Yesterday a new version of GSAS-II (v5.9.0) was released which incorporates a new sequential refinement mode — that allows for histograms to be grouped. The histograms within a group are used for a combined refinement, but groups are fit sequentially. See https://urldefense.us/v3/__https://github.com/AdvancedPhotonSource/GSAS-II/discussions/376__;!!G_uCfscf7eWS!av9crI3k8sbzWf1NygqCI2mzuLtaFLZNgpWx2C8Vc4JiztqY_g0TvRB8JamkfmLkcrBh-arraORFK3s$ <https://urldefense.us/v3/__https://gcc02.safelinks.protection.outlook.com/?url=https*3A*2F*2Fgithub.com*2FAdvancedPhotonSource*2FGSAS-II*2Fdiscussions*2F376&data=05*7C02*7Cgsas-ii*40aps.anl.gov*7Cfdc587f4be88487b9e8908df35151a7e*7C0cfca18525f749e38ae7704d5326e285*7C0*7C0*7C639288036905424997*7CUnknown*7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ*3D*3D*7C0*7C*7C*7C&sdata=k*2B*2Bm25o75cWqgWPpzYLv9mAsnu8oYUT6A*2Fj4pcY3kcI*3D&reserved=0__;JSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUl!!G_uCfscf7eWS!av9crI3k8sbzWf1NygqCI2mzuLtaFLZNgpWx2C8Vc4JiztqY_g0TvRB8JamkfmLkcrBh-arr60AZ554$ > for more discussion. The reason for this e-mail is that this version also fixes a bug that could affect the results of TOF fits, and we want to alert people about this bug.
In version c1ced7b of GSAS-II (June 2025, ~v5.4.7/#5814) a computation was introduced that applied the hydrostatic strain tensor terms (Dij) terms twice for TOF datasets (only). Thanks to two users who pointed out inconsistent results to us, we found the bug and fixed it. To explain why this is about, the hydrostatic strain tensor terms provide offsets for lattice constants. They are required for sequential fits and are occasionally used in combined fits. They are found in the Data tab for each phase. With this bug, when these values were fit, then the refinement would converge to lattice constants that are incorrect, but again only with TOF data. If you were using x-ray or CW neutron data, this bug would not have affected you. Likewise, if you used TOF data in GSAS-II, but did not refine the hydrostatic strain (which we often call the “Dij” terms) you also would not have seen this bug, but if you performed a sequential fit with TOF data or used these terms to deal with difference in lattice constants between histograms (such as in the PbSO4 tutorial) with TOF data, then your fit would provide inaccurate lattice constants and less than an ideal fit, though usually only to a minor extent. To correct this, if this bug affected you, please update GSAS-II to the latest version and simply repeat the last refinement. The refinement will then provide the right results. Questions: reply here or contact either of us directly. Brian & Bob
_______________________________________________ GSAS-II mailing list [email protected] https://mailman.aps.anl.gov/mailman/listinfo/gsas-ii
