Direct Meassurement With XRF-Sampling but No Sample Preparation

Direct measurement with XRF –

sampling but no sample preparation?

Bertil Magnusson

Content

an overview of direct XRF measurement will be

presented using standard wavelength dispersive and

energy dispersive instruments as well as handheld

instrument

• What we all know about XRF

• Direct measurement versus normal analysis in laboratory

• Elemental range

• Systematic effects

• Application of direct measurements

1. Quality control in nuclear power plant

2. Analysis of sediment and wood cores

3. Determination of thickness of coating

Direct measurement with XRF 2

We all know that XRF

can be regarded as a non-destructive

technique

but in most cases in order to get more

reliable results with a lower

measurement uncertainty a sample

preparation step is performed for solid

samples. The preparation can e.g. be

homogenizing and subsequent

briquetting or fusion into a bead.

For liquid samples with no particles direct

measurement is possible with reliable

results


We also know that XRF analysis….

can be performed directly on many samples

with additional benefits such as

• less sample handling

• no sample transport to laboratory

• better spatial resolution

• lower response time

• higher sample throughput


But in most cases with a

drawback of much higher

measurement uncertainty

Direct measurement versus ”normal” analysis

”normal” analysis

Direct

measurement

”Normal” analysis

Atmosphere Air/He Vacuum or

helium

Sample

preparation

Normally not Yes

Sample size Few restrictions < 5 cm

Spatial resolution Yes No


Drawbacks with measurement in air

and no sample preparation…

Today light elements possible in air with short

pathlengt – example handheld XRF

6 Direct measurement with XRF

XRF measurement with no sample preparation

Measuring depth is dependent on energy of analytical line and

mean atomic number in the matrix. With prepared sample that

are homogeneous - no systematic effect


Direct measurement without sample preparation:

Systematic effects for inhomogeneous samples!

Sn Ka

Zn Ka

S Ka

Si Ka

Na Ka

Measuring

depth

Measurement cycle

Client Issue

Client

Define issue

Decision on

measurement

Client

Sampling

Analysis

Measuring scientist

Interface Report on

measurement

Decision on result

Evaluation

Data presentation

Measuring

scientist


Can i sell this

product?

Does the

product fulfil

the

regulations?

Direct measurement - building materials


Direct measurement – consumer goods


Metals in

• Plastics

• toys EU Toy directive (last March 2012)

• Presence of flame retardants e.g Br, Cl, Sb

• Electronics – ROHS directive Cd, Cr(VI), Ni, Hg, Pb

and org Br

• Jewellery e.g. Ni, Cd, Cr, Pb Co

• Textiles

• Isolation material

• Impregnated wood e.g As. Cr

Today these tests are performed routinely at SP

Applications presented here

• Quality control in nuclear power plant -

Analysis of deposits (crud) on nuclear

fuel rods – sampling on a filter.

• Analysis of sediment and wood cores

– XRF footprint along the core

• Determination of thickness of coating

on welding wire – direct measurement


Measurement cycle – nuclear power plant

Client Issue

Client

Define issue

Decision on

measurement

Client

Sampling

Analysis

Measuring scientist

Interface Report on

measurement

Decision on result

Evaluation

Data presentation

Measuring

scientist


Corrosion situation

and…..?

Study crud

deposits

Quality control – analysis of deposits, crud

• Crud sampling

• Direct measurement with XRF

• XRF analysis on active samples

• Validation with synthetic samples and active

test samples


Crud* sampling

The crud is scraped or brushed from a defined

area

of the fuel rod

The crud is collected on a filter membrane

The crud/filter sample is analysed to determine:

