olli1183
Dear all.
I tried to refine a tetragonal perovskite compound, which shows different strain along the two different crystallographic axes, i. e. (00l) reflections are far broader than the corresponding (h00) reflections with h=l.
Therefore,
1) I modified John's macro for the sample dependent TOF peakshape
macro tof_sample_peakshape_aniso1(lor,lor_val,dsp,dsp_val,dspsq,dspsq_val)
{
prm dsp dsp_val del = 0.05 Val + 1; min 1
prm dspsq dspsq_val del = 0.05 Val + 1; min 0
peak_type pv
pv_lor lor lor_val
pv_fwhm = sh1 dsp D_spacing + sh1 dspsq D_spacing^2;
}
and used the following in the sample description
str
r_bragg 4.92165139
phase_name "nuclear"
scale scale_P4mm_NPD_1 152209.112`
space_group P4mm
Tetragonal(=a_P4mm_1; : 3.927561`,=c_P4mm_1; : 4.192942`)
site ....
spherical_harmonics_hkl sh1
sh_order 8 load sh_Cij_prm {
y00 !sh1_c00 1.00000
y20 sh1_c20 1.00013`
y40 sh1_c40 0.69319`
y44p sh1_c44p 0.06425`
y60 sh1_c60 -0.27250`
y64p sh1_c64p 0.04774`
y80 sh1_c80 -0.27287`
y84p sh1_c84p -0.07328`
y88p sh1_c88p 0.03123`
}
tof_sample_peakshape_aniso1(lor1b1, 0.423882015`,dsp1b1, 475.14475`,dspsq1b1, 0.00000`_LIMIT_MIN_0)
This works fine and I can now obtain a good fit to the shape of the reflections. The sample furthermore shows also magnetic ordering. When I used Johns macro and constrained the shapes of the nuclear and the magnetic part, this worked totally fine. However, if I introduce a magnetic phase using my modified macro and try to constrain it to the shape of the nuclear phase, this doesn't work out so well and the problem gives me a strange error: "Non-linear least squares (c) 1992-2012 Alan A. Coelho
Loading xyz's for p4mm from file C:\TOPASAV5\sg\p4mm.sg
Trying to load space group name, number etc... but no space group file."
Strangely, if I use this constraint with my modified macro for a non magnetic impurity phase with a different space group (which makes of course not necessarily sense, but I just tried if it works in principle; e. g. P63/mmc instead of P4mm), this works also perfectly fine!
macro tof_sample_peakshape_aniso2(lor,lor_val,dsp,dsp_val,dspsq,dspsq_val)
{
prm dsp dsp_val del = 0.05 Val + 1; min 1
prm dspsq dspsq_val del = 0.05 Val + 1; min 0
peak_type pv
pv_lor lor lor_val
pv_fwhm = sh2 dsp D_spacing + sh2 dspsq D_spacing^2;
}
str
mag_only_for_mag_sites
phase_name "mag 1"
scale =scale_P4mm_NPD_1/16; : 9513.06951`
mag_space_group 1.1
a =(2^0.5)*a_P4mm_1;
b =(2^0.5)*a_P4mm_1;
c =2*c_P4mm_1;
al 90
be 90
ga 90
site ...
spherical_harmonics_hkl sh2
sh_order 8 load sh_Cij_prm {
y00 =sh1_c00; : 1.00000
y20 =sh1_c20; : 1.00013`
y40 =sh1_c40; : 0.69319`
y44p =sh1_c44p; : 0.06425`
y60 =sh1_c60; : -0.27250`
y64p =sh1_c64p; : 0.04774`
y80 =sh1_c80; : -0.27287`
y84p =sh1_c84p; : -0.07328`
y88p =sh1_c88p; : 0.03123`
}
tof_sample_peakshape_aniso2(lor1b1, 0.423882015`,dsp1b1, 475.14475`,dspsq1b1, 0.00000`_LIMIT_MIN_0)
Any idea why this happens? Thanks in advance, Olli.
alancoelho
Hi Olli
Just had a look and at present spherical_harmonics_hkl cannot be used with magnetic space groups.
The reason is that the magnetic space group typically comprises hkls that are outside of the symmetry of the nuclear space group and as such the symmetrized spherical harmonics of spherical_harmonics_hkl will be broken.
If you used P1 (ie. magnetic space group 1.1) then there is no symmetry and therefore spherical_harmonics_hkl should be allowed to work for P1. For P4mm then spherical_harmonics_hkl should not be allowed to work unless the magnetic hkls are the same as the nuclear hkls. This is something that needs more thought.
Would you be able to e-mail me the INP and data file and I will have a closer look.
cheers
alan