XII. The Software
As with most modern instruments, software is used to convert raw signals to useful results. As this lab is being written, the software is incomplete. Rather than write pages and pages of documentation (which, if everything's written correctly, should be needed anyway!), let me outline what's there. You can play with it, let me know what does not work (or what works badly), and then we'll rewrite it.
There are 3 graphical insets and 2 control sections. In the upper LEFT of the screen, put the JPG for the SAMPLE (the picture from which you want to find I(λ)). In the upper RIGHT of the screen, put the JPG for the REFERENCE (the picture from which you want to find I0(λ)). I have not designed this for "drag and drop." You must use the directory functions in the middle of the screen to select the files by name. If you regard this as a serious limitation, let me know, and after we make sure the science parts of the software are error-free, I'll try to install that capability.
After the JPGs are in place, you have to tell the software where the spectrum is. Point to either the red or blue end of the spectrum with the mouse. When you click, you'll get to choose whether it's the red or blue end. Once you have both ends marked, you can use the data fields in the lower right of the screen to tell the software what wavelengths correspond to your clicks. You won't know exactly, but give the machine your best estimate. As long as you are consistent between the two frames, your results will be precise, though perhaps not accurate.
"Spectrum width" lets you use more than one row of pixels at a time. How does this influence signal-to-noise ratio? Experiment and find out! Does it influence resolution?
Once you've marked where the spectral data is, you can plot it. In addition, you can export the data as a comma-separated variable (.CSV) file, readable by most spreadsheets. That way, you can improve on the very limited data processing that the software performs (when it computes absorbance, it only ratios intensities without correcting for any of the problems that you have found in looking at the data).
Notice that plots can be as a function of pixel number (bluest pixel, corresponding to the blue wavelength in the lower right of the screen, is numbered 0) or wavelength (interpolated from knowing the dispersion implicit in the data in the lower right entries). Also notice that the graphics window, while fairly small, has the ability, independent of the rest of the program, to save data to a file or print.
XIII. Reprocessing data in Excel or some other spreadsheet
The software described in the previous section will export the red, green, and blue intensity values for each wavelength setting, identifying each wavelength with a pixel number as well. What if you want to get information about a part of the image that isn't really spectral data? Just click into the software a line along which to report intensity values and export those lines of numbers. The software is dumb – it will believe that ANY line of pixels corresponds to a spectrum, whether it does or not! Once the lines of data are in the spreadsheet, you get to take over and figure out how to combine intensity values, subtract background, and so on, to get transmittance or absorbance. From these, you can make a working curve.
XIV. Attempt to make a working curve
Choose a wavelength where absorbance seems to be a strong function of concentration. Make a working curve. Is it a straight line? If not, can you figure out why? Can you obtain data that allow you to combine your extra information with the raw I(λ) and I0(λ) to make it straight (or, at least, straighter)? If you have an unknown, determine its concentration. CHALLENGE: how in the world do we determine precision or accuracy with this system?