15June2016

You can download a PDF version from the link at the foot of this page (use the blue down arrow on the right) or view it in a new page here... (you might find it easier to read.)

XRF Meeting Report 15^th June 2016

A joint BCA/ RSC Atomic Spectroscopy Group Meeting at University of Leicester.

Group Photograph.

Overview of Event

The XRF group is a sub-committee of the British Crystallographic Association (BCA) which covers a range of instrument techniques and interest areas. Further details can be found at http://www.crystallography.org.uk/

Annual meetings are held by each sub-committee which gather together industrial users, academics and instrument vendors in one place. This year’s XRF meeting was held at the University of Leicester in collaboration with the Royal Society of Chemistry and was attended by 47 delegates along with 10 vendor stands presenting their products.

Morning Sessions.

Session 1.

The welcome to the BCA XRF meeting of 2016 at Leicester University was given by Nick Marsh.

The chair of the session, David Beveridge, then introduced Professor Andrew Shortland who gave a fascinating talk on XRF analysis of historical glass and glazes. After a short introduction to Cranfield Forensic Institute, Professor Shortland described the work performed on Meissen porcelain glazes. Differentiation between visually similar items of porcelain is critically important to auction houses and collectors. Provenance dramatically effects the value of porcelain, reducing the £100K price tag of a genuine 18^thC piece to just £1000 for a fake. Later 19^thC reproductions produced by the Meissen factory from original moulds and later repairs or conservation of original 18^thC pieces complicate the situation further. A bench top XRF instrument was used to non-destructively analyse over 300 original objects, and 800 fragmented pieces of known origin. This created a database and cluster plots were generated to characterise early porcelain. The elemental composition of enamel and gilding also allowed for differentiation between restored and original pieces accurate to within a 10-30 year timeframe for the Meissen factory. Professor Shortland went on to describe the in-situ HH-XRF analysis of a medieval glass window at Christ Church Cathedral, Oxford. Using early fragments of 17^thC glass found in a coal hole (!) for comparison and balancing precariously on a ladder (!!) to do the measurements, the entire window was mapped for 12 elements. Using compositional groupings, the results showed that much of the original window had been taken down but later restored by a contemporary artist (and an English Civil War document supported this finding) while Victorian and modern glass inclusions were also identified.

The exhibitor’s forum followed with representatives giving 3 minute overviews of their latest instruments, innovations and products:

SciMed (Paul Vanden Branden)

· Introduced the Rigaku Nex-DE ED-XRF with 60kV tube and SDD high speed detector; the WD-XRF Primus and Supermini 200 bench top XRF. The Hitachi SE1000 large capacity mapping spectrometer was also mentioned.

Spectro (Graham Hibberd)

· Presented the latest Spectro XEPOS benchtop ED-XRF; this direct excitation instrument features a bandpass filter and new detector to handle high countrates

Niton UK (Ken Granger)

· A live real-time demonstration of HH-XRF described the features and applications of the different Niton models available.

Spex Europe (Dave Speake)

· XRF sample preparation equipment (mills, grinders and cryogenic mills), pellet presses and the latest 1, 3 and 6 place Katanax electric fusion machine were shown. A fusion method development service is available in addition to binders, fluxes and ICP liquid standards. The presentation closed with a tribute to Fernand Claisse (1923-2016).

Bruker (Simon Bailey)

· A new WD-XRF crystal, the XS-400 was described. The performance of this surface-treated crystal was compared to the LiF200 and LiF220; it gives good signal-to-noise, increasing sensitivity for transition metals. A lab report on this new crystal is to be released shortly.

Specac (Allan Finlay)

· Presses and dies for XRF pressed pellet preparation were featured. Manual, electrical, auto-touch and hand press models are available; some with evacuable dies. Heated presses are also available and suitable for producing thin films from polymers.

PANalytical (Michael Brogan)

· The new Zetium WD-XRF was presented; different Zetium models feature a Theta free lime channel, small spot mapping and WD/ED core combinations. In 2014, Claisse merged with PANalytical bringing in-house fusion systems and fluxes into the PANalytical portfolio.

Shimadzu (Lee Parry)

· The product range of HPLC, GC-MS and LC-MS was described. Shimadzu offer applications, technical support and FOC training in conjunction with the RSC. The EDX-7000 and 8000 bench top instruments were featured.

Datech Scientific Limited (Adam Housley)

· An extensive range of XRF sample preparation equipment (Herzog) and automated lab solutions (Nucomat) were presented. New additions to the portfolio include LIBS and LA-ICP-MS instrumentation from Applied Spectra.

XRF Scientific (Frederic Davidts)

· Socachim fluxes, platinumware and the Phoenix II cold-cold fusion machine were featured. XRF sample preparation equipment for crushing, milling and pressing are available. The xrWeigh was presented; this carousel can weigh up to 30 vials of flux in 20 minutes with one-touch operation.

