Data+Analysis+

The typical procedure for data collection is, to first set up the beamline for the desired absorption edge. This involves tuning the monochromator to the required energy, and filling the ion chambers I0 and It with the correct mixture of gases to ensure sufficient absorption.

This is typically around 20% absorption for the first ion chamber (I0)and between 50-80%

absorption for the second ion chamber (I_t). Once the monochromator has been set-up and ion chambers filled it is ideal to run either a standard material or a foil, e.g. for Cu edge running either CuO powder or Cu metal foil.²² This allows users to determine whether the data is of a reasonable quality and to check the edge position to ensure the monochromator is calibrated properly.

Once standards have been run, everything is ready to start experiments. If in situ measurements are being performed the cell would need to be aligned at this point to ensure that all data taken is of a good quality. Care needs to be taken to ensure that when the cell is mounted on the beamline the beam does indeed pass through the sample and does not clip the cell. This would result in vast abnormalities in the data, in addition to an unnatural loss in

2.4.5.1 Multiphase+and+MultiHCluster+Analysis&

For data collected on more than one absorption edge, it can be beneficial to fit the data sets simultaneously. This allows for terms and interactions present in both files to be fitted together increasing the integrity of the fitting model. The same principle can also be applied to setting up a multi-cluster model, where by the same edge but two or more different structures are fitted simultaneously.

It is a relatively simple task to achieve in Artemis, in the top window simply select add data set to add another data window to the calculation. This can then be set up as outlined in the previous section with the data from another absorption edge and using a separate .cif file. The real power of using multiple absorption edges becomes visible in the parameterisation step.

There will be paths, for example Au-Cu in copper gold bimetallic clusters that are present in both the Au LIII-edge and Cu k-edge data. By setting up these paths with the same parameters it provides a link between the results of the two models. This can be used to verify the presence of a species in the system and also to improve the accuracy of the filling model.

Aside from the ability to fit multiple datasets the other most powerful aspect of using Artemis for data analysis is the throughput rate. Once a model has been set up it is possible to simply read in data sets and analyse using the same model. User discretion is required to monitor the quality of the results and the integrity of the fitting model, however this provides a powerful method to process large datasets in a short amount of time. This is one of the main reasons for the significant use of Artemis analysis in the thesis work.

2.4.5.2 EXAFS+Analysis:+Excurve+Software+

Excurve is another software package used for analysing EXAFS. The program is able to read any input file as long as the column layout is known. As such it is possible to convert Athena, Viper and potentially other software outputs into a form readable by Excurve. As Excurve is a much older program and does not have the same community support and updates that the Athena and Artemis packages have, however the versatility of the software means that it is updated periodically by experienced users to the benefit of the EXAFS community. In contrast to Artemis, Excurve is an Ab-initio based software package, which has its own advantages and disadvantages.

The program is command driven, and it can generate graphical outputs to assess the quality of the fitting model. When starting an analysis, the first set is to input the types of atoms. Taking Au metal as the example again, the first step would be to input two types of Au atoms, one for

the first and second shells respectively. Following this the potentials between the atoms are calculated, followed by the phase shifts. Also the symmetry of the system can be set as this will aid in generating the theoretical model. The most important feature of Excurve lies here;

a model can be created without symmetry using atomic coordinates to manually enter a model. This is a distinct difference to Artemis and it’s reliance on crystallographic information or model building, and allows for a much more interactive analysis of amorphous materials.

Once the model has been constructed and the number of shells selected the experimental data can be read in. it is essential to start with a standard such as a foil or material of precisely known composition as this will be used to obtain the AFAC term. This value is constant through out all experiments on the same edge during a beamtime and is analogous to the amplitude (amp) term in Artemis, it typically has a value of between 0.6 and 1. When running fits it Excurve it is possible to set a limit for the number of iterations; this is a maximum value, though the software will ask for permission to continue if a stable answer isn’t obtained within the iteration limit. It is possible to save all of the calculations, such as the phase shifts, in addition to the fit data so that simply reloading the files in the future can perform a perfect reconstruction.

Another powerful feature of Excurve is the ability to utilise multiple scattering in the calculations. Whilst Artemis takes into account multiple scattering from the FEFF calculation, Excurve calculates them all separately. Activating the function in the software converts every degenerate path into separate on atom paths and calculates the positions and interactions independently. Whilst this is very beneficial in terms of the integrity of the fit, it does increase the calculation time. Another drawback to this feature is the number of independent parameters required. This usually requires some creative fitting strategies in order to produce good results.

This is a short overview in the Excurve EXAFS analytical software; whilst there are many more features this is sufficient information to communicate the benefit of the software as an alternative to Artemis and also to outline the limitations. Like any technique or data processing method it is not flawless, but by selecting the right software, based on the nature of the material, the best possible interpretations for the experimental data can be obtained.