ferrite structure.

firkin

firkin

Joined
Jan 26, 2002
Messages
2,737
Shgeo,

sorry, if you looked earlier and found nothing--I got delayed.

For (Fe)3 (C), as far as I can tell, the periodic solid structures are going to be a cubic lattice of Fe atoms, with either face-centered or body centered carbon atoms. And that's been around for a long time.

In fact, somebody, probably Mete, posted links with some pictures a while back.

If you're interested I can post a qualitative description of how that makes sense, chemistry wise. I get a charge of -1 on the carbon and + 1/3 on the iron atoms.

Coming up with a cyrstallographic unit cell for the body-centered case will be easiest.

First, just to try to get on the same page, a unit cell is derived from a space lattice or internal co-ordinate system that depends on which of the seven crystal systems the material exists in. This space lattice is totally different from the crystal lattice that the atoms ocupy. In very simple structures of pure elements the space lattice points occupy the same position as the nuclei. This is very, very unusual.

The general definition of a unit cell is a volume defined by three angles and three pairs of lines whose ratios need not be 1:1:1. Fortunately, life is a little easier in the cubic crystal system that we are discussing. The three angles are 90 degrees. The lengths of the lines need not be equal.

I think I have the unit cell for the body-centered cell, but I have to make a model to check and figure out how to describe.

The face-centered instance is more complicated. I may think about it, maybe not :)

PS: "Tinkertoys" make stuff like this easier, and are not real expensive.

I have a couple old sets of these, but it looks like nobody sells the sets anymore except Aldrich chemical supply--dunno if they sell to individuals :( add a little flexible tuping or a few pipe cleaners and they will handle most stuff.

http://www.edumodel.com/2parts.html

http://www.sigmaaldrich.com/cgi-bin/hsrun Distributed/HahtShop/HahtShop.htx;start frmCatalogSearchPost?ProdNo=Z203467&Brand=ALDRICH

I've used these, many more atom types available , but I like the former better if lots unusual co ordinations aren't needed . One of the crystal lattice sets should do nicely.

http://www.indigo.com/models/orbit-molecular-model sets.html

Or you can use min-marshmallows/playdoh and toothpicks!
 
Firkin,

The seven crystal systems are defined by 32 point groups that can be developed by translations into the 230 space groups. See Crystallography and Crystal Chemistry by F. Donald Bloss for details. Unfortunately I loaned the book out to someone who never returned it. Mineralogy texts in general cover the symmetry of point and space groups, but Bloss is the best-if it is still in print.

The Isometric (cubic) system has all angles = 90ºand A1 = A2 = A3. All the lines must be equal.
The 5 point groups are 4/M, /3, 2/M; 4,3,2; /4 3 M; 2/M, /3; and 2, 3. These are developed into 15 space groups by the various symmetry translations.

If all angles = 90º and C≠ A1 = A2, It is in the Tetragonal system.

If all angles are 90º and C≠ B ≠ A, It is in the Orthorhombic system.

The geometry of the unit cell ie. FCC vs BCC depends on the bond lengths and energy in the material and can change as in the Austenite FCC vs the Ferrite BCC. What I was trying to simplemindedly figure out was the large scale structure of the cementite layers, based on the simple formula.

This is pretty academic anyway as I use high alloy steels with structures I never will figure out.
 
shgeo,

OK, OK...

we are on the same page as to what a unit cell is!!

And it looks like I misread my crystallographic text. :o The few structures I did a long time ago were just discrete chemical componds of very low symmetry--the really high symmetry stuff was always more confusing to me, maybe because I never had any experience with it.

BTW, the point groups are extensively used in the consideration of bonding.

Anyway, I thought I could pick a unit cell out of this stucture--but I can't...Maybe you can do better:

http://cst-www.nrl.navy.mil/lattice/struk/d0_11.html
 
Firkin, on your link click to see several perspectives and check out the right view. Each C atom appears to be in coordination with 5 Fe atoms. The 4 sided pyramid is on its side in the view with its basal 4 Fe atoms shared while the apex is not. This would give the Fe3C unit cell. A mirror plane translates to a Fe6C2. Each of these is surrounded by Fe, possibly metallically bonded?

The surrounding Fe would not be part of the cementite structure if this is right. I was trying to simplify it too much into discrete layers of cementrite and ferrite. I still lean towards the C bonds being at least partially covalent.

