# Choosing the right wavelength for calibration curve and for analysis of actual results

Introduction:

I am currently conducting a series of experiments in which I am studying the transport of polystyrene nanoparticles (PNPs) through sand and soil. This is done in sand/soil columns, into which I am injecting the PNPs solution using a peristaltic pump. The PNPs are fluorescent, so I am using a fluorimeter for detection. During the experiment, I am collecting the fluid that is leaving the column in a fraction collector, and as a result I am constructing a breakthrough curve that is supposed to show me the behavior of the PNPs compared to an inert tracer (bromide) that is also found in the inlet solution.

In order to construct the breakthrough curve, I need to prepare a calibration curve that will allow me to convert the intensity of the collected fractions into concentrations. For the calibration curve I am preparing several samples with known concentrations.

Question:

For preparation of the calibration curve, I am choosing the wavelength that gives me the max intensity in each sample's emission spectrum, and then for the breakthrough curve I am taking the intensities in that same wavelength from each fraction for the breakthrough curve. In the example below it is 635 nm.

Then, when I came to take the appropriate intensities of my actual results, I obtained this graph:

The X-axis range is the same: 575-700 nm. I need to choose a certain wavelength and take the data pertaining to it.

Can I actually use these intensities, or would it be wrong because at 635 nm I don't have a max point?

• It looks like you have some material that is suspended in solution and scattering and hence giving a sloping background that is obscuring your data. Is it possible that your sample has decomposed? – porphyrin Feb 17 at 10:05
• Adding to @porphyrin, applying to standards for your calibration and a typical sample containing your NPs eluted: if it takes 2 min to record your spectrum once, does a next spectrum looks reasonably the same (regarding position of the maximum recorded, and counts/s on the detector) if recorded a twice / multiple times just after the initial recording / 10 min later? Different to potential decomposition of your NPs: perhaps your soil samples equally contain a quencher (e.g., oxygen, chloride)? Does tonic water (as a source of quinine) pass the soil samples still a little bit fluorescent? – Buttonwood Feb 17 at 11:47
• Your peak is buried by scattering as @porphyrin suggested. You can see two inflections around the expected maximum. Consider that fluorescence quantitative analysis is tricky. Dealing with np makes it potentially difficult as there can be a concentration depending scattering that shift your baseline continuously though I do not see this in your calibration. It could simply be extremely fine sand. Actually you should make a calibration using water filtered to just sand and see what you get. I suspect something like the current and actual measurements. – Alchimista Feb 18 at 12:07
• Thanks everyone. I tried a few more things in the meantime: I added bromide (same concentration as in inlet solution) to my standards to see if it was interacting with the NPs, but the intensities seem to not be affected (i.e. the graph looks similar to the first graph above). Then I tried adding some sand (clean sand, i.e. sand that has been treated with HCl and lots of water) to the standard samples WITH the bromide. The results of this step were a bit different, since high concentrations gave the bell-shape, while lower concentrations gave the descending, more-or-less linear shape. – Don_S Feb 18 at 12:42
• You should really use water previously passed through the sand while doing the calibration. You should run into similar problems if sand is responsible. Another point I did not want to bother before. Wouldn't better work with absorption? Or absorption of your np stay buried in water cutoff? – Alchimista Feb 18 at 14:11