol light scattering intensity by

nephelometry and aerosol mass concentration by beta attenuation

SHARP Beta Gauge

• 14C beta emission source

– Low energy electrons, inelastic scattering with the atomic nucleus of

the absorbing material... Generally follows Beer’s Law and is

relatively insensitive to composition of the absorber

– In contrast, high energy electrons susceptible to bremsstrahlung

radiation (inelastic scattering with the nuclear column field) which

can be very sensitive to composition of the absorber

– thus, can we rule out site-to-site variations in SHARP response

(compared to a reference method) being driven by differences in

aerosol composition?

• Calibrations performed with muscovite films, mass absorption

properties similar to ammonium sulfate and other major components in

ambient aerosol

Sample Conditioning

• Relative humidity measured immediately upstream of the

beta gauge filter tape

– User-selected setpoint RH

• Typically 40% RH USA and 65% RH EU

– Sample stream is conditioned upstream of the

nephelometer

• PID controller for applying a heater duty

– Above setpoint RH, heating is applied subject to a

constraint for the maximum allowable temperature

difference

– Below setpoint RH, heater is turned off

Field Tests

• New Haven, CT

• East St. Louis, IL (St. Louis – Midwest Supersite)

• Bakersfield, CA

• Dayton, CA

Field Testing – New Haven, CT

After statistical conditions are

met, the site specific calibration

factor begins to be continuously

applied to the nephelometer in

real-time

Field Testing – New Haven, CTJanuary 25 -February 28, 2005

24-hour SHARP & FH62C14 PM2.5 Concentrations vs FRM - Criscuolo Park, New Haven

Winter Evaluation : January 25 through February 28, 2005

y = 0.99x + 0.82

R2 = 0.97

y = 0.94x + 1.32

R2 = 0.97

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30 35 40

FRM (µµµµg/m3)

SHARP & FH62C14 (µµ µµg/m

3)

1:1 SHARP FH62C14 Linear (SHARP) Linear (FH62C14)

SHARP = 0.99 x FRM + 0.8

BETA = 0.94 x FRM + 1.3

Field Testing – New Haven, CTApril 1 – June 15, 2005

24-hour SHARP & FH62C14 PM2.5 Concentrations vs FRM - Criscuolo Park, New Haven

Spring Evaluation : April 1 through June 15, 2005

y = 0.97x + 1.26

R2 = 0.96

y = 0.93x + 1.62

R2 = 0.96

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30 35 40

FRM (µµµµg/m3)

SHARP & FH62C14 (µµ µµg/m

3)

1:1 SHARP FH62C14 Linear (SHARP) Linear (FH62C14)

SHARP = 0.97 x FRM + 1.3

BETA = 0.93 x FRM + 1.6

Field Testing – New Haven, CT (2005)

• Hourly comparisons between Thermo SHARP, Thermo FDMS TEOM

and MetOne 1020 BAM [posted materials will show scatter plots]

• Regression of Monitor “X” on SHARP…

0.830.620.80R2

-0.013.082.00Intercept

1.020.941.03Slope

0.910.800.90R2

0.24-0.490.08Intercept

1.050.980.90Slope

6/16/-9/294/1-6/151/25-2/28Period

FDMS

TEOM

MetOne

1020 BAM

East St. Louis SHARP Deployment

• Beta testing with prototype units commenced in spring 2004

– Numerous hardware and software revisions

• One-year field performance evaluation with production units

commenced April 2005

– No revisions to the hardware or software were implemented

– Two units: a primary monitor and collocated monitor

• Primary monitor used for comparisons to FRM

• Collocated unit typically operated identical to primary unit

for estimating precision

– Operating conditions periodically changed, e.g.

dynamic zero sampling (HEPA filter on inlet), different

beta gauge tape advance schedules

• Default operating conditions:

– Both units outdoors in environmental shelters but with no

HVAC control

– One/day beta gauge tape advances (midnight CST)

East St. Louis SHARP Deployment

• Virtually no field operations issues

– Routine maintenance performed weekly; much lower

frequency likely needed

– One instance of the tape advance system not resealing

after a tape advance event at very cold temperature (not

an issue for deployments inside a HVAC-controlled

shelter)

• Beta gauge recalibrated after one year in the field –

0.5% change

Challenging the Moisture Control System:

Environmental Conditions in East St. Louis

-20

-10

0

10

20

30

40

01/01/05 02/15/05 04/01/05 05/16/05 06/30/05 08/14/05 09/28/05 11/12/05 12/27/05

10m temperature, degrees Celcius

0

10

20

30

40

50

60

70

80

90

100

01/01/05 02/15/05 04/01/05 05/16/05 06/30/05 08/14/05 09/28/05 11/12/05 12/27/05

relative humidity, %

sustained, subfreezing

temperatures in the winter

daily maximum RH

routinely exceeds 85%,

regardless of season

Hourly Average Collocated SHARP

excellent agreement

between collocated

SHARP instruments

1-hour average primary SHARP, µg/m3

0 20 40 60 80 100

1-hour average collocated SHARP, µg/m

3

0

20

40

60

80

100

April 2005 - January 2006 (N=4158)

collo precision = 1.3 µg/m3 (8%)

collo SHARPmean

/ primary SHARPmean

= 1.02

Daily-Integrated FRM and Daily-Average SHARP

Collocated Performance

24-hour integrated WUSTL FRM, µg/m3

0 10 20 30 40 50 60

24-hour integrated IEPA FRM, µg/m

3

0

10

20

30

40

50

60

January 2005 - December 2005 (N=53)

collo precision = 1.3 µg/m3 (8%)

