\(\renewcommand\AA{\unicode{x212B}}\)

FindUBUsingLatticeParameters v1

Summary

Calculate the UB matrix from a peaks workspace, given lattice parameters.

See Also

SetUB, FindUBUsingFFT, FindUBUsingIndexedPeaks

Properties

Name

Direction

Type

Default

Description

PeaksWorkspace

InOut

IPeaksWorkspace

Mandatory

Input Peaks Workspace

a

Input

number

Mandatory

Lattice parameter a

b

Input

number

Mandatory

Lattice parameter b

c

Input

number

Mandatory

Lattice parameter c

alpha

Input

number

Mandatory

Lattice parameter alpha

beta

Input

number

Mandatory

Lattice parameter beta

gamma

Input

number

Mandatory

Lattice parameter gamma

NumInitial

Input

number

15

Number of Peaks to Use on First Pass(15)

Tolerance

Input

number

0.15

Indexing Tolerance (0.15)

FixParameters

Input

boolean

False

Do not optimise the UB after finding the orientation

Iterations

Input

number

1

Iterations to refine UB (1)

Description

Given a set of peaks, and given lattice parameters (\(a,b,c,alpha,beta,gamma\)), this algorithm will find the UB matrix, that best fits the data. The algorithm searches over a large range of possible orientations for the orientation for which the rotated B matrix best fits the data. The search for the best orientation involves several steps.

During the first step, a reduced set of peaks typically at lower |Q| are used, since it is easier to index peaks at low |Q|. Specifically, if there are at least 5 peaks, the peaks are shifted to be centered at the strongest peaks and then sorted in order of increasing distance from the strongest peak. If there are fewer than 5 peaks the list is just sorted in order of increasing |Q|. Only peaks from the initial portion of this sorted list are used in the first step. The number of peaks from this list to be used initially is specified by the user with the parameter NumInitial. The search first finds a list of possible orientations for which the UB matrix will index the maximum number of peaks from the initial set of peaks to within the specified tolerance on h,k,l values. Subsequently, only the UB matrix that indexes that maximum number of peaks with the minimum distance between the calculated h,k,l values and integers is kept and passed on to the second step.

During the second step, additional peaks are gradually added to the initial list of peaks. Each time peaks are added to the list, the subset of peaks from the new list that are indexed within the specified tolerance on k,k,l are used in a least squares calculation to optimize the UB matrix to best index those peaks. The process of gradually adding more peaks from the sorted list and optimizing the UB based on the peaks that are indexed, continues until all peaks have been added to the list. Finally, one last optimization of the UB matrix is carried out using the full list of peaks.

Usage

Example:

ws=LoadIsawPeaks("TOPAZ_3007.peaks")
print("After LoadIsawPeaks does the workspace have an orientedLattice: {}".format(ws.sample().hasOrientedLattice()))

FindUBUsingLatticeParameters(ws,a=14.131,b=19.247,c=8.606,alpha=90.0,beta=105.071,gamma=90.0,NumInitial=15)
print("After FindUBUsingLatticeParameters does the workspace have an orientedLattice: {}".format(ws.sample().hasOrientedLattice()))

print(ws.sample().getOrientedLattice().getUB())

Output:

After LoadIsawPeaks does the workspace have an orientedLattice: False
After FindUBUsingLatticeParameters does the workspace have an orientedLattice: True
[[ 0.04542062  0.04061954  0.01223576]
 [-0.00140377 -0.00318446 -0.11654506]
 [-0.05749773  0.03223779 -0.02737294]]

Categories: AlgorithmIndex | Crystal\UBMatrix

Source

C++ header: FindUBUsingLatticeParameters.h

C++ source: FindUBUsingLatticeParameters.cpp