Silica analysis: XRD and FTIR · by direct-on-filter XRD and recent findings from workplace samples Martin Mazereeuw, John Volpato, Akemi Ichikawa TestSafe Australia –SafeWork NSW

15/12/2019

1

State-of-the-art analysis of respirable crystalline silica by direct-on-filter XRD and recent findings from workplace samples

Martin Mazereeuw, John Volpato, Akemi IchikawaTestSafe Australia – SafeWork NSW

3/12/2019 2

Silica analysis: XRD and FTIR

XRay Signal

Lamp Detector

Filter

Lamp Detector

IR Signal

XRD and FTIR used for RCS analysis

➢ Direct-on-filter approach

➢ Different technology

➢ Different performance

Diffraction

technique

Transmission

technique

1

2

15/12/2019

2

3/12/2019 3

XRD Analysis

Method: NH&MRC (1984), HSE(2014)

Response factor: Peak intensity of Q (101) diffraction

Criteria: Q(101)/average*= 90-110%, Dust<2mg

Q (100), (112) when Q(101) interfered

Overload correction implemented when overloadedQ

(10

0)

Q(1

12)

Q(1

01)

Ag(1

11)

Ag(2

00)

Measurement area

(24mm diameter)

*average= {Q(100)+Q(101)+Q(112)}/3

3/12/2019 4

FT-IR Analysis

Method: HSE(2014)

Blank subtraction implemented

Response factor: Absorption peak height of Si-O vibration

Criteria: PH800/PH780 = 1-1.4, Dust<1mg

PV

C

Measurement area

(8mm diameter)

Q 8

00

Q 7

80

*

PV

C

PV

C

PV

C

PV

C

PH 800 PH 780

3

4

15/12/2019

3

3/12/2019 5

Equipment

XRD

3/12/2019 6

a-quartz calibration curves

XRD r2=0.998 FT-IR r2=0.996

pure quartz = linear result

5

6

15/12/2019

4

3/12/2019 7

XRD vs FT-IR (α-quartz samples)

r2=0.991n=43(valid results)

Pure α-quartz (No matrix)

Good agreement

r2=0.99

FT-IR values: XRD values

y=x

y=0.8x

y=1.2x

3/12/2019 8

Real workplace samples

n=253

Sampled 34 different workplaces

in Australia (2014-18)

Industry:

- Road construction/ Tunneling (47%)

- Coal mining (23%)

- Kitchen benchtop (25%)

- Others (5%)

Compositions depend on

individual samples

7

8

15/12/2019

5

3/12/2019 9

XRD vs FT-IR (real workplace samples)

n=253(all measured

results: Included

invalid results)

y=x

y=0.8x

y=1.2x

Outliers

3/12/2019 10


n=171(valid results)

y=x

y=0.8x

y=1.2x

32% of FT-IR data failed

(PH criteria, Dust >1mg)

r2=0.97

FT-IR values: ~10% higher

than XRD

9

10

15/12/2019

6

3/12/2019 11

FT-IR Spectra (α-quartz vs Silicates)

Q(7

80

)

Q(8

00

)

Peak overlap on Quartz (800) Blank subtracted

3/12/2019 12

XRD Spectra (α-quartz vs Silicates)

Q(1

00

)

Q(1

12

)

Ag(1

11)

Q(1

01

)

Ag(2

00)

No major peak overlap on Quartz (101)

11

12

15/12/2019

7

3/12/2019 13

Recovery of α-quartz (with Silicates)

FT-IR showed positive bias

w/ Kaolinite (>60%),

Albite (>90%)

Cristobalite(>10%)

3/12/2019 14

Cristobalite

α-Quartz Tridymite Cristobalite

T > 800 ºC T > 1100 ºC

Quartz transforms to Cristobalite (Tridymite) by heat process

>~800degC and influenced by alkalinity.

Pure Quartz transforms to Cristobalite at ~1400degC.

Quartz most commonly present (stones, soils).

Cristobalite (Tridymite) presence in nature related to volcanic activity.

Cristobalite (Tridymite) presence in engineered stones.

13

14

15/12/2019

8

3/12/2019 15

Cristobalite

Engineered stone

Cristobalite found regularly

in dust samples

Sometimes at high %

10-15 % of cristobalite found in all jobs offered.

