In this first step we will try if we can reproduce the measured transmission spectra with a model assuming a single homogeneous silver layer on glass. We have already developed such a model in tutorial 1, example 3. To make a quick start, we simply use that model. Load the SCOUT configuration called tu1_ex3_step3.sc2 which you should find in the directory of tutorial 1.

Since we are now dealing with transmission instead of reflection spectra, we have to make a few changes. Open the spectrum list and rename the spectrum that was called 'Reflectance' to 'Transmittance'. Open this spectrum and set the spectrum type to transmittance. The angle of incidence must be changed to 0°. Then select the Recalc command to recompute the model spectrum.

Now we Import (using standard format) the transmission spectrum of the silver layer with 10 s deposition time. The file is called 10.std. The spectral range of the measured data is 1.1 ... 5 eV. Please check the unit in the range dialog that is displayed after the data are imported. It must be eV. Now you are probably warned that the spectral units of measurement and simulation do not match. Change the spectral range of the simulation with the Range command to 1.1 ... 5 eV also. After a proper setting of the graphics parameters you should now have a picture similar to this one:

Now you can start the fit with the Start command in the main window. The only fit parameter is the silver layer thickness. The agreement between simulation and measurement is quite good and the thickness is found to be 12.7 nm:

This nicely confirms the sputtering rate of 1.3 nm/s that was obtained in tutorial 1, example 3.

Now load the spectrum 5.std that contains the transmission spectrum of a sample after 5 s sputtering time. The best fit is this one:

This is still ok with a thickness of 6.5 nm.

With 4.std the first problems arise:

You can easily check that all spectra obtained on samples with sputtering times less than 5 s cannot be described with a model of a homogeneous silver film. In the following we will investigate in particular the 1 s spectrum 1.std. Here the 'best' homogenouse fit is this one:

Save the current program configuration and proceed to the next step.