Indeed.
eu·tec·tic/yo͞oˈtektik/
Adjective: Relating to or denoting a mixture of substances (in fixed proportions) that melts and solidifies at a single temperature that is lower than the melting points of the separate constituents or of any other mixture of them.
i.e. eutectic = molten metal
The steel beam probably melted when an eutectic mixture of molten metal containing sulfur dripped on it.
NIST did not do as FEMA recommended and determine what caused Sample #1 to melt.
NIST lied when they said that NO steel was recovered from WTC 7.
OK - I'm going to try and be gentle because the subject of high temperature corrosion is quite specialised. There's maths and chemistry and solid state physics involved, but this is a brief (ha!) explanation in the simplest terms I can think of for this particular example. I'm going to use diffusion to explain it but there are other chemical examples.
All metals when they oxidise increase in weight. This is because atoms from an element (species) in a gas environment (oxygen in air) diffuse into the metal's surface. The metal is still solid, but a reaction is taking place. Once that scale forms the surface of the metal has been changed to an oxide and therefore it's more difficult for oxygen atoms to diffuse into this new surface.
The mechanics that determine how fast this occurs and the depth of oxide layer are temperature, the concentration gradient of the diffusing species across the metal/gas interface, diffusivity of species and time (surface area, D - Diffusion coefficient, pressure, alloy composition, etc are also factors but it gets complicated fast). Imagine it as a flow of atoms from the gas into the solid across the metal/gas interface. Increase heat increases the rate. Increasing the concentration of the gas increases the difference between concentration of the gas at the metals surface and the concentration of the diffusing species in the metal - concentration gradient goes up. Increase the time, more atoms flow across the interface. Diffusivity of species is how mobile or how easy it is for that atom to move in the solid - solid state diffusion.
So for example carbon is "highly mobile" in iron and is an example of interstitial diffusion. Remember BCC, FFC, HCP etc crystal arrangements? Well if you take a BCC iron then you've got 1 Fe atom in the centre and 8 Fe corner atoms. Well the carbon atom being small will fit into the sites on the faces of the cube and midway along the edges. Heat is the driving force that will move these interstitial carbon atoms through a crystal lattice of iron atoms. Arrrgh, getting technical!
We can measure the rates of diffusion because we can weigh samples that have been exposed to air or any other gas at a given temperature. Over time the samples gain weight and then we can plot weight gain verses time.
We can then see if there is a linear relationship or parabolic (curved) one. And from these curves using solutions to f-i-c-k's laws (profanity filter picks this one up) we can determine rates etc.
We can also do other things like cut up a steel that's been carburised and measure how deep the carbon has penetrated for a given time, temperature and partial pressure to make sure the process parameters are giving us what the equations say it should give us!
I hope that is reasonably clear for everyone. So lets move on a bit.
Imagine a piece of low carbon steel say 1/2" thick, inch wide and 5 inches long suspended in a furnace that allows a gas mixture to be pumped in. In this case we'll use SO
2 and CO
2. The temperature will be held at 1000°C and the concentration (partial pressure) of SO
2 and CO
2 will be a reducing one. What happens?
I'm going to ignore the austenite phase change, spheroidization of cementite, dissolution of pearlite, recrystallisation, grain growth (especially columnar), etc because it's just too much and isn't required.
First of all there are 2 corrosion effects - oxidation and sulphidation. Both are atoms moving from the gas into the solid, and it's quite usual to find both.
Secondly we are going to get decarburisation - that is carbon atoms diffusing out of the steels surface. This happens in a reducing atmosphere.
Now as oxygen and sulphur are diffusing into the steel's surface and carbon is going out the composition of that steel at the surface
is changing.
Metals are made up of crystals (grains) and where these crystals join there is a boundary - grain boundary. These boundaries are "weak" areas mechanically, thermally and chemically. For example if you heat a metal up to it's melting point you will find that melting occurs at the grain boundary first. Cracks often follow grain boundaries and grain boundaries are susceptible to chemical attack.
Diffusion of a species such as sulphur, in high temperature corrosion, will be preferential, that is easier, at these grain boundaries where these boundaries meet the steel's surface.
So if the steel has been attacked at the grain boundary by S and O and the composition of the steel changes to the Fe-O-S eutectic composition then this material will melt above the eutectic melting point.
In our furnace example at 1000°C we would expect this to occur (if we've got the SO2/CO2 gas concentration correct).
Right so now we have a liquid that is sitting on the steel surface and along grain boundaries as the result of inter-granular melting (liquation). Remember these boundaries are weak and are very, very small. The liquid is around 20µm (microns) thick - average human hair thickness is 100µm. This liquid is now going to be allowing a couple of things to happen. Firstly because it is below the surface level the interior is now closer to sulphur and oxygen. This means the sulphur and oxygen can now penetrate deeper than if they were just at the solid metal/gas interface so the corrosion rate increases. Secondly this liquid is causing liquid metal embrittlement akin to hot shortness. Sulphides at grain boundaries are not good news since they cause embrittlement so the steel is more susceptible to cracking.
As the corrosion penetrates deeper along grain boundaries the material is being weakened as grains are effectively being circumvented by the liquid, differences in thermal expansion between the oxide layer and the steel will cause the oxide layer to spall (become detached) exposing fresh steel below that is then subject to diffusion of S and O. The mechanism continues.
There are other complications, eg: silicon in the steel and other "impurities" lowering the eutectic further, the possibility that the liquid eutectic is acidic, etc but that's by the by.
I hope that helps Chrismohr with his debate with Gage and ofcourse anyone else who had the stamina to read it!