3/12/2019 16

SDS – engineering stones

15

16

15/12/2019

9

3/12/2019 17

Quartz-Cristobalite

3/12/2019 18

Estimated LOD

For “Direct-on-filter” using XRD with old instrument

Major limitation: Signal to Noise Ratio (SNR)

XRD can achieve

2 µg LODwhen needed

17

18

15/12/2019

10

3/12/2019 19

Estimated LOD

For “Direct-on-filter” using XRD with new instrument

XRD can achieve

800 ng LODwhen needed

3/12/2019 20

Blanks

Lowering exposure limit Lowering reporting limit

Analysis usually done

against a blank signal

When lowering detection limits,

quality of blank becomes important

Found: Not all blanks are blank

➢ Contamination control

➢ Blank variability

19

20

15/12/2019

11

3/12/2019 21

Transfer/ Handling the sample

Do not use flat plastic bag Do not tape the sample

Dust came off from the filter

Dust came off when tape removed/

Interference from tape

3/12/2019 22

Recommended casePlace the sample and clip the edge,

collected surface (to be analysed) should be upside

Label on the top

Label on the top

21

22

15/12/2019

12

3/12/2019 23

Chemical Analysis / TestSafe Australia

Level2, Bldg1, 9-15 Chilvers Rd,

Thornleigh, NSW 2120

P: 02 9473 4000

E: [email protected]

http://www.testsafe.com. au

Booth 48

3/12/2019 24

Questions/ Discussions

23

24

mailto:[email protected]

15/12/2019

13

3/12/2019 25

Conclusion

Comparison between “Direct-on-filter” using XRD and

“Direct-on-filter” using FT-IR

“Direct-on-filter” using XRD showed better results:

1) Less interference from matrixes found in Australian

workplace samples.

2) 32% of workplace samples failed by FT-IR for valid

analysis

3) XRD could handle up to twice the sample loading

(2 mg) and could correct for overloading

4) XRD can achieve lower LOD

3/12/2019 26

Overloaded samples (Dust:1~3mg)- Pure α-quartz samples

- Simulated samples

XRD can handle

overloaded samples

25

26

15/12/2019

14

3/12/2019 27

Crystalline silica

Chemical name CAS No.Current TWA

(mg/m3)

Quartz 14808-60-7 0.1Stone, cement,

soil

Cristobalite 14464-46-1 0.1

Engineering

stones, heated

condition

(>800degC)

Tridymite 15468-32-3 0.1heated condition

(>800degC)

3/12/2019 28

Quartz-Cristobalite-Tridymite

Quartz present in nature (stones, soils).

Cristobalite (Tridymite) presence in nature at volcanos.

Cristobalite (Tridymite) presence in engineer stones.

Quartz transforms to Cristobalite (Tridymite) by heat

process >~800degC with Alkaline elements (Li, Na, Mg,

Al, K, Ca) .

Pure Quartz transforms to Cristobalite at ~1400degC.

27

28

15/12/2019

15

3/12/2019 29

Respirable crystalline silicaSamples from stone manufacturing

3/12/2019 30

Transfer/ Handring the sample

Do not use flat plastic bag Do not tape the sample

29

30

15/12/2019

16

3/12/2019 31

Discussion: Possibility to achieve lower LOD

For “Direct-on-filter” using XRD

Major limitation: Signal to Noise Ratio (SNR)

XRD can achieve

lower LODWhen analyse with

slower scan or better

detector

323/12/2018

Respirable crystalline silica analysis

Method

Quick

analys

is

Precision(pure QTZ)

Interference

from matrixSensitivity

Effect of

PVC film

Effect of

Fe

“Direct on filter”

using FI-IR


using XRD

“Indirect-KBr”

using FT-IR

“Indirect-filter”

using FT-IR -

“Indirect” using

XRD - -: Good or Not affected : Not good or Affected

Less likely

(Fe-oxide>90%)

31

32

15/12/2019

17

3/12/2019 33

XRD Spectra (α-quartz vs non Silicates)Q

(10

0)

Q(1

12

)

Ag(1

11)

Q(1

01

)

Ag(2

00)

No peak overlap on Quartz (101) Noise increased slightly

3/12/2019 34

FT-IR Spectra (α-quartz vs non Silicates)

Q(7

80

)

Q(8

00

)

No peak

Broad absorption

For calcite& graphite, No peak but absorption observed at Quartz (800)

No peak at 800

Broad absorption

Blank subtracted

33

34

15/12/2019

18

3/12/2019 35

Recovery of α-quartz (with non Silicates)

FT-IR showed negative bias

w/ Calcite (>95%)

XRD showed negative bias

w/ Fe-oxide (>90%)

3/12/2019 36



A

y=x

y=0.8x

y=1.2x



r2=0.97


than XRD

Z

B

C

35

36

15/12/2019

19

3/12/2019 37

Spectra (Sample A) FT-IR = XRD

Q(1

00

)