IEPAmean

/ WUSTLmean

= 1.02

24-hour average primary SHARP, µg/m3

0 10 20 30 40 50 60

24-hour average collocated SHARP, µg/m

3

0

10

20

30

40

50

60

April 2005 - January 2006 (N=188)

collo precision = 0.7 µg/m3 (4%)

SHARPcollocate

/ SHARPprimary

= 0.98

FRM SHARP

Comparison of SHARP to FRM

data is highly correlated

SHARP biased ~15%

high with respect to FRM

circled data likely invalid

for FRM based on

comparisons to other data

streams (not shown)

0 10 20 30 40 50 60

0

10

20

30

40

50

60

24 reported hours21-23 reported hours18-20 reported hours

24-hour integrated WUSTL FRM, µg/m3

24-hour average SHARP, µg/m

3

April 2005 – February 2006 (N=276)

collo precision = 2.3 µg/m3 (14%)

SHARPmean / FRMmean = 1.15

SHARP Performance and FEM

Slope and Intercept Acceptance Limits (STL)monthly data sets meet

proposed FEM test

sample sizes

SHARP collocated

precision is acceptable

(6% < 15% criterion)

correlation between

SHARP and FRM is

acceptable (98% > 95%

criterion)

in most cases, SHARP

performance falls outside

the proposed FEM slope

and intercept (versus

FRM) acceptance limits

-5

-4

-3

-2

-1

0

1

2

3

4

5

0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4

regression slope

regression intercept

monthly data (N>22)

all data

Potential Refinements to the SHARP

• Refined moisture control

– Careful examination of 1-minute data shows there potential

issues with the moisture control system

• Cannot adequately respond to rapid RH transients (in part

due to heater duty control algorithm which in STL was a

constant heater duty, not PID)

• Heater duty constraints to prevent overheating might be too

stringent

– any changes to the moisture control system must consider

implications to dynamics of volatile species!

• It is not clear that aerosol water can explain the differences

between the SHARP and FRM in St. Louis

– Same relationship between SHARP and FRM (~15%

difference) observed by Missouri DNR at a separate St. Louis

area site… are the beta absorption properties different for the

St. Louis aerosol compared to the muscovite calibration film?

Thermo SHARP PM2.5 Mass Monitors at

East St. Louis and Arnold

0

10

20

30

40

50

60

70

80

02/28 12:00

02/28 18:00

03/01 00:00

03/01 06:00

03/01 12:00

03/01 18:00

03/02 00:00

03/02 06:00

03/02 12:00

PM m

ass, wind speed

ESL SHARP

Arnold SHARP

ESL wind speed, m/s * 10Small excess at Arnold under

advective conditions at ESL;

often observed with winds

from the east/southeast

Thermo SHARP PM2.5 Mass Monitors at

East St. Louis and Arnold

0

10

20

30

40

50

60

70

80

02/28 12:00

02/28 18:00

03/01 00:00

03/01 06:00

03/01 12:00

03/01 18:00

03/02 00:00

03/02 06:00

03/02 12:00

PM m

ass, wind speed

ESL SHARP

Arnold SHARP

ESL wind speed, m/s * 10

ESL wind speed < 0.5 m/sec

Large jump in fine PM mass at

East St. Louis compared to

Arnold

Summary and Conclusions - I

• Field testing at exhibited:

– New Haven, CT

• Excellent regression slope and intercepts compared to FRM

– East St. Louis, IL

• Excellent durability with very low maintenance requirements

• Excellent collocated precision

• Highly correlated to FRM but biased about 15% high

Summary and Conclusions - II

• Efforts underway to close the gap on SHARP comparison to FRM

– Refinements to moisture control system, examine β absorption

• Very difficult to extrapolate performance in one geographic area to

other areas due to variations in aerosol and environment conditions

– SHARP: East St. Louis, IL versus New Haven, CT (the former is

outside, the latter is in a climate-controlled shelter)

– FDMS TEOM: Bronx and Queens, NYC (Schwab et al. 2006)

– potentially presents a challenge to FEM testing and subsequent

deployment of designated instruments

• A preliminary examination of hourly data streams using matched

instruments suggests we can gain substantial insights into intraurban

variability

– Previous efforts on St. Louis area data using monitors of different

makes and models were challenging to interpret

Acknowledgements

• New Haven field and data support– Paul Norton, Pete Babich and other Connecticut DEP staff

• East St. Louis field and data support– Jason Hill, Eric Ryszkiewicz and other Turner group staff

– Illinois EPA

• Ron Stockett (Missouri DNR) for Arnold, MO data

• Proceedings of the 99th Annual Meeting of the Air & Waste

Management Association, New Orleans, LA, June 20-23, 2006

– K. Goohs, P. Lilienfeld, W. Harmon and J. Wilbertz, “A Synchronized

Hybrid Ambient Real-Time Particulate Mass Monitor”

– J.S. Hill, E.R. Ryszkiewicz, J.R. Turner and K.J. Goohs, “Field

Performance Evaluation of the Thermo Electron Model 5030 SHARP

for Measurement of Ambient PM2.5 Mass Concentration” (paper #533)

For More Information