Overall, the supplier’s representatives kept to their allotted 3-minutes slots, much to relief of the chair. The presentations were very informative and welcomed by the delegates who were then encouraged to visit the stands during the lunch break.

Heather Harrison and Racheal Ige, British Gypsum.

Session 1 Speakers: left to right Simon Bailey, Frederic Davidts, Chris Calam, Allan Finley, Ken Granger, David Beveridge (Chair), Adam Housley, Lee Parry, Michael Brogan, Paul Vanden Branden, Thierry Theato and Dave Speake.

Session 2.

New super eruptions from Yellowstone hotspot: XRF fingerprinting

Leicester University’s Dr Tom Knott, told us about the super eruptions which can cover 10,000 km³. No need to worry though, as we have only had 42 in the last 36 million years! Looking at Yellowstone in USA, he studied Snake Plain as part of an international team to understand effects and frequency. Snake Plain formed at 900 C to 1000 C. With multiple eruptions it can be difficult to identify one from another. XRF is used to look at trace element ratios. A 2 km bore hole revealed analysis of Ti, Si, Al & Fe consistent throughout the core. However problems arose when looking at rhyolites. When samples were analysed by two separate labs there was a 20% discrepancy, this was due to the USA lab running samples outside the calibration range.

Tom found that the biggest single continual eruption was at Castleford measuring 1900 km3, this being the 2^nd largest in Snake River. The frequency of these eruptions has tailed off. Tom assured us that Yellowstone is safe to visit and the pictures look truly amazing.

Recent ground breaking developments in ED XRF.

“The more counts the better!” This was the message from Thierry from Spectro. Traditional ED XRF gives around 100,000 cps and WD systems giving 1,000,000 cps. This all leads to fewer errors for us the user. The ED XRF XEP05 offers nearly ten times more counts than their previous ED XRF and is 3 times more sensitive by using direct excitation. The new Pd/Co alloy X-ray tube can use Pd for (Na-Cl, Fe-Mo, Hf-U) with the Co for K-Mn; the advantages are like having 2 tubes in 1. This new instrument uses polarized and direct excitation, with a band pass filter with a 60 KV tube. The Co/Pd tube is difficult to manufacture so tube targets can’t be changed for different elements but the Co/Pd alloy is doing a great job at delivering the magic million cps for ED XRF.

DOT 3 Update

So it’s time to find out how good we are at analysing the unknown. Ros provided feedback on the results obtained for the DOT 3 sample which was a gypsum based cement that consisted of mainly calcium and sulphur. The data was variable but normalising the data to CaO gave better grouping and removal of IR combustion analysis made trending sulphur easier. With a number of fused bead results low in sulphur, these are likely to be from loss on fusion. Calibration errors may well be adding to the variability. CaO and SO₃ were good on pressed pellet, but high results on SiO₂ and Al₂O₃.

The best approach to analyse the sample was explained by Heather from British Gypsum. Heat sample at 45-50 ºC to remove any free water. Fusion was then carried out by pre-igniting at 1000 ºC for 1 hour, using 66 % lithium tetraborate and 34% lithium metaborate in a ratio 7:1. Raw materials and certified reference materials were used as standards. To note, because these samples contained sodium, these wouldn’t make a good plaster, as these migrate to the surface causing stains. Thanks to all that carried out analysis.

Judith from Alfred H Knights gave us an insight into how to carry out the analysis of unknown samples. Start by checking if your sample is wet? Think about what the loss on ignition (LOI) will tell us with regards to carbonates, volatiles or gains. Trial a fusion at 9: 1 g (lithium tetraborate/lithium metaborate) and utilise an internal standard. Judith then invited us all to analyse a DOT4 sample. We were all given a clue about the composition being a ferromolybdenum sample that requires oxidation…….but what else can we find in this? Let’s get busy and send our results in for next year.

Adam Grayson, Johnson Matthey.

Session 2 Speakers: left to right Judith Bain, Iain Howland (Chair), Tom Knott, Heather Harrison, Thierry Theato and Ros Schwarz.

Afternoon Sessions.

Session 3.

Session 3 Speakers: left to right Paul Vanden Branden, Steve Davies and Judith Bain (Chair)

Calibration Strategies for Quantitative XRF Analysis, Paul Vanden Branden SciMed.

Paul started from first principles, that quantification is the conversion of measured intensity into concentration, according to:

which he expanded to

Where

are calibration constants x intensity

And (1 + "Matrix") is the matrix correction coefficient.

He began by showing a basic linear empirical calibration, equivalent to “y = mx + c”,

The slope and intercept of this line form the calibration coefficients.

The effect of changing the sample matrix was discussed using PbO in glass as an example. Adding another component to the matrix gave a PbO calibration line with a different slope, because the MAC (mass absorption coefficient) of the sample had changed.

Paul described the different types of matrix effects, absorption and enhancement, and strategies for compensating for them.

Alpha Corrections.

In the general formula for quantitative calculation this is the “1 + Matrix” term:

Where α is a calculated correction coefficient for the absorption / enhancement effect of other components in the sample.