Q(1

12)

Ag(1

11)

Q 8

00

Q(1

01

)

Ag(2

00)

No significant matrix

detected

α-quartz = 0.17mg

Blank subtracted

α-quartz: 82wt%, Matrix: 18wt%

Dust :0.21mg

No significant matrix

detected

I800/I780 = 1.2

α-quartz = 0.17 mg

3/12/2019 38

Spectra (Sample B) FT-IR > XRD

Q(1

00

)

Q(1

12)

Q 8

00

Kaolinite

Q(1

01

)

Ag(1

11)

Ag(2

00)

Blank subtracted

No interference

α-quartz = 0.26mg

Fe-oxide: 4wt% (XRF)


Dust: <0.72mg

Kaolinite detected

Al4Si4O10(OH)8

I800/I780 = 1.4

α-quartz = 0.32 mg

37

38

15/12/2019

20

3/12/2019 39

Q(1

00

)

Q(1

12)

Ag(1

11)

Q 8

00 Albite

Q(1

01

)

Ag(2

00)

Spectra (Sample C) FT-IR > XRD

Blank subtracted

Q(1

12) No interference

α-quartz = 0.14mg

Fe-oxide: 5wt% (XRF)


Dust: <0.88mg

Albite detected

NaAlSi3O8

I800/I780 = 1.0

α-quartz = 0.32 mg

3/12/2019 40

Spectra (Sample Z) - FT-IR data failed

Q(1

00

)

Q(1

12)

Q 8

00Calcite

Q(1

01

)

Ag(1

11)

Ag(2

00)

Blank subtracted

No interference

α-quartz = 0.20mg


Dust: 1.97mg

Quartz peaks deformed

Calcite detecte

Silicates detected

I800/I780 = 1.6

α-quartz = 0.28 mg

39

40

15/12/2019

21

3/12/2019 41

Simulated respirable dust samples

- a-Quartz (0-100%) & common Matrix in Australian workplaces

(Known α-quartz & matrix, pre-weighted)

Silicates (Kaolinite: Al4Si4O10(OH)8,

Albite: NaAlSi3O8,

Cristobalite: SiO2 )

Non silicates (Calcite: CaCO3, Graphite: C)

Fe-oxide (Fe2O3)

=> Measured simulated samples

=> Calculated from spectrum for matrix

3/12/2019 42

a-quartz calibration curves

XRD r2=0.998

Linear to ZERO

41

42

15/12/2019

22

433/12/2018

Standard deviation of α-quartz samples

LOD

3/12/2019 44

Respirable dust samples

1) Pure α-quartz samples (n=43), blanks (PVC filters)[Prepared with NIST SRM 1878a, A9950, WASP, LGC samples]

- Calibration curves

- Standard deviation/ LOD

- Relationship between both methods

2) Real workplace samples (n=253)

- Relationship between both methods

- Spectra analysis

3) Simulated samples (α-quartz with known matrix)

- Interference form Matrix

- Fe-effect

43

44

15/12/2019

23

3/12/2019 45

PVC filters (Blanks)

- Suitable for gravimetry

(low moisture, light weight)

- Low cost

- Durable to water, acids, metals, oils

- Large differences : 4-8mg (25mm filter)

25mm filter weight

GLA-5000

Filter weightFT-IR Spectra of blanks

*before blank subtraction

Differences

463/12/2018


MethodPreparation for

specimen

film/pellet when

analised

Analysis

method

No. of

labs


using FI-IR Not requiredPVC filter FT-IR 9(28%)


using XRDPVC filter XRD 12(38%)

“Indirect-KBr”

using FT-IR Ashing/

Acid treatment/

Alkaline

treatment

if needed

KBr pellet FT-IR 7(22%)


using FT-IR

Re-deposited

(PVC) filterFT-IR 1(3%)


XRD

Re-deposited

(PVC) filterXRD 1(3%)

Other - - - 2(6%)

LGC (2017) AIR PT Scheme Report (Round 19) 2017

45

46

15/12/2019

24

473/12/2018


Method

Quick

analys

is

Precision(pure QTZ)

Interference

from matrixSensitivity

Effect of

PVC film

Effect of

Fe


using FI-IR ? ? ? ?


using XRD ? ? ? ?