• Quantity (g/m2) - XRF

• Elemental composition (g/m2): XRF

Fe, Co, Cr, Mn, Ni, Cu, Zn, Pt Sampling from fuel rod?- measurement of Zr

• Gamma activity of radioactive nuclides: Co-

60 …

*deposits on the nuclear rod

This application is presented

courtesy to

Eva Fredriksson & Pentti

Hietala, Westinghouse

Zircalloy

ZrO2

crud

Crud* sampling

SP 1

SP 2

SP 3

SP 4

SP 5

SP 6

2

3

4

1

BC

AD

A B C D E F GH I J

876

910

4321

5

4

1

2

3

A

D

B

Marking

Marking

2

3

4

1Channel Center

Side OrientationX

Subbundle OrientationX

SP Spacer

C

Sampling

level

Fuel crud is

sampled between

the spacers -

normally 5 or 6

levels

Water reference

samples are taken

both before and

after each

assembly

3

2

5

4

1

6

7 Top plate

Spacer 1

Spacer 2

Spacer 3

Spacer 4

Spacer 5

Spacer 6

HOW


*deposits on the nuclear rod

Crud sampling

HOW

The fuel

crud

sampling

equipment

Crud sampling process in a nuclear power plant

1. Sampling under water on a filter (Millipore)

2. Removing water by pressing air throught

the filter

3. Removal with tool the upper part of filter

holder

4. Mounting av cup with mylar film on top of

filter

5. XRF-analysis using either

• Uniquant semiquantitative software

or

• a calibration based on active standards

Analysis of crud in a nuclear power plant

XRF-

container

ARL Optim’X in a container

XRF measurement on radioactive samples

Issues with radioactive samples

The radioactivity of the sample can lead to:

• Sample excitation resulting in XRF signal for

several elements.

• Here we test X-ray tube off

• Higher background – especially for lighter

elements

• Samples in sample changer increasing

background for measured sample


Sample excitation from active samples

ED-XRF with X-ray tube on&off?

Spectrum presented courtesy to Kari Moum and Torill Solheim

Institutt for Energiteknikk/Halden Reaktor Prosjektet


Tube off

Tube on

Sample is a filter.

Sampling 10 l water

from primary cooling

circuit in a nuclear

power plant. Activity 25

µS/cm close

ED-XRF - tube off

we see Cr Mn Fe Ni but no Co Cu and Zn

Cr Mn Fe Co Ni Cu Zn


Tube off

Tube on

Iron

background

≈ 2 µg/cm2

WD-XRF with X-ray tube off&on


Tube off Samples are crud from a nuclear

power plant

Tube on

NOTE:

log

scale

WD-XRF with X-ray tube off


Tube off Samples are crud from a nuclear

power plant

Tube on

Cr Ka Mn Ka

Green is ca

30 ug/cm2

Background increase with more crud


Red 20 mg crud or 5 mg/cm2

Blue 6 mg crud or 1.5 mg/cm2

The background mystery

2009 was the first time we used XRF an a nuclear powr plant

For a water blank sample we got very much higher background

especially for Al & Si.

A scan showed low background???


New scan of water blank with samples and

without samples in sample changer


With no samples in sample changer less than 0.3 kcps

No radioactive samples in sample changer

Validation

Instrument calibrated with standardless software

1. Validation with synthetic crud samples prepared from pure

metals – Low ”recovery” - < 80 %

2. Validation with synthetic crud samples grinded further –

”recovery ” close to 100 %

3. Validation with active samples analysed with ICP-MS –

”recovery” close to 100 % for most elements


Validation with synthetic crud samples

Fe, Co, Cr, Mn, Ni, Cu, Zn, Pt


Figure 1 Particle size distribution for crud 1. Surface weighted mean 3 µm

Figure 2 Particle size distribution for crud 1 after sample was further grinded. Surface weighted mean 1.2 µm

Low Recovery < 80 %

Recovery close to 100 %

1 µm

Particle

size

distribution

Validation Fe up to 1 000 µg or 250 µg/cm2

XRF vs ICP-MS on active samples

2010-08-17 30 crud samples

Validation Zn up to 600 µg or 150 µg/cm2

– XRF vs synthetic crud


Residuals

Validation Zn up to 100 µg or 25 µg/cm2

XRF vs ICP-MS on active samples year 2009


-50

0

50

100

150

0 20 40 60 80 100 120Res

iid

ua

l u

g

Similar results were obtained 2010 –

Decision to use active crud samples for calibration

Conclusions

Systematic effects – yes.

