[jamesrb]

For an undergraduate Instrumental Analysis lab I have been given the task of designing two experiments. I will design and execute experiments for determining the amount of zinc in a sample of sheet brass and the amount of copper in the same sample of sheet brass but this must be done via two different instrumental methods. We will also be given pure copper metal for, I assume, in generating a calibration curve.

The instrumental lab here is not exactly decked out with instruments and we are basically limited to the use of a PerkinElmer ICP-AES and a UV-2450/2550 Spectrophotometer.

Now the determination of copper in brass by UV-VIS spectroscopy is pretty well documented. I have several references and a pretty good idea of how to go about the experiment.

However that leaves ICP as the only method left for the determination of zinc in the sample. I haven't had much luck finding literature relating to analysis of brass by ICP. I am wondering if I am heading in the right direction. If I can't find any literature to help me this could be difficult. I'm pretty sure ICP can be used for this type of analysis but I'm wondering why it isn't typically.

Furthermore the absence of any type of zinc standard alarms me. How am I supposed to generate a calibration curve to determine the concentration of the unknown. I am curious as to what direction the instructors want us to go but I am not ready to turn to them. They want their help to be minimal and so I turn to you all  .

[Corribus]

Short of laser ablation, your sample has to be a dilute solution in order to inject into an ICP-AES. Which basically means your sample has to be digested, usually in acid. For a metal alloy, this is usually simple enough. Your standards are prepurchased (or, if you feel confident, made yourself) metal ion solutions of known concentration.

[jamesrb]

I didn't think this through enough but the determination of zinc will prove difficult by ICP due to spectral interference from copper. I need to separate the copper or the zinc first.

[Corribus]

Most modern instruments can measure many lines simultaneously. Complete interference of one element by another is very rare. Usually you can find a combination of lines that works for you. Of course, optimizing the method for best simultaneous detection limits will be a challenge, especially when one element is much more abundant than another. But you should be able to detect both of these at one time I would think.

You need to look at the various lines for each element, and pick one for each element that is free from intereferences by the other element.

You can find an incomplete list of some lines and potential interferences here:

https://www.ecn.nl/docs/society/horizontal/hor_desk_19_icp.pdf

Most instruments have a much more comprehensive list of lines and interferences built into their software.

[jamesrb]

I'm really tempted to attempted this Multiwavelength Linear Regression Analysis method described here:

http://pubs.acs.org/doi/pdf/10.1021/ed066p178

It really kills two birds with one stone and I think my instructors might allow it, part of my grade is based on innovation and I think maybe they want us pitch a curve-ball. I am hesitant because I fear other impurities present in brass may interfere with the spectra.

[Corribus]

Ok, but: why would you try to do a deconvolution of overlapped peaks if you can find two peaks that aren't overlapped? It's not just a matter of workload, either. Deconvolution introduces additional sources of error and uncertainty into a measurement. Thankfully here the lineshapes should be pretty simple, but still.

[jamesrb]

I had assumed a spectrum of a zinc/copper solution would not have any peaks free from overlap but I'm getting the feeling that is incorrect. I had assumed brass was chosen for this fact in order to prevent the analysis being done with one instrument. However we can't use one method for to for both determinations so I guess I am stuck with describing two experiments. I could prepare an experiment for the multiwavelength linear regression method as a contingency plan.

[Corribus]

I had assumed a spectrum of a zinc/copper solution would not have any peaks free from overlap but I'm getting the feeling that is incorrect. I had assumed brass was chosen for this fact in order to prevent the analysis being done with one instrument. However we can't use one method for to for both determinations so I guess I am stuck with describing two experiments. I could prepare an experiment for the multiwavelength linear regression method as a contingency plan.

Honestly, I don't know. I don't have all lines for each element memorized. But it seems unlikely every line for zinc will interfere with every line for copper. 

If you are required to develop two methods, obviously you have to develop two methods, though. ICP-MS is another option, though probably overkill. AAS also would work, but all those are essentially the same method. Have you considered some kind of X-ray methodology (SEM with EDS)? Or, there are likely chemical ways of quantifying the elements, maybe involving spectrometry or selective chelation.

E.g., this paper may be useful to you:

http://pubs.acs.org/doi/abs/10.1021/ed100259b

[jamesrb]

I am pretty much limited to ICP-AES and UV-VIS. I am at a small public university and the teaching facility is limited. I encountered that paper you linked in my research but I think the analysis needs to be performed instrumentally.

The determination of copper via UV-VIS is pretty well documented and I am starting to create a procedure for that.

I was hesitant to use ICP for zinc because I couldn't find any references and didn't know what I was doing. However the link you provided from the National Institute of Public Health and the Environment indicates that the only interference present at the 206.191 nm line should be from chromium, which won't be present in the sample. I also have an article that describes the separation of zinc from copper alloys with 1-(2-tiazolylazo)-2-naphthol (TAN) and then retention of that complex onto a SEP PAK C18 cartridge. The complex is then eluted with ethanol. So that's a method of separation that could be used. I wonder if there are some simpler methods for separation or ones that use common laboratory reagents.

[Corribus]

Oh, so you actually have to perform these expts? That changes things.

What ICP-AES do you have? If you know how to get around the software, you should be able to find the lines for each element and what potential interferences are. As long as you find two that are mutually compatible, and are both sensitive enough*, you should be golden. You can also call the instrument manufacturer, ask for an applications specialist. Tell them what you're trying to do, and they can help you find a suitable pair of lines. I don't know if you are allowed to do this in your class, of course.

*And even then, you are OK if you are willing to do some concentration steps and run the experiments twice, one for the less sensitive element, one for the more sensitive element.

[jamesrb]

It is a PerkinElmer Optima 8x00. I emailed my instructor to ask if it is 8000 or 8300.

[Corribus]

So, if you are handy with the software, you can investigate for interferences yourself. If not, call Perkin Elmer, ask to speak with an applications scientist for ICP, ask if they have any application notes for brass analysis. If not, they can still probably help you identify a good set of lines for your two elements.

[jamesrb]

So if I can find a line for zinc without any interference from copper is the analysis pretty straightforward (i.e. just generating and using a calibration curve)? I am wondering why I need a set of lines, one for each element, when I am only determining the amount of zinc in brass.

Supposedly 206.191 nm is a good line for zinc free of copper interference and copper has two good lines free of zinc interference (324.754 and 327.396). Hopefully there will be no spectral overlap at 206.191 nm from 213.856 nm. That's pretty close but I'm hoping using low concentrations will prevent any overlap. If not I can use PerkinElmer's Multicomponent Spectral Fitting technique as a contingency plan:

http://www.perkinelmer.com/CMSResources/Images/44-130615TCH_OptimaMSF.pdf

I am curious in the article below why there is no mention of using the 206.191 nm line. I am worried typically there is too much spectral overlap. However it is from 1997 and they did use a different ICP.

http://ac.els-cdn.com/S0039914097003937/1-s2.0-S0039914097003937-main.pdf?_tid=51b92dea-d49f-11e4-bf30-00000aab0f01&acdnat=1427474326_1429980107c7027338cc5dfe3c6f5fff

Ultimately I am pretty confident I can do a pretty straightforward analysis of brass. I'm hoping I can find some good lines to use for quantification but I am a tiny bit worried the instrument may not be able to provide me with such.