Sometimes the Alpha Correction approach cannot be used, if for example, the full compositional information on our calibration standards is not known.

Paul presented some possible alternative approaches to matrix correction as follows;

Internal Standardisation

• With internal standardisation, we add a selected component to the sample

• The addition is maintained at the same concentration for all standards / samples

• Measure the analyte and internal standard peak intensity and ratio analyte to internal standard

• Calibration is concentration vs. intensity ratio

The calibration is independent of matrix effects.

Scatter Correction - Compton and Background

In the earlier example Paul explained how the intensity of a measured peak was affected by changes in the matrix absorption coefficient. We also find that other spectral lines are affected the same way, in particular, tube or secondary target lines which are scattered by the sample. Hence there is a consistent relationship between the analyte intensity and the Compton scatter peak from the X-ray source

This method can be useful when the matrix is “unknown” e.g. waste materials.

However, if there are any major absorption edges between the analyte line and the Compton scattering, the correlation to the sample matrix is broken and the correction does not work well.

Background in XRF is also caused by scattering . Paul used an example of Zn peak intensity ratioed to its own background.

Standard Additions

This method is effective when we don’t have calibration reference samples and can’t generate a calibration curve.

A known amount of the element to be analyzed is added and then we estimate the concentration of the element by the increasing ratio of x-ray intensities.

We need to pay attention to the following items:

- Since it is assumed that the calibration curve is linear, this method is applicable to concentrations below 1 %.

- We have to use net intensities (with background subtraction). Any inaccuracy in background correction = error in result

- In the case of powder sample, we need to add the element by solution; adding the solution as the whole sample is soaked, mixing well and drying it, then pressing the dried sample

- Standard additions analysis may be prone to mineralogical errors

Conclusion

• Many options to calibrate

• Choose carefully based on sample type and amount of CRMs available

Small Spot Mapping - The Latest ‘Must Have’ Steve Davies PANalytical.

It is becoming popular because of the following reasons; Small spot analysis with element distribution mapping is an ideal tool for materials research and production process troubleshooting. No longer confined to research facilities, this technique is now available anywhere you need it.

The latest developments in XRF analysis allow bulk analysis combined with small spot analysis and elemental mapping on a single device. The added capability of small area analysis and elemental mapping on a fine scale extends the scope of possible applications for basic investigations in materials research and production control.

Wavelength dispersive (WD) XRF and energy dispersive (ED) XRF are used in combination with two detectors for light and heavy elements. This combination expands the scope of applications of the instrument. In addition to the accelerated data acquisition, using the ED core for small spot analysis provides other benefits like allowing the WD core to exclusively perform high accuracy and precision bulk analysis​.

Unlike other elemental mapping techniques such as electron microprobe analysis or scanning electron microscopy which require extensive pre-treatment of samples the XRF method requires minimal sample preparation.

Applications include; Metals analysis, Geological, Small samples, Inclusions, minerals.

Steve said that he was looking to achieve several points:

What’s the precision of measurement of a small spot, homogeneity of prepared specimens? Is the sample the same at different points on the surface?

He used a bead made from DOT-3 to test for sensitivity of light elements and reproducibility over 10 measurements on one spot. To examine the homogeneity of fusion he used a pre-prepared specimen of ECRM776-1 measuring 5 separate spots down a radius of the bead.

He found that the precision values were as expected, although he was disappointed about light element results. A different flux would have been more appropriate for the fused bead. Homogeneity was good but he concluded that he should look after his beads better!

Richard Morris, Morris Analytical X-ray Ltd.

Session 4.

Session 4. Speakers: left to right Jonathon Prus, Ros Schwarz (Chair) and Charles Shand.

Trace element analysis of whisky by TXRF (Dr Charles Shand)

This was an interesting topic and concept, which was to see if TXRF could be used in order to group different whiskies from different regions of Scotland, as well as be able to identify counterfeit whisky. All the elements in question (11 in total) had to tell apart the raw materials, product process equipment, storage vessel and any additives that were used in order to paint a full picture of the final product for good traceability. From the results it was clear to see that there were noticeable groups with only few overlaps. What was interesting was the major difference in the counterfeit whisky which was in a clear group by itself with considerably more Br present. These results were compared to ICP and were in good agreement. This proved that although TXRF is not commonly used for trace analyses of elements in alcohol, that it can play a role in this process.

Quantitative aspects of iron extraction in the Sussex Weald. (Dr Jonathan Prus)

The final talk of the day was based on iron minerology which was close to many people’s area of work but also had a historical twist.

The efficiency of the bloomery iron smelting technique to extract iron from its oxides and slags was discussed. PXRF was used to compare the larger samples from different sites and then compared with electron microscope which fell within the tolerance and matched fairly well. From the results it found that around 44% of iron was extracted leaving behind a considerable amount which went to show that this method used was not for efficiency but more because it was a simple way to smelt.

Robert Wood, Alfred H Knight.

....and finally some scenes from the Exhibition