What is PDFgetX2?

PDFgetX2 is a user friendly program to obtain the atomic pair distribution function (PDF) from X-ray powder scattering data. The interaction between users and data is facilitated by a rather extensive graphical user interface (GUI). The supported platforms include Linux, Windows, and Mac OSs. Please refer to Appendix A for detailed installation instructions. The supported file formats include SPEC, CHI, and free data format. Please refer to Appendix B for details.

The PDF, G(r), is a real space function telling the probability of finding pairs of atoms separated by distance r in the material. The experimental PDF can be thought of as the sum of snapshots of the instantaneous atomic arrangements over the data collection time. Accordingly. the PDF can tell us about both the local structure (low r region) and average structure (high r region). The PDF method has been used as early as 1931 by Debye [1] to study liquid mercury, and has mostly focused on glass and amorphous materials [2,3]. Not until recently, the PDF analysis has been successfully applied to crystalline and nano-crystalline materials [4]. We would like to direct you to a recent book by Egami and Billinge (2003) [5] for more technique details and recent PDF applications.

The PDF, G(r), is obtained by a direct sine Fourier transformation of the structure function S(Q), where Q is the magnitude of the scattering vector. However, the collected raw data during X-ray scattering experiments subject to various instrument and sample effects, which have to be properly corrected to extract the S(Q). This program, PDFgetX2, lives to serve this purpose. Standard corrections [7,6,5] implemented include subtraction of background scattering, sample absorption, X-ray polarization, unwanted Compton intensity, and normalization by the average atomic scattering power. Particularly for the RA-PDF experiments recently developed by Chupas et al.  [8] utilizing the image plate area detector, additional corrections due to oblique incident angle dependence [9] and the detector energy dependence [10] are also implemented. Statistical uncertainties due to limited intensity counts are estimated and are propagated up to the PDF, G(r), which becomes very important when as the S(Q) and G(r) data are increasingly being modeled using regression algorithms. In addition, simple data smoothing and damping are also possible.

A standing alone functionality in this program is to preprocess the raw SPEC data to a single 4-column data set. You have the options to apply the detector dead time corrections on monitor and detector counts; measured intensities from different scans can be either normalized by their collection time or monitor counts. Scaling between different scans when merged into a single set is automatically computed to achieve the best overlap. Error propagation is also carried out.

All features of the program is accessible from the built-in GUI, however, manually editing data processing parameters are possible through a text window. One noteworthy feature is the automatic addition of data processing parameters into the S(Q) and G(r) files as header history information. The history information can also be saved alone. Restoration of the history information will reproduce the original S(Q) and G(r) which may become much desirable when careful examination of data processing is found necessary at a later stage.