Edited to add: This unit cell would have 432 symmetry. I will try to find out if this is the proper point group for Cementite-more later

Edited again: The Fe atoms each share 1+ with the surrounding Fe structure and the basal four of the pyramid share 1+ each through the mirror plane with the other pyramid ie, 1/2+ charge times 4 plus 1+ from the Apex Fe gives the Fe3C unit cell. I thought it sounded too simple above.
 
Shgeo,

follow the links at that site and download RASMOL!! I just updated my older version.

Then save the co-ordinates in XYZ format they havelisted on your machine. May have to copy from the web page into a text editor--

Then you can rotate and visualize in 3D. Stereo mode too if you can control your eyeballs--you can set how "splay-eyed" or "cross-eyed" you want.

It works fine on my ancient box for files of this size--not slow--should be great on a newer machine.

I haven't used RASMOL much and that was a long time ago...I'm trying to figure out if I can hide atoms other than by editing the input file.

Sadly, the crystal symmetry info isn't part of an XYZ file, or RASMOL would pick out the unit cell for us.

But, if you look again at the link, you'll see that they provide a lot more information further down the page that I never learned how to use or have forgotten...I'm pretty sure that the required crystal symmetry information is there, but I don't know if RASMOL could accept and use it. Look at the "Cartesian" and "lattice coordinate" tables--I think they might be saying that 8 atoms are in the unit cell???

What I can figure out right now is that the space group is listed as Pnma, which my text says is in the point group mmm. Also, the carbons away from the edges are formally depicted as being 8-coordinate by this program--I haven't looked at bond lengths yet or if this display program allows criteria to be set for their dislay.

Try RASMOL, it's fun until you get a headache form making your eyes look at stereo pictures.:)
 
Shgeo, just saw your last edit.

"The surrounding Fe would not be part of the cementite structure if this is right.."

:confused: I thought it was just an artifact of what the authors selected to display the periodic structure.

Anyway, I just checked the the XYZ coordinate file that I downloaded. Maybe I missed one or two, but there are 37 C atoms and 99 Fe atoms. Sure looks "clean" to me.

Pity that it is only an XYZ file, I think I can see distortions in the structure--The RASMOL display program's algorithm (whatever it is) did not draw 8 "bonds" to some of the C atoms (also left out some expected Fe-Fe "bonds"), even though there there were Fe atoms in the approprate locations. I haven't measured the distances, but it looks like some local C environments were distorted compared to others.

I'd bet that the thermal anisotropy factor for the C atoms (maybe all atoms) is pretty high. I suspect that you're right that the bonding looks more like a 4-sided pyramid, and that the C atom bounces around inside the 8 (maybe only 6 are bonding possiblities since to me this is orthorhombic symmetry because the space group is Pnma) nearest Fe atoms but is bonded to only four of them at a time. Carbon just doesn't have energetically accessable orbitals to expand its valence shell from what I know.

"I still lean towards the C bonds being at least partially covalent."

:confused: To me, "partially ionic" is an oxymoron. Covalent bonds run the gamut from symmetrical bonds between diatomic molcules to covalent bonds between unlike atoms that have a substantial, permanent dipole. Maybe just semantics.

As far as "metallic" Fe-C bonds go, my concept is that electrons in Fe-C bonds are either in a filled band that is energetically separated from an unfilled band, and are thus "non-metallic", or they are in an unfilled band or in a band very, very close in energy to an unfilled band. "Metallic" or "non-metallic", an either-or situation. If there are two kinds of Fe-C bonds there could be two bands for Fe-C bonding electrons with an energy sepatation. Then the electrons in one type could be "metallic" and the ones in the other type of bond could be "non-metallic", again depending upon occupancy and energetic proximity to partially filled bands. As far as I know, complex computer (and necessarily approximate) calculations are the only way to address this. Predicted vs measured magnetic moments are the experimental "verification" of the calculations.

"I was trying to simplify it too much into discrete layers of cementrite and ferrite."

My impression was that the lammelar structure was more than a few atoms in thickness, and dependent upon the temperature regimen.

Which leads to:

Where did they get a crystal of this stuff anyway??
I'd think must have been big by metalurgy standards, even if they did neutron diffraction.
 
If you take Fe3C and throw out the Fe then you could take the carbon and make carbon nanotubes which would be far stronger than steel.
 
I'll take my carbon as diamonds, use some to buy a cabin in the woods, an what's left over to buy some nice knives, thanks.
 
I am going to take this opportunity to give this up and do something fun. I have six blanks cut out of D2 that I need to grind and deer season starts two weeks from Wednsday.
 
You must log in or register to reply here.
Back
Top