Large standard deviations!

pengguo2

Dear topas community,
I am currently working on refining a novel crystalline porous material. During the process, I've noticed a larger standard deviation for the coordinate values on the y-axis, particularly for the P2(1) space group.

prm x1 0.48798_0.00077 prm y1 0.7652219831.28516
prm z1 0.91204_0.00051 prm x2 0.287280.00068
prm y2 0.27214_19831.28516 prm z2 0.732070.00050
prm x3 0.14471_0.00074 prm y3 -0.2514219831.28516
prm z3 0.48627_0.00055 prm x4 0.547010.00074
prm y4 -0.23088_19831.28516 prm z4 0.609000.00048

site P1 num_posns 2 x = x1; y = y1; z = z1; occ P = 1 ; beq = b1 ;
site P2 num_posns 2 x = x2; y = y2; z = z2; occ P = 1 ; beq = b1 ;
site P3 num_posns 2 x = x3; y = y3; z = z3; occ P = 1 ; beq = b1 ;
site P4 num_posns 2 x = x4; y = y4; z = z4; occ P = 1 ; beq = b1 ;

Furthermore, I tried increasing the symmetry from P2(1) to P2(1)/m. Although this constrains most of the y-axis values, the resulting bond angles and lengths are not reasonable.

Could you possibly guide me on how to rectify this? Your expertise would be greatly appreciated.

Peng

johnsoevans

P2(1) has a floating origin. If you look in international tables you'll see it can lie anywhere on the y-axis. There's therefore nothing to stop your whole asymmetric unit moving along y, which leads to the large standard uncertainties. It's common to fix the y-coordinate of one atom (usually the heaviest one) to fix this.

Ramon13

Just out of curiosity because I wasn't able to figure it out from the technical manual but I also calculate the errors in JEdit, when you have "prm y1 0.76522_19831.28516" is that to say the value is 0.77 +/- 19831 ? Similarly with "prm x1 0.48798_0.00077" it's 0.49 +/- 0.0008?

johnsoevans

They're the standard uncertainties which would normally be written in crystallography as 0.4880(8) or similar. More or less what you say, but different communities interpret +/- differently. The IUCr dictionary has an entry on this.