The EDS can detect nitrogen (N), present in many cleaners; as well as carbon (present in many oils); oxygen (usually not terribly informative because it is in rust, some oils and many industrial grinding media); silicon (Si); and aluminum (Al), which can be present as abrasive residue or “dirt.” EDS can’t detect hydrogen (H), helium (He) or lithium (Li). Helium is unlikely to be present in very many things, but hydrogen is. That’s where we switch to the FTIR to detect the carbon-hydrogen bonds. Hydrogen and helium will never be detectable by EDS methods because the atoms are not big enough. Lithium theoretically could someday, but it would take great advances in electronics. So when I am looking for lithium-grease residue, I am out of luck unless it has PTFE in it. In that case, I might detect the fluorine (F).

See Figure 1, which shows an EDS spectrum for a dark stain on a shiny electroplated steel stamping.

The EDS (energy dispersive spectrum) method is considered semi-quantitative rather than fully quantitative. One reason for this is because the penetration depth of the electrons that generate the X-rays, which allow the atomic elements to be identified, varies as a function of the density of what is being analyzed. If it's pure zinc, it is denser than if it is a mixture of zinc, zinc oxide, iron, carbon, sodium and chlorine.

Another important thing to understand about the EDS method is that, when used alone, it does not allow anyone to prove that the oxygen is present as zinc oxide. It could be a mix of zinc and iron oxides, or it might be oxygen held in an oil molecule along with the carbon. I give my client my best guess based on many factors, including appearance to the eye, appearance in the optical microscope and how it behaves in the electron microscope as it interacts with the electron beam.

Furthermore, the depth of penetration also depends on how fast we "shoot" the electrons at the surface we are analyzing. In the case of Figure 1, the electrons were accelerated using a voltage of 15 kV. The electrons are likely penetrating approximately 1 micron. There are computer programs to simulate this, but it is pretty difficult to get a realistic or believable value when the surface is not perfectly flat.

It's possible that there was some residual stamping lubricant on this part that did not get cleaned up before it was immersed into the plating bath. Carbon with traces of chlorine are often found in stamping-lube residue. The chlorine could have gotten stripped from the oil molecule and transformed, in the presence of moisture, to hydrochloric acid. That is aggressively corrosive.

I'm not sure where the sodium (Na) came from, but note that the Na peak is on top of one of the zinc (Zn) peaks. I'm not sure how, exactly, the computer figures out the concentration of sodium versus the zinc, but it likely has to do with the exact position of the tall mixed Na/Zn peak at approximately 1 eV. Pure Zn would be at 1.012, and pure Na would be at 1.041 eV. Older detectors may not be able to distinguish two compounds whose peaks are so close together. The computer may also somehow take account of the relative peak height of the zinc peak at 8.63 eV.

This gives a little taste of how challenging this type of work can be. Every once in a while, there is an obvious answer.

The EDS can detect nitrogen (N), present in many cleaners; as well as carbon (present in many oils); oxygen (usually not terribly informative because it is in rust, some oils and many industrial grinding media); silicon (Si); and aluminum (Al), which can be present as abrasive residue or “dirt.” EDS can’t detect hydrogen (H), helium (He) or lithium (Li). Helium is unlikely to be present in very many things, but hydrogen is. That’s where we switch to the FTIR to detect the carbon-hydrogen bonds. Hydrogen and helium will never be detectable by EDS methods because the atoms are not big enough. Lithium theoretically could someday, but it would take great advances in electronics. So when I am looking for lithium-grease residue, I am out of luck unless it has PTFE in it. In that case, I might detect the fluorine (F).

See Figure 1, which shows an EDS spectrum for a dark stain on a shiny electroplated steel stamping.

The EDS (energy dispersive spectrum) method is considered semi-quantitative rather than fully quantitative. One reason for this is because the penetration depth of the electrons that generate the X-rays, which allow the atomic elements to be identified, varies as a function of the density of what is being analyzed. If it's pure zinc, it is denser than if it is a mixture of zinc, zinc oxide, iron, carbon, sodium and chlorine.

Another important thing to understand about the EDS method is that, when used alone, it does not allow anyone to prove that the oxygen is present as zinc oxide. It could be a mix of zinc and iron oxides, or it might be oxygen held in an oil molecule along with the carbon. I give my client my best guess based on many factors, including appearance to the eye, appearance in the optical microscope and how it behaves in the electron microscope as it interacts with the electron beam.

Furthermore, the depth of penetration also depends on how fast we "shoot" the electrons at the surface we are analyzing. In the case of Figure 1, the electrons were accelerated using a voltage of 15 kV. The electrons are likely penetrating approximately 1 micron. There are computer programs to simulate this, but it is pretty difficult to get a realistic or believable value when the surface is not perfectly flat.

It's possible that there was some residual stamping lubricant on this part that did not get cleaned up before it was immersed into the plating bath. Carbon with traces of chlorine are often found in stamping-lube residue. The chlorine could have gotten stripped from the oil molecule and transformed, in the presence of moisture, to hydrochloric acid. That is aggressively corrosive.

I'm not sure where the sodium (Na) came from, but note that the Na peak is on top of one of the zinc (Zn) peaks. I'm not sure how, exactly, the computer figures out the concentration of sodium versus the zinc, but it likely has to do with the exact position of the tall mixed Na/Zn peak at approximately 1 eV. Pure Zn would be at 1.012, and pure Na would be at 1.041 eV. Older detectors may not be able to distinguish two compounds whose peaks are so close together. The computer may also somehow take account of the relative peak height of the zinc peak at 8.63 eV.

This gives a little taste of how challenging this type of work can be. Every once in a while, there is an obvious answer.