Incremental Sampling Methodology (ISM) for Metals: Number ...

Incremental Sampling Methodology(ISM) for Metals: Number of Increments and Milling NecessityJay Clausen, Ph.D. Thomas Georgian, Ph.D.

April 29, 2014

Innovative solutions for a safer, better worldBUILDING STRONG®

Soil Sampling Issues• Unrepresentative results using conventional grab soil

sampling at military ranges with energetic & metallic residues

• Poor precision for grabs (e.g., for duplicates)

• Large uncertainty at concentrations near decision limits

• Increasing State regulatory insistence for Incremental Sampling Methodology (ISM) for soils

• USEPA SW-846 work group to update Method 3050B to Method 3050C with ISM. By late 2015?

2


Demonstration Sites Kimama, ID

Small Arms Range21 ISM30 Grab

Fort Wainwright, AKSmall Arms Range

63 ISM52 Grab

Camp Ethan Allen, VTSmall Arms Range

43 ISM36 Grab

Fort Eustis, VASmall Arms Range

27 ISM33 Grab

3


Metallic Residues

4


Incremental Sampling Methodology (ISM)

5


Decision Units (DUs)Volume of soil to be sampled for which a decision needs to be made using the sample results.

Exposure AreasSource AreasSize, shape, and type of DU determined during systematic planning and depend on Data Quality Objectives (DQOs).

6


Incremental Sampling

increment

7


Can ISM samples be split in the field to reduce the mass that needs to be

shipped to and processed in the laboratory?

8


Field Cone-and-Quartered Splits

4321

1200

1000

800

600

400

200

Split

Cu (

mg/

kg)

Boxplot of Cu (mg/kg)

4321

3500

3000

2500

2000

1500

1000

Split

Pb (

mg/

kg)

Boxplot of Pb (mg/kg)

4321

30

25

20

15

10

Split

Sb (

mg/

kg)

Boxplot of Sb (mg/kg)

4321

120

110

100

90

80

70

60

50

Split

Zn (

mg/

kg)

Boxplot of Zn (mg/kg)

9


Background Metals

4321

8.8

8.6

8.4

8.2

8.0

7.8

7.6

7.4

7.2

Split

Co (

mg/

kg)

Boxplot of Co (mg/kg)

4321

40

35

30

25

20

Split

Ba (

mg/

kg)

Boxplot of Ba (mg/kg)

4321

12.0

11.5

11.0

10.5

10.0

Split

Ni (

mg/

kg)

Boxplot of Ni (mg/kg)

4321

16.5

16.0

15.5

15.0

14.5

14.0

Split

V (

mg/

kg)

Boxplot of V (mg/kg)

10


Number of Increments

11


Number of Increments vs. Relative Standard Deviation

12

ISM Percent Relative Standard Deviation (RSD)

m Al Cr Cu Fe Mn Ni Pb Sb V Zn

5 3 10 22 4 4 3 25 25 6 9

10 8 6 162 4 4 4 32 63 5 154

20 27 121 26 22 18 26 30 50 32 15

30 3 7 15 10 4 3 14 15 6 6

50 3 15 21 10 2 4 11 11 6 10

100 3 7 26 4 2 2 17 17 3 15200 6 3 18 4 5 2 4 7 1 11m = number of increments per MI sample


Lead Concentration vs. Number of Increments of ISM Field Replicates

13


Number of Increments (m) for Small Arms Metals (Sb, Pb, Cu, & Zn)

10050301

80000

70000

60000

50000

40000

30000

20000

10000

0

Num be r o f F ie ld I ncr e m e nts

Pb

(m

g/k

g)

10050301

6000

5000

4000

3000

2000

1000

0

Num be r o f F ie ld I ncr e m e nts

Pb

(m

g/k

g)

Boxplot of Pb (m g/kg) Boxplot of Pb (m g/kg)

Number of Increments Number of Increments1 1

0 0

30 3050 50100 100

1000

2000

3000

4000

5000

6000

10000

20000

30000

40000

50000

60000

70000

80000

Lead (mg/kg) Lead (mg/kg)