“Indirect-KBr”

using FT-IR


using FT-IR -


XRD - -: Good or Not affected : Not good or Affected

3/12/2019 48

Questions/ Discussions

47

48

15/12/2019

25

3/12/2019 49

Crystalline silica information

http://www.safework.nsw.gov.au/silica

https://www.safework.nsw.gov.au/hazards-a-z/hazardous-

chemical

3/12/2019 50

Indirect method with FT-IR (3mm-KBr)

Ojima (2003)

Determining of Crystalline

Silica in Respirable Dust

Samples by Infrared

Spectrophotometry in the

Presence of Interferences

J Occup Health 2003; 45

94-103

Measurable range:

7-37 µg

49

50

http://www.safework.nsw.gov.au/silica

15/12/2019

26

3/12/2019 51

Indirect method with FT-IR

NMAM 7603

Quartz in Respirable Coal

Mine Dust, by IR

(Redeposition)

Measurable range:

10-500 µg

3/12/2019 52

Overload correction* (XRD)

*Overload correction by using

Ag-sheet

Altree-Williams (1977)

Quantitative X-ray

diffractometry on respirable

dust collected on nuclepore

filters

Ann Occup Hyg 1977; 20

109-126.

Able to handle

more than 2mg

51

52

15/12/2019

27

3/12/2019 53

Bulk vs filter samples (XRD)

X-ray

Sample

Fe

X-ray

Bulk Filter

Absorbed Transmitted

3/12/2019 54

Simulated samples (Kaolinite)

FT-IR

XRD

Q(7

80

) Q

(80

0)

Acceptable range

Blank subtracted

53

54

15/12/2019

28

3/12/2019 55

Simulated samples (Cristobalite)

FT-IR

XRD

Q(7

80

) Q

(80

0)

Acceptable range

Acceptable range

Blank subtracted

3/12/2019 56

Fe-effect (bulk samples, XRD)

Increase of

background noise

Decrease of peak intensity – Decrease of quartz amount

Increase of mass absorption

Fe effect

60% Fe-oxide

No Fe-oxide

55

56

15/12/2019

29

3/12/2019 57

Fe-effect (filter samples, XRD)

Increase of background noise

Increase of mass absorption

Fe effect

60% Fe-oxide

No Fe-oxide

Workplace samples: 0-6%

Fe-effect negligible

3/12/2019 58

Respirable dust samples

Pure α-quartz samples (n=43)

- NIST SRM1878a, A9950

deposited onto a PVC filter (GLA5000) from a dust generator

through a cyclone sampler

- PT samples (WASP HSL, LGC UK)

57

58

15/12/2019

30

593/12/2018

Standard deviation of blank filters (LOD)

ConditionXRD

(µg/filter)

FT-IR

(µg/filter)

Repeatability

(within day)

10 measurements of same

blank filter2.5 0.030

Intermediate precision

(between days)

10 measurements of same

blank filter over 3 months3.1 0.34

Variability between

filters

10 different blank filters2.8 3.45

Estimated LOD

from blanks (3 x s) 10 10

Major limitation to

determine standard

deviations

Signal to

noise ratio

(SNR)

Absorption &

difference of

blank filter

3/12/2019 60

XRD Spectra (α-quartz vs Silicates)

Q(1

00

)

Q(1

12

)

Ag(1

11)

Q(1

01

)

Ag(2

00)

No peak on Quartz (101)

59

60

15/12/2019

31

3/12/2019 61

XRD Spectra (α-quartz vs Polymorphism)Q

(10

0)

Q(1

12

)

Ag(1

11)

Q(1

01

)

Ag(2

00)

No peak on Quartz (101)

3/12/2019 62

FT-IR Spectra (α-quartz vs Silicates)

Q(7

80

)

Q(8

00

)

Peak observed on Quartz (800) Blank subtracted

61

62

15/12/2019

32

3/12/2019 63

FT-IR Spectra (α-quartz vs Polymorphism)

Q(7

80

)

Q(8

00

)

Strong peak observed on Quartz (800) Blank subtracted

3/12/2019 64



A

y=x

y=0.8x

y=1.2x



Almost all outlier excluded

r2=0.97


than XRDY

63

64

15/12/2019

33

3/12/2019 65

Spectra (Sample Y) - FT-IR data excluded

- I800/I780 = 18

- Calcite detected

- Several silicates

detected

- Matrix/Dust :97 wt%

Q(1

00

)

Q(1

12)

Q 8

00

Calcite

Q(1

01

)

Ag(1

11)

Ag(2

00)

65

Silica analysis: XRD and FTIR · by direct-on-filter XRD and recent findings from workplace samples Martin Mazereeuw, John Volpato, Akemi Ichikawa TestSafe Australia –SafeWork NSW

Documents