Can someone help me interpret xrf of a magnetic material?

[Larry Smith]

Hi. I'm a retired helicopter engineer that's taken on a project to analyze concretions found in the bay. I can tell they contain iron because they attract to a magnet. One looks like it might be a ship spike, the other was on a small plank (either side of nail hole). I imagine that they might be from an old shipwreck, I just don't know how old. I had them xrf analyzed and results are confusing me since, in addition to lots of iron oxide Fe2O3 and silicon oxide SiO2, there's many elements present including Cu, Zn, and Sn plus several race amounts of REEs. Is it possible that my xrf data can be interpreted to correlate when items were made? My target period is 16th century. It's beyond my area of expertise so I'm hoping someone on this forum has understanding of what was in compositions of metals. Thank you.

Sample 1/Plank Concretion: Oxides (sample plus duplicate test results)
mass% mass% mass% mass% mass% mass% mass% mass% mass% mass% mass% mass%
Na2O MgO Al2O3 SiO2 P2O5 SO3 K2O CaO TiO2 MnO Fe2O3 Total
ND ND 1.56 66.9 0.087 4.14 0.139 ND 0.172 ND 27.0 100.0
ND ND 0.834 68.1 0.076 0.198 ND 0.0176 0.042 ND 30.7 100.0

Sample 1/Plank Concretion: Elements (sample plus duplicate test results)
mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg
Cl V Cr Ni Cu Zn Ga As Se Br Rb Sr Y Nb
243 ND < 54.8 ND 34.8 65.7 ND 11.0 ND ND ND 6.32 34.3 ND
712 ND < 43.8 ND 34.1 15.9 ND 18.2 4.52 7.58 ND 6.76 33.3 ND

Sample 1/Plank Concretion (cont.): Elements (sample plus duplicate test results)
mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg
Mo Ag Cd In Sn Sb Cs Ba La Ce Hg Pb Th U
ND ND ND ND 28.5 ND ND ND ND ND ND 19.3 30.1 ND
ND ND ND ND 20.6 ND ND ND ND ND ND 27.8 < 28.5 < 16.5

Sample 2/Spike Concretion: Oxides (sample plus duplicate test results)
mass% mass% mass% mass% mass% mass% mass% mass% mass% mass% mass% mass%
Na2O MgO Al2O3 SiO2 P2O5 SO3 K2O CaO TiO2 MnO Fe2O3 Total
ND ND 0.952 30.2 0.061 5.03 ND 0.163 ND 0.360 62.0 98.8
ND ND 0.929 29.3 0.061 5.04 ND 0.176 ND 0.372 62.9 98.8

Sample 2/Spike Concretion: Elements (sample plus duplicate test results)
mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg
Cl V Cr Ni Cu Zn Ga As Se Br Rb Sr Y Nb
1550 ND 182.0 ND 1250 4650 28.2 133.0 ND ND ND ND ND ND
1560 ND 162.0 < 38.9 1250 4820 31.8 140.0 ND ND ND 12.2 ND ND

Sample 2/Spike Concretion (cont.): Elements (sample plus duplicate test results)
mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg
Mo Ag Cd In Sn Sb Cs Ba La Ce Hg Pb Th U
ND ND ND ND 235 ND ND ND ND ND < 24.2 206.0 < 101 ND
ND ND ND ND 225 ND ND ND ND ND < 19.9 215.0 ND ND

I am grateful for any feedback that you may provide. My conjecture is that this is some very dirty iron from a crude smelting process.

 

[EdStainless]

If one of these was sectioned would there be any solid metal to analyze?
The microstructure can also tell a lot about how and where the material was made.
Some of the ones with Cu/Zn/Sn are likely from being used to fasted brass or bronze parts into place.


= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Larry Smith]

Attached photo is of sample 2. It does appear to have solid metal on the inside. It broke open and in pieces when I accidentally dropped it. I sent off half of the tip end for xrf testing. The other half is available for microstructure evaluation. Who can do this? I'm a novice and what I've learned is because I found a university with fee-for-service xrf capability. Fwiw, there was no solid metal left in sample 1. At best, it (and sample 2) might reveal its inside with a CT scan, but that's another bridge too far as finding who does it. Sample 2 had higher mg/kg values of Cu, Zn, and Sn than did sample 1 (since sample 1 appeared in majority like the "corn dog" casing in photo of sample 2). I'm also curious about the other trace elements like As and Pb, plus Y, Th, and U, if these are residuals from ore. Another curious observation, the small rock on the upper spike section glows orange under ultraviolet 365nm light. I don't know why. Thanks again for thinking about puzzle. Any thought on next step is welcome.

 

[MFAMET]

To simply the main composition for these two samples:

Si: 60-67%; Fe: 27-32%, S: 4-5%.
They appear as Ferrosilicon alloys. They are electric, magnetic, and brittle. Many residues may come from the scraps used during EAF smelting.

Note: XRF is mainly used to determine the composition of elements only. They are incapable to analyze any form of oxide compounds. Your lab reports are not the professional type.

 

[EdStainless]

One of the issues that you will have is that the concretion that has formed will contain a lot of entrapped material as well as corrosion products and biological residues. Considering all of the minerals in seawater this could introduce almost anything.
You would learn a lot more by sectioning the larger piece, finding some solid metal, and then looking at its composition and microstructure.
You should be able to rather quickly decide if it is recent or older.
And you want chemistry done by OES or some similar method. XRF will not give you light elements (such as C) which are critical.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Larry Smith]

Thank you MFAMET. You are right. The XRF method is capable to detect and quantify the element contents only. Not oxides. The reports were provided in oxides after recalculation. Geological reports are commonly reported in major oxides. I did not know to stipulate otherwise.

Thank you EdStainless. I will see what I can learn from sample 2 since it has solid metal. I'll look at composition and microstructure. Good to know what chemistry to do (Optical emission spectroscopy). I appreciate clues and insight to limits of xrf data for my project.

 

[MFAMET]

My thought was the lab used a computing system to calculate out the compound/oxides composition with the obtained data of element composition. The conversion formula is readily figured out after taking into the account of oxygen element.

 

[Larry Smith]

Yes, MFAMET, the calculation to back out oxygen is pretty straightforward. Thanks again for your contribution. You helped me to learn.

 

[MFAMET]

Other than microstructure and composition studies, you may want to ask lab to perform the microhardness measurement prior to etching the mounted sample. This could also provide the info of mechanical properties.

 

[Larry Smith]

MFAMET, will do. Thanks again.

 

[MagBen]

16th century? this is ancient metallurgy! ancient China, India and Middle East, people used smelting process with iron ore and silica, and added coke as heat source and reducing agent to make Fe-Si alloy. similar to cast iron. Si increase fluidity, decrease melting temp. while mechanically IMPROVE HARDNESS AND WEAR RESSTANCE. iron ores contain lots of residuals Al2O3, S, P As, Hg and some heavy metals.

oxide calculation may not be accurate, Fe has at least three oxide forms, FeO, Fe2O3 and Fe3O4 among which FeO is very weak magnetic.

by the way, attracting to a magnet does not mean it is iron base (e.g. other elements Ni, Co, Mn, and RE can be magnetic), iron-based alloy doensot necessarily mean it is magnetic (e.g. austenitic stainless is non-magnetic)

 

[Larry Smith]

Thanks, MagBem. Roger that on your magnetism comment and of possibility for other ferrous oxides. It will be interesting to see what I can learn beyond a list of elements from xrf. So far: I'd like to CT scan the head on the concretion to see if it reveals any clues, then (as EdStainless and MFAMET suggest), perform analyses for composition and microstructure, chemistry by optical emission spectroscopy, and a microhardness measurement. Have I missed anything? Also, I only know one company in Pensacola who can do it all and they need to agree. I'm open to using other providers with capability and public-facing services to support my independent research project initiative. fwiw