Q&A Recap: XRF solutions for automotive catalyst recycling

For the determination of precious metals platinum, palladium, and rhodium in spent automotive catalytic convertors, we’ve created an easy to use and accurate method. During our webinar March 24th, I had the pleasure to introduce it together with XRF application specialist Youhong Xiao.

During the live demo at your desk, we not only introduced how to analyze platinum, palladium, and rhodium in spent automotive catalyst. We also showed how straightforward an XRF measurement can be with our Epsilon 1 and Epsilon 4. At the end of the webinar, we did a Q&A session, but although we did our best, we couldn’t answer all questions. Therefore, we would like to take the opportunity here to answer a few of them.

If you did not join the live webinar on your desk, but are interested in, here you are able to find the recording of the session.

How can surface roughness affect the measurement?
XRF is a surface sensitive technology. It measures the elements in the surface of the sample. Therefore, accuracy will improve by a better-defined measurement surface when it is a densely packed sample with a flat surface.

What is optimum mesh size of powder for accuracy of analysis?
For the autocatalyst application a mesh size below 100 um, i.e. around 75 μm, should be sufficient to determine the precious metals accurate. More important is to keep the particle size constant from sample to sample.

What grinding tool do you recommend for spent autocatalysts?
There are different mill technologies available, with each their own advantages and disadvantages. A swing mill have the advantage to be fast and easy to clean, but investment is a bit higher. You can also opt for a ball mill, which would be a bit slower, but is more economic.

Could you please share the calibration powder again?
To create the method presented and offered as ready-to-use solution for the Epsilon 1, we used fully characterized Spent Automotive catalyst samples supplied by different customers worldwide. Besides the three precious elements, also other compounds and elements are included in the calibration.

The major elements Al, Si, etc. were also in the calibration for matrix corrections?
Suppose handheld ones do not give good results for light elements? How important are these light elements for accuracy and repeatability of the PGM metals?
If the main focus is on the three precious metals, it is not directly necessary to include an accurate reading for Al and Si. Those light elements often are present in larger quantities and can be taken into account as balance without affecting the final accuracy for the PGM.

Do you just calibrate platinum, palladium, and rhodium or all base metals as well so that the instrument calculates the final PGM results in the best way?
The application included calibration for other base metals heavier as potassium.

What is the difference in the accuracy of analysis if we use compressed pallet or loose powder of autocat?
If you are interested in the complete composition of the catalyst including the lighter compounds, the accuracy will be better chosen the pressed pellet method as the surface is better defined.

Which binder is best for making pallets and what is the optimum ratio?
For the preparation of the method described in the Epsilon 4 application note, we mixed 1 gram of UltraWax binder with 10 grams of sample.

How can Epsilon 1 or Epsilon 4 manage Foreign material (such as tantalum for cheating platinum reading) mixed up in sample to pull the platinum peak up?
Our software will deconvolute a measured spectrum, which means it fits a theoretical model with the actual measurement. Using deconvolution, the Epsilon software is capable to separate the separate contributions of individual elements to one larger peak, and therefor can make a separation between the contribution of tantalum on platinum.

Is OMNIAN on Zetium / Axios based on certified standards a suitable method to quantify spent automotive catalysts in a similar quality as your Epsilon? Or is OMNIAN not accurate enough?
Omnian at Zetium will provide more accurate results compared to Omnian at a benchtop if you look at the full periodical table. The question if Zetium + Omnian will provide comparable results for precious metals vs a dedicated calibration on Epsilon 4 for the precious metals will possibly not directly be the case, as in the dedicated calibration on the benchtop different interelement effects are included and focus is on heavier elements the benchtop has good sensitivity for.

Does the Epsilon XRF only measure one sample at a time?
Yes, in XRF you measure 1 sample at a time. If you want to load more samples at once, the Epsilon 4 has a sample changer for 10 samples at once.

What is the minimum level of magnesium in a metal that can be accurately measured by the Epsilon 1?
This depends on the metal, but as Epsilon 1 will measure Mg in air, a LLD can be expected between a few thousand ppm to 1 wt-%.

For the iron scan, the voltage and amperage were altered, is there guidance to these variables for different metals/alloys?
In our Epsilon software recommendation for the kV setting and filter per group of elements is given. To determine the final ampere / power setting, there is a function available who will advise the recommended setting regarding current and therefore applied power.

How do you determine the analysis time? In production, we are pressed to use the minimum amount of time.
The chosen analysis time is determined by many different factors, based on making a choice between the required throughput and the accepted accuracy and precision. Accuracy and precision will benefit by measure longer, but this will affect the total throughput.

The final requirements will also be the optimal XRF instrument for your needs; if you best can opt for an Epsilon 1, a more powerful Epsilon 4, or even to a high power Zetium.

What is RMS and how does it play a role?
RMS stands for Root Mean Square, which is comparable to the standard deviation, and shows in what range a result can be expected when repeating a measurement.

When we put results on normalization, why does accuracy of results differ?
When you apply normalization, the software expects you will see 100% of the composition of your sample. For metals this is straightforward, as most element in metals, can be seen by XRF and therefore an accurate reading is given. For mineralogical samples this already become more difficult as most elements are present in their oxide form, where the oxide will not be analyzed directly, but indirect calculated form the counter element in the compound.

But if there is water or carbonate present, it can contribute to the total weight of a sample, but this is not into account when calculating the final composition.

Do moisture and LOI affect the analysis?
This will depend on how the application will be setup, but if you look at the situation where normalization is used, elements who can’t be analyzed with XRF, will not be included when determining the final concentration.

What are key factors for analysis accuracy?
Final accuracy is determined by several factors,

• The spectrometer and chosen technique. Maybe the complexity of the samples and elements to analyze could require WDXRF
• Quality of the standards used to set up the calibration
• Are all measured peaks recognized and are interelement effects assigned for?
• Sample representation, and is the method chosen how to prepare the sample suitable for the required accuracy
• How alike are the samples to be analyzed or are there strong matrix and mineralogical effects expected.

Are the detectors the same in the Epsilon 1 and Epsilon 4?
The type of detectors is the same, namely an SDD detector, but for the Epsilon 4 there is an option for a larger surface area detector.

Further reading

• Exploiting X-ray diffraction techniques for catalytic materials characterization
• Customer story: FJ Church & Sons – Scrap Trading Company
• Using fusion to prepare petroleum refineries catalysts for XRF analysis
• Aluminum trace elements analysis by Epsilon 1 for small spot analysis

Stay tuned for more content next year by following our Advanced Materials specialists on LinkedIn.

You can also be the first to read our industry updates by subscribing to our Advanced Materials newsletter here.