Instructions for powgen pdf data Reduction in Mantid and pdfgetN

Instructions for POWGEN PDF Data Reduction in Mantid and PDFgetN

Katharine Page ([e-mail])

Ashfia Huq ([e-mail])

Here are instructions for using the Mantid data reduction framework in combination with PDFgetN to produce total scattering datasets for POWGEN data. These procedures are intended to be temporary solution that will be replaced with a fully automated Mantid PDF data reduction in the near future.

We have found the combinations of Frame 1 and Frame 5 to provide high quality and high resolution total scattering data for an oxide powder in a 10 mm Vanadium Can. We will publish these quantified results soon, but wish to inform the user community as soon as possible.



Figure 1: BaTiO3 data collected on POWGEN in a 10 mm Vanadium Can in frame 1 and Frame 5 (6 hours total time).



In Mantid:

Find the run numbers corresponding to:

1. 
   Sample Bank 2 (0.533)
   
2. 
   Sample Bank 5 (2.665)
   
3. 
   Vanadium rod (normalization) Bank 2 (0.533)
   
4. 
   Vanadium rod (normalization) Bank 5 (2.665)
   
5. 
   Vanadium sample can (sample background) Bank 2 (0.533)
   
6. 
   Vanadium sample can (sample background) Bank 5 (2.665)
   
7. 
   EmptyPAC (empty instrument) Bank 2 (0.533)
   
8. 
   EmptyPAC (empty instrument) Bank 5 (2.665)
   

Locate the POWGEN instrument calibration file (*.cal):

Locate the POWGEN Characterization Run File (*.txt):

Select SNSpowderreduction from the dropdown menu on the right.

Insert the string of eight run numbers corresponding to the above list.

PushDataPositive: None

For PDFgetN we want to normalize to proton charge only. The following selections should be made:

Back Num -1

VanNum -1

VanBackNum -1

Calibration File:

Characterization File: Most of the entry will be ignored, only the tmin and tmax will be read in from this file.

RemovePromptPulseWidth: 50

Binning: -0.0008 for most samples (- means log binning)

We want to filter out anything that is less than 10% of the maximum proton charge and scale the data by 100.

FilterBadPulses: 10

ScaleData: 100

We also want to save data in gsas format for reading into PDFgetN.

SaveAs: gsas

Finally, we need to choose an output directory, elect Momentum Transfer as the data display option, and XXX.

OutputDirectory

See Screenshot below



Figure 2: Correct Mantid selections in SNSPowderReduction for PDFgetN input files.

At this point the individual wavelengths are reduced.

You will need four files to complete the POWGEN total scattering reduction in PDFgetN:

1. 
   Sample
   
2. 
   Vanadium rod (normalization)
   
3. 
   Vanadium sample can ()
   
4. 
   EmptyPAC (empty instrument)
   

Each set of data should be combined in a single file. There is a small routine on our analysis computers that can be used to accomplish this. This routine is called mergegsas. In a terminal type

mergegsas input filename #1.gsa input filename #2.gsa -o outfile file name.gsa

Alternatively you can write a shell script with all the files and run that. A shell script is a text file with the extension .sh that can be executed in the terminal with a command

sh filename.sh

Following is an example of what a shell file content

mergegsas PG3_22014.gsa PG3_22016.gsa -o PG3_Vrod_2015A.gsa

mergegsas PG3_22041.gsa PG3_22042.gsa -o PG3_EmptyPAC_2015A.gsa

mergegsas PG3_22029.gsa PG3_22030.gsa -o PG3_Empty8mm_2015A.gsa

mergegsas PG3_22845.gsa PG3_22846.gsa -o PG3_7LMNO-SC.gsa

In PDFgetN:

The POWGEN team collects data for (2)-(4) at the beginning of each cycle, or when instrument configurations change. Make sure you use the normalization and instrument background that correspond to the configuration of your experiment. Also make sure you select the appropriate V can for your measurements.

We have provided a PDFgetN history file and data for a BaTiO3 and LaB6 samples for your reference: filename. We have tried to keep instructions here specific to POWGEN and normal (cylindrical Vanadium can) measurement configuration. Please refer to the PDFgetN manual for further options and explanations.

You will also need to save and install the special SNS version of PDFgetN.

Sample Configuration Tab

Enter sample information in the typical way. A screenshot for our BaTiO₃ sample is provided in Figure 2.

Figure 2: PDFgetN settings for Sample configuration tab.

Experimental Configuration Tab

The V cans on POWGEN have the following dimensions:

Diameter	Inside Diameter (cm)	Outside Diameter (cm)	Inside Radius (cm)	Volume of 4 cm high Sample (cm^3)
6 mm	0.63	0.59	0.295
8 mm	0.786	0.746	0.373
10 mm	0.942	0.902	0.451

All cans have a wall thickness of 0.02 cm.

If you used a different configuration you should change dimensions appropriately.

Be sure to select the POWGEN instrument parameter file matching the run cycle and instrument configuration of your experiment. Here we select: PGHR_60-2014B.iparm. A screenshot of the Experimental tab including the information from our measurement is provided in Figure 3.

Figure 3: PDFgetN settings for Experimental configuration tab.

Detectors Configuration Tab

If you

A screenshot of the Detectors tab including the data cuttoffs we used is provided in Figure 4.

Figure 4: PDFgetN settings for Detectors configuration tab.

Hints: Absent data reduction aberrations and significant incoherent inelastic contributions (for instance, from H), the baselines from the separate banks of data should overlap. If your baselines do not match, check the sample chemistry you entered and the sample, vanadium, and container dimensions. These factors will influence normalization, absorption and multiple scattering corrections, and may impact the normalized total scattering structure factor intensities. If you do have significant incoherent inelastic signal, you may wish to complete an external correction for the baselines, like those outlined in the Kpage or Ilevin.

Plotting Configuration Tab (bottom right buttons in Figures 2-4)

There are a few settings you should ensure are correct in Edit All Mode. These should already be correct if you use the provided history file as a template.

In the dropdown menu accessible from the ‘Plot’ icon in the bottom right, select ‘Edit All’ and select ‘Do it’. This opens exert edit mode. Ensure that the Vanadium can cross section information is correct. The scattCS=5.100 absorpCS=5.080. Also ensure that the Vanadium rod peak killing is on. The values should read vanKillThresh=4.0 and vBackKillThresh=4.0.

Figure 5: Expert Edit mode options.

Results:

Figure 6: Impact of Data Collection Times: BaTiO₃ in a 10 mm can

The standard normalization and background measurements completed at the beginning of each cycle and after changes to instrument configuration are completed for 3 hours. To match the statistics presented above

Figure 6: Effect of Q_max on 3 hr Data

Other Configurations: If you set out to do a total scattering measurement on POWGEN, we recommend the combination of Frames 1 and 5, providing Q max between ~1 and up to 40 Å^-1. Higher Q maximums require substantially increased data collection times, and will not be supported without significant additional justification. The POWGEN frame 2 provides a Q range of 1.4 to 22.8 Å^-1, which may be appropriate for some situations. If your system does not have Bragg Scattering below 1.4 A-1 in Q (or above 1.4 A in d).

Figure 7: Full Q range of combined banks 1 and 2.

Note: We are very pleased to help you with data reduction . Please do not hesitate to contact us if you encounter difficulties. We are actively developing a Mantid-based total scattering data reduction routine for POWGEN. The routine above is intended as an interim solution. Please contact Katharine Page ([e-mail]) and Ashfia Huq ([e-mail]) with questions or suggestions.