10050301

90

80

70

60

50

40

30

20

10

0

Number of Field Increments

Sb (

mg/

kg)

Boxplot of Sb (mg/kg)Antimony (mg/kg)

Number of Increments1 30 50 100

10

30

50

20

0

40

60

70

80

90

10050301

900

800

700

600

500

400

300

200

100

0

Number of Field Increments

Cu (

mg/

kg)

Boxplot of Cu (mg/kg)


100

300

500

200

0

400

600

700

800

900

Copper (mg/kg)

10050301

120

110

100

90

80

70

60

50

40

30

Number of Fie ld Increments

Zn (

mg/

kg)



40

60

80

50

30

70

90

100

120

130Zinc (mg/kg)

14


Bootstrap Simulation: Number of Increments per ISM Sample

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 10 20 30 40 50

Boo

tstr

ap M

ean,

Pb

(mg/

kg)

No. Increments (sample size) m

Mean of 48 grab samples = 432 mg/kg

300 bootstrap replicates for each value of m

15


Distributions of Simulated Bootstrap Lead Means

0

50

100

150

200

250

0 100 200 300 400 500 600 700 More

Num

ber o

f Boo

tstr

ap M

eans

Mean of Pb from m Increments (mg/kg)

m = 5m = 7 m = 10

m = 15m = 20

m = 35

m = 25

m = 30

Distribution of means becomes Gaussian & variability decreases as mincreases, consistent with central limit theorem (CTL)

16


Bootstrap Standard Error of Mean Simulation

y 2000 x-1

r² 0.99

0

50

100

150

200

250

300

350

0 50 100 150 200Stan

dard

Err

or o

f Mea

n (m

g/kg

)

Number of Grab Samples or Increments per ISM Sample

Grab Samples

ISM Samples

Power (Grab Samples)

17


To mill or not mill: Whether ‘tis nobler in the mind to not mill and

suffer the slings and arrows of outrageous fortune or to mill to obtain reproducible results…

18


Milled and Unmilled Comparison

Sample #1 Sample #2

19


To Mill or Not To Mill (1)

0

500

1000

1500

2000

2500

3000

3500

4000

0 5 10 15

Con

cent

ratio

n (m

g/kg

)

Replicate

Lead - Unmilled

20



0

500

1000

1500

2000

2500

3000

3500

4000

0 2 4 6 8 10 12 14 16

Con

cent

ratio

n (m

g/kg

)

Replicate

Lead Unmilled

21



22

0

500

1000

1500

2000

2500

3000

3500

4000

0 2 4 6 8 10 12 14 16

Con

cent

ratio

n (m

g/kg

)

Replicate

Pb Unground Pb Ground #1 Pb Ground #2


Digestion Mass for Milled & Unmilled Samples

Mass (g)

Type

10.05.02.01.00.5

Unmi lle

dMille

d

Unm ill e

dM ille

d

Unm ille

dMi lle

d

Unm ille

dMill e

d

Unmille

dMille

d

6000

5000

4000

3000

2000

1000

Pb (

mg/

kg)


23


Probability of One Lead Particle by Aliquot Mass for Unmilled Soil

24


Probability of One Lead Particle by Aliquot Mass for Unmilled Soil

25


Puck Mill Precision

0

10

20

30

40

50

60

70

80

0 30 60 90 120 300

Rel

ativ

e St

anda

rd D

evia

tion

(%)

Milling Interval (seconds)

CuPbSbZn

Acceptance Criteria

26


SAR Metal Concentration by Puck Mill Milling Time

3001209060300

4500

4000

3500

3000

2500

2000

1500

1000

Puc k M ill Gr ind Time (s e c )

Pb (

mg/

kg)

B oxplot of P b (mg/ kg)

3001209060300

1200

1100

1000

900

800

700

600

500

400

300

Puc k M ill Gr ind Time (s e c )

Cu (

mg/

kg)

B oxplot o f C u (mg/ k g)

3001209060300

60

50

40

30

20

10

0

Puck Mill Grind Time (sec)

Sb (

mg/

kg)

Boxplot of Sb (mg/kg)

