8.3.1. Calculating models on grids

Model spectra are generally calculated for an energy or wavelength grid. In SPEX, the default energy grid has 8192 bins between 0.001 and 100 keV. This grid is used when calculating models without an instrument defined, for example, when you plot a model spectrum using the plot type model command (Plot types). The resolution of this grid is usually enough, but when it is not, the egrid command can help to increase the resolution or broaden the energy range (Egrid: define model energy grids).

8.3.1.1. Energy grids when loading data

When data is loaded into SPEX, then the energy grid is automatically adapted to the loaded response matrix. Because response matrices already contain a model energy grid and a channel grid, the SPEX energy grid should match the one from the response, at least within the energy range defined in the response matrix. This way, the model is always calculated at a resolution that is set by the response matrix.

8.3.1.2. Special cases

For some models, like PION, the broad-band spectrum is very important for the calculation. In those cases, the energy band should be extended. Recommended is a logarithmic grid between  keV with a step size of 0.005 (see Pion: SPEX photoionised plasma model). Thus, the default grid has 200 bins per decade. Which energy grid should be used depends on your settings and wishes. The range (lower and upper energy) matter for the physical outcome of the PION model, due to the effects of (inverse) Compton scattering of high- and low energy photons.

When an observed spectrum and response are loaded, we could stick with just the energies contained in the matrix. However, for the PION model also energies outside this range are relevant, because the photons there contribute to heating/cooling/ionisation etc. That is the reason why the energy grid is extended for this model to span the full 1E-6 to 1E6 range. We realise that the 200 bins per decade of that extension is an arbitrary choice, but we think for most cases that is accurate enough to represent the IR or gamma-ray range.

This extension of the energy grid happens in SPEX always, as soon as you read data, even if you do not have the PION model. This is because SPEX cannot anticipate what model components the user may want to add or delete to the full model that is being used. But of course it is harmless for all other models.

You might argue that sometimes you may also want to have higher resolution than 200 bins per decade in the optical or UV range, for instance in those cases where you might have joint HST/COS data with high resolution. Fortunately, whenever the corresponding UV data (response & spectrum) is added, the way SPEX determines its energy grid as described above would include the high-resolution UV part automatically at the desirable high resolution.

8.3.1.3. Checking the energy grid

Whenever you read data (a single spectrum, or from multiple instruments), SPEX creates a new energy grid on which the model is evaluated, before folding with the response matrices. This energy grid is the combination of all energies on the photon axis (not to be confused with the count axis) of the response matrices of all instruments involved, supplemented with a grid with a spacing of 0.005 in log energy spanning between 1E-6 and 1E6 keV.

To see what happens in a specific case, you can plot the model spectrum, put on the x-axis the energy in keV, and choose plot uy bin which gives you the bin number. Also issue the command plot ux kev to re-determine the min/max energies of the grid. This should show the binning of your response matrix/matrices as well as the extension.

Another method to inspect the energy grid is running the egrid save command (Egrid: define model energy grids). This will save the full energy grid to file.