Trace level Cu, Ni, Co, & V standards

[Samuel Saltzman]

Hi,

Anyone know of any good standards for the aforementioned elements.  Anything up to 2 wt% would be great and all information is much appreciated. 

Thanks
Sam

[Probeman]

Hi,

Anyone know of any good standards for the aforementioned elements.  Anything up to 2 wt% would be great and all information is much appreciated. 

Thanks
Sam

Hi Sam,
This is a classic request and the answer is: these so called trace element standards are extremely difficult to come by.  And for a very simple reason:  how does one determine a small amount of an element with accuracy?  It can be done, but it's quite difficult.

What is much easier however is to find a standard with a matrix similar to your unknowns which contains a zero (below detection limit) concentration of your element of interest.  If your pyrite standard is FeS2, and you've checked for impurities using say ICP-MS, and the concentration is below the detection limit of your EPMA, then that is a good blank standard for measuring all sorts of trace elements in pyrite.

Why is this useful?  Because then one can answer the most important question of all: how well can I measure *zero* in a matrix similar to, or the same as, my unknown...

This is what is often called the "blank" correction in EPMA and a paper on this method is linked here:

http://probesoftware.com/Ti%20in%20Quartz,%20Am.%20Min.%20Donovan,%202011.pdf

Using this method we documented 2 to 3 PPM sensitivity for Ti in SiO2.  Then please read here and you will probably have more than a few questions by then:

http://probesoftware.com/smf/index.php?topic=29.msg387#msg387

Happy to help with any questions.
john

[UofO EPMA Lab]

Hi,

Anyone know of any good standards for the aforementioned elements.  Anything up to 2 wt% would be great and all information is much appreciated. 

Thanks
Sam

Besides the fact that it is very difficult to obtain trace element standards in a suitable matrix with well characterized (whatever that means!) concentrations, we must also keep in mind, that in trace element analysis, we are essentially only concerned with the background correction.

Yes, the matrix correction extrapolation from the standard is involved, but the accuracy errors associated with background measurements are easily larger than the matrix correction physics errors, which at this point are better than 2% in most cases. Also, the precision of the standard intensity is actually more import for analytical sensitivity as the sensitivity equation wants a high std intensity with the smallest matrix correction.  Which generally means a pure element or oxide or other compound with a high concentration of the standard element.

So, testing to be sure that one can indeed measure zero, in a similar matrix, is actually a good sanity check. As Mike Jercinovic has said about the "blank" correction and other attempts to measure zero:

"If you can't measure something, trying measuring nothing. Because if you can't measure nothing, you can't measure anything."

 

[Samuel Saltzman]

John and Oregon,

Thanks for the information on the blank correction method.  Seems like it will be very helpful in many of the upcoming tasks for this lab and I can indeed measure 0 in standard provided. 

Some background on this project. I am the interim lab manager at the University of Iowa with about a year of experience working in the lab under my belt and a month as manager.  I am working on developing a procedure for measuring the amount of Cu, Zn, Ni, Co, Fe, Mn, V, Al, and Cr that have been included (doped)in a synthetic uranium crystal lattice.  Crystals are grown in a organic acid so I have applied a glow discharge and gold coat to stop sample degassing (this was a huge issue).  The concentrations of these elements in the crystal is expected to be between .5-2 wt%.  The uranium standard that I was provided is reproducing excellently but is rather heterogeneous and void of all the aforementioned doped elements.  I expect this to be reflected upon the doped unknowns so that is not an issue.     

I have been able to measure 0 for Cu in my standard and I got an intensity of -3.7 cps and I am not sure what value to apply for Cstd and Clevel.

Also, is this method appropriate for values this high.

Cheers and Thanks,
Sam

[Probeman]

John and Oregon,

Thanks for the information on the blank correction method.  Seems like it will be very helpful in many of the upcoming tasks for this lab and I can indeed measure 0 in standard provided. 

Some background on this project. I am the interim lab manager at the University of Iowa with about a year of experience working in the lab under my belt and a month as manager.  I am working on developing a procedure for measuring the amount of Cu, Zn, Ni, Co, Fe, Mn, V, Al, and Cr that have been included (doped)in a synthetic uranium crystal lattice.  Crystals are grown in a organic acid so I have applied a glow discharge and gold coat to stop sample degassing (this was a huge issue).  The concentrations of these elements in the crystal is expected to be between .5-2 wt%.  The uranium standard that I was provided is reproducing excellently but is rather heterogeneous and void of all the aforementioned doped elements.  I expect this to be reflected upon the doped unknowns so that is not an issue.     

I have been able to measure 0 for Cu in my standard and I got an intensity of -3.7 cps and I am not sure what value to apply for Cstd and Clevel.

Also, is this method appropriate for values this high.

Cheers and Thanks,
Sam

Hi Sam,
Welcome to the field!   You've come to the right place for questions and answers on EPMA...   

Interesting problem. You say the uranium was grown in a synthetic acid so the material contains some organics that are (were) out gassing?  I would be concerned about beam damage.  You might want to perform some time dependent intensity (TDI) measurements to see if the x-ray intensities are stable over time.

The blank correction could be performed by hand, but it is intended to be included in the matrix iteration for best accuracy (it is built into the Probe for EPMA quantification code).  But performing the blank correction off-line in Excel is better than not performing any correction at all.

Basically the "blank level" is the known concentration in your blank standard and the "blank value" is what you actually measure in your unknown.  The difference between these can be applied to your unknown concentration results, in your case the standard would be your pure uranium crystal and the unknown would be your doped uranium crystal lattice materials.   

In PFE the concentration difference is converted to an intensity and applied during the matrix iteration, since technically the matrix is being adjusted, so the matrix corrections should be re-calculated. In practice the changes to the matrix composition are trace to minor levels in magnitude, so they might be ignored depending on the physics details and performed in Excel using just concentrations. 

Basically the blank correction is intended for correction of continuum artifacts such as the holes in some PET crystals under the Ti Ka peak or the curved background under Al Ka in quartz.  However, it can be applied as an accuracy correction even for major elements, though you want to be careful that it is being applied appropriately.

Attached below is a white paper I produced some time ago on using the blank correction for determining water in glasses by measuring the oxygen content directly.

See also here for more details on the blank correction:

http://probesoftware.com/smf/index.php?topic=204.msg3467#msg3467

But I would suspect that, due to the delicate nature of the organic matrix in your materials, you may have bigger problems than just measuring the background.  That is, the TDI correction might be much more important.  You should check out the method described here and perform some of your own tests to check the stability of your materials:

http://probesoftware.com/smf/index.php?topic=11.0

For example, see the last few posts in this topic on measuring the absorbed current during the analysis to check for beam damage over time.

The other method we use in PFE for beam sensitive trace analysis is the "on and off-peak alternating" backgrounds described here:

http://probesoftware.com/smf/index.php?topic=29.msg387#msg387

This allows the intensity measurements to be distributed over time between the on and off-peak positions...

Feel free to call me any time if you want to chat off-line.
john

541-346-4632