3001209060300

110

100

90

80

70

60

50

Puck Mill Grind Time (sec)

Zn (

mg/

kg)


27


Ball Mill Precision

0

50

100

150

200

250

300

0 8 12 16 20

Rel

ativ

e St

anda

rd D

evia

tion

(%)

Time (hours)

CuPbSbZn

Acceptance Criteria

28


SAR Metal Concentration by Ball Mill Milling Time

20161280

8000

6000

4000

2000

Ba ll Mill Gr ind T im e (hrs)

Pb

(m

g/k

g)

20161280

60

45

30

15

0


Sb

(m

g/k

g)

20161280

1200

900

600

300


Cu

(m

g/k

g)

20161280

100

90

80

70

60


Zn

(m

g/k

g)

Boxplot of P b (m g/kg) Boxplot of Sb (m g/kg)

Boxplot of C u (m g/kg) Boxplot of Zn (m g/kg)

29


Milling by Device Type

Unmilled #2Unmilled #1Puck & RingPuck #2Puck #1MPBall Mill

5000

4000

3000

2000

1000

Grinder Type

Pb (

mg/

kg)


30


Performance AssessmentActivity Yes NoISM, 30+ increments √Grab Sampling √Field Splitting √Sieving √Milling necessity √Increased digestion mass √ √Increased digestion time √Subsampling √

31


No ISM-Lite

ISM

32


Cost ComparisonGrab ISM

Systematic Project Planning $ $Site Preparation $$ $Field Sampling $$ $$$Laboratory Sample Processing $ $$$Analysis $$ $Data Validation $$ $Per Sample Cost $-$$ $$-$$$Number of Samples ### #Total Project Cost $$-$$$ $-$$

33


Conclusions Field splitting not recommended when metal particles

expected Minimum of 30-increments per ISM sample Number of increments is a function of expected degree

of heterogeneity Larger digestion aliquot in lieu of milling does not work Milling is necessary when metal particles are present Marginal improvements in data quality with larger

digestion aliquots for milled samples Recommend 5 minutes milling with Puck and 8 hours

with Roller Mills

34


References Clausen. 2015. Sampling of Soils with Metallic Residues Collected

from Military Small-Arms Ranges. PhD Dissertation. University of New Hampshire.

Clausen et al. 2013. Cost and Performance of Incremental Sampling Methodology (ISM) for Metallic Residues, ESTCP Project ER200918.ERDC/CRREL TR-13-10. http://acwc.sdp.sirsi.net/client/search/asset/1030100

Clausen et al. 2013. Demonstration of Incremental Sampling Methodology for Soil Containing Metallic Residues. ERDC/CRREL TR-13-9. http://acwc.sdp.sirsi.net/client/search/asset/1030080

Clausen et al. 2013. Incremental Sampling Methodology (ISM) for Metallic Residues. ERDC/CRREL TR-13-5. http://acwc.sdp.sirsi.net/client/search/asset/1029240

Clausen et al. 2012. Evaluation of Sampling and Sample Preparation Modifications for Soil Containing Metal Residues. ERDC TR–12-1. http://acwc.sdp.sirsi.net/client/search/asset:asset?t:ac=$N/1006020

35


References (Continued)• Hewitt et al. 2012. EPA Federal Facilities Forum Issue Paper: Site

Characterization for Munitions Constituents. EPA-505-S-11-01. http://www.epa.gov/fedfac/pdf/site_characterization_for_munitions_constituents.pdf

• Clausen, J. et al. 2012. Metal Residue Deposition from Military Pyrotechnic Devices and Field Sampling Guidance. ADA562327. http://handle.dtic.mil/100.2/ADA562327

• Clausen, J. and N. Korte. 2009. The distribution of metals in soils and pore water at 3 U.S. military training facilities. Soil Sed Cont J.18(5):546-563. http://www.tandfonline.com/doi/abs/10.1080/15320380903085683

36

Incremental Sampling Methodology (ISM) for Metals: Number ...

Documents

Transcript of Incremental Sampling Methodology (ISM) for Metals: Number ...