• Validation with synthetic standards - The sample

particles were finer than the synthetic crud standards

(before further grinding)

• Validation with active test samples

• For most elements the ”recovery” was close to 100

% comparing XRF Uniquant with ICP-MS results

• The Zn gave 150 % ”recovery” when comparing

with ICP-MS. No understanding of this effect so

far. The ICP-MS were validated with synesthetic

crud and gave excellent results


Conclusions

• New calibration performed with active crud samples,

analysed with ICP-MS, to minimise systematic

effects.

• For elements not found in crud synthetic samples

were used.

• Interference from active samples

• Increased background with radioactive samples.

• No characteristic peaks found with WD-XRF

• Fitness for intended purpose


Direct analysis now possible for

elemental composition and amount

using XRF for elements of interest

Measurement cycle – Increasing depth in harbour area

Client Issue

Client

Define issue

Decision on

measurement

Client

Sampling

Analysis

Measuring scientist

Interface Report on

measurement

Decision on result

Evaluation

Data presentation

Measuring

scientist


Can we dredge in

the harbour area?

Will there be

an

environmental

impact?

Analysis of sediment and wood cores

Shale (norsk: skifer) sample was scanned with steps of 0.2

millimeters at 10 seconds using XRF. On top of the optical image is

superimposed Si and Fe graphs.

Si (blue) Fe (red)

50 mm

Applications

This

application is

presented

courtesy to

Anders

Rindby


Combined m-XRF, radiography and optical scanning

The Core Scanner moves the sample in steps.

Sample is irradiated with an X-ray tube with a

well-defined beam. The fluorescence (XRF) and

absorption (radiography) are measured.

An camera system provides sample images.


Combined m-XRF, radiography and optical scanning

XRF

detector

profiler

Camera

Sample


K profile Ti profile Fe profile

As profile

Br profile

Average atomic number

Sample with courtesy of Dr. Bernard Dennielou, Ifremer, France

820 mm

XRF scan examples


Illustrations/graphs area

Determination of thickness of coating on welding wire

This application is presented

courtesy to

Göran Säwemark, ESAB


Client Issue

Client

Define issue

Decision on

measurement

Client

Sampling

Analysis

Measuring scientist

Interface Report on

measurement

Decision on result

Evaluation

Data presentation

Measuring

scientist

Client need

reporting in

mg/dm2

Production

need direct

reporting


Application of surface additives

Lubricant

•MoS2 - thickness on wire 0.001-0.01µm ( up to 0,3 mg/dm2)

•Ca (soap) 0,07 mg/dm2

Wet Chemical Analysis

Analysis of Mo and Ca

4 m sample is leached in acetone in an ultrasonic bath to remove

the MoS2 powder.

The acetone is evaporated and the MoS2 is dissolved in aqua regia

and Mo is determined by ICP.




Sample Preparation of wires for XRF

Sample holder on the insert Sample ready for XRF-analysis

Opening: 27 mm

Analysed wire: 0.4-0.5 m

Wire diameter: 1.0 -1.2 mm


Mo and S scan WD-XRF Intensity versus 2


Calibration curve Mo Lb Intensity versus Mo (mg/dm2)

Mo (mg/dm2) measured with ICP


Calibration curve Ca Intensity versus Ca (mg/dm2)


ED-XRF


Mo and S scan ED-XRF Intensity versus energy (channel number) 1.6 to 7.5 keV

We measure S k alfa + partly Mo L alfa


Handheld XRF for screening

Using a general metal calibration

The apparent sulfur concentration

is used for screening if the

product fulfils the specification


Production control before respooling

Conclusions - Application welding wire

At factory no wet chemical analysis

XRF is a fast tool for deposited amount on the surface of MoS2

and Ca on welding wires

For accurate analysis: Separate calibration curves / diameter

• WDXRF: Can separate Mo and S

• EDXRF: Analysis of S (Mo overlap)

• Handheld XRF: Analysis of S (Mo overlap)

Fast production control on line Pass/Fail


Direct measurement with XRF –

sampling but no sample preparation?

YES

when we know our analytical

technique

&

when we can demonstrate fitness for

intended purpose

Bertil Magnusson

And I also would like to pay homage to Wilhelm

Conrad Röntgen


Documents

Direct Meassurement With XRF-Sampling but No Sample Preparation