24小时热门版块排行榜

返回列表

【奖励】本帖被评价29次，作者staphlee增加金币 24.5 个

当前主题已经存档。

当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖

staphlee

金虫 (正式写手)

应助: 0 (幼儿园)
金币: 1520.8
帖子: 615
在线: 68.3小时
虫号: 615616

[资源] 簡單的教你如何在解晶體的時候辨認原子.....

简单的教你如何在解晶体的时候辨认原子.....
看到一份对初学者很有帮助的文章
对于如何在解结构的过程中指定原子...大家看看吧...
找好资料很辛苦阿....请帮我回文评分....谢谢

ATOM IDENTIFICATION USING X-RAY DIFFRACTION DATA

Peaks from Fourier maps don't come labeled as to atom type, so it is up to us to make the correct assignments. Assuming reasonably good quality diffraction data, it is usually possible to distinguish between atoms of similar atomic number, such as C vs. N, N vs. O, S vs. P, etc., using several lines of reasoning. With luck, all clues will agree, and the atom type will be unambiguous. Comments will follow about each clue.

Ellipsoid Size
Recall that X-rays are scattered by electron density, and least-squares refinement tries to fit the electron density as well as possible, given the assigned elements. In a correct, well-refined model, we expect the ellipsoids of the various atoms of the molecule to have similar sizes, unless there is a good reason to expect otherwise (a librating t-butyl group, perhaps). What do we expect to happen if we assign an atom with fewer electrons than are actually present, perhaps putting a C atom where O belongs? The diffraction data "know" that the electron density is higher there, so the leastsquares tries to increase the density by the only means it has, by decreasing the volume of the atom. So, if we see an ellipsoid which is much smaller than those surrounding it, it is a clue that perhaps a heavier atom is needed there. Conversely, if we put too many electrons at an atom site, the refinement will smear them out over a larger volume to adjust the density, and the ellipsoid will appear larger than expected.

Residual Electron Density
In the case above, with O misassigned as C, when the volume of the atom is decreased by shrinkage of the ellipsoid, a shell is left around the center of the atom in which the electon density is not properly accounted for. So, we expect to see difference electron density peaks around that atom.

H Atoms
Since X-ray scattering is done by electron density and H atoms don't have much of it, they don't always show up too clearly. Low-T helps. With a reasonably good data set for an organic compound, we expect to be able to at least see H atoms in difference maps, even if we can't refine them very well. So a significant clue in distinguishing an ether link RO-R' from R-CH2-R' or an amine R-NH-R' is direct observation of the H atoms. The same can be said of distinguishing a methyl group from an OH group, etc. Be sure to use a PLAN N command in SHELXL with N large enough to provide enough peaks to see all the H atoms. It's always a good idea to make sure we know where the H atoms are before placing them in calculated positions, anyhow. Also, normally H atoms on N or O will be involved in hydrogen bonds, so contact distances to H can be an important consideration. Bond Distances Covalent radii decrease B>C>N>O, so bond distances to an atom of uncertain identity are also a significant clue. C-C single bonds are typically around 1.5 Angstroms, while C-O single bonds are typically about 1.4.

R Values
A single misassigned atom results in a measurably poorer fit to the intensity data. So, you can do refinement experiments in which you assign an atom with different elements and see what happens to the R value. Changes won't be huge, but should be significant.

When the above clues all point to the same atom type, it creates that warm, fuzzy feeling that we know what we're talking about. The process can be messed up by disorder, so it doesn't necessarily work every time. Agreement with spectroscopic data is a good thing, too. Chemists do sometimes know what they are talking about, so talking to them when ambiguities arise can be helpful.

[ Last edited by staphlee on 2009-7-15 at 10:51 ]

回复此楼