# Tutorial:What to do with data

Revision as of 10:31, 1 December 2017 by KevinYager (talk | contribs)

A common question from new GISAXS users is: "What do I do with my data?" Of course, the answer is: "It depends on what scientific question you're trying to answer!" The scattering you observe is essentially the Fourier transform of the realspace structure of your sample. So, in principle any structural question you have about your sample can be answered by analyzing the scattering. Of course, in practice some questions are easier to answer than others; your data may or may not be sufficient to get the answer you want.

The main ways of using scattering data are:

**Qualitative inspection**: By just looking at the scattering, and using some 'rules of thumb', you can infer much about the structure of your sample (e.g. differentiating between ordered vs. disordered; isotropic vs. anisotropic). This is especially valuable when you're measuring samples, to get a rough idea of what's going on. For some purposes, this may even be sufficient for publishing (i.e. you simply qualitatively describe your data in the paper).**Linecuts**: A more sophisticated analysis involves converting your data from the raw detector image into*q*-space (reciprocal-space), and then taking 'linecuts' through this space to learn more about your sample. For instance, you can quantify a peak position (by fitting to a Gaussian), and thereby measure the realspace repeat-spacing of the associated structure. You can also quantify grain sizes, orientation distributions, etc.**Indexing**: For ordered materials (atomic crystals, superlattices, etc.) that generate many peaks on the detector, indexing can be used to prove that you understand the structure of your sample. That is, you compare the predicted peak positions for a candidate unit cell to the experimental data, and show that your model is correct. The peak positions can also be fit, in order to quantify the unit cell parameters, and the unit cell orientation.**Modeling**: You can use various pieces of modeling software to predict the scattering patterns for candidate structures. By comparing these to your experimental data, you can eliminate some possibilities, while demonstrating that others are consistent with your data. This doesn't*prove*that you have identified the correct realspace structure, but it provides good evidence. This is most successful if you have identified a likely structure using other means (e.g. electron microscopy), in which case comparing the experimental scattering to the output of a theoretical model gives you good confidence that the proposed structure was present in the measured sample.**Fitting**: Theoretical models can be quantitatively fit to your experimental data. This allows you to quantify parameters of interest (lattice spacing, disorder, etc.), assuming you've selected the right model!