June 25: UV-Vis Spectrophotometer Broke

Yesterday morning when I took out the round bottom flask from the sonicator after nanoparticles were formed from the pre-nano solution, dye properly injected, and the entire solution duly mixed by sonication, I noticed that the liquid in the flask (VL 61) appeared more reddish than normal. Usually the dominant species, PFBT nanoparticles, would give the mixture a pale yellow color. A UV-Vis scan after THF was removed confirmed that the nanoparticles were doped with an amount of APESO much more than what was designed according to stoichiometric calculation. My first judgement was that the nanoparticles filtered out, a problem frequently encountered in the early days of this project. But that problem has since been solved through improvements on every detail of the preparation process. The procedure is now greatly streamlined and formularized. It has been proven to reduce the loss of nanoparticles during the process to a bare minimum.

However, now that the “old problem” cropped out again, I resorted back to the old scheme of reasoning that I used to streamline the procedure. The most important factor that affects the size of polymer nanoparticles formed by the re-precipitation method in water is generally understood to be the concentration of free polymer chains in an organic, precursor solution. I suspected that the large volume of APESO dye used during this particular preparation somehow triggered some erratic behavior during nano formation and doping. So I repeated VL 61 with a newly prepared pre-nano solution of the same composition as that used before. This time I injected only 800 uL instead of 1000 into water the form nanoparticles. Proportionately the amount of dye injected was also reduced. But the same low absorbance was observed on the UV-Vis scan again (VL 62).

Then I prepared a blank sample (VL 63), with yet another pre-nano solution of the same composition but completely without doping. Again lower than usual absorbance was observed. Throughout yesterday I did not doubt the reliability of the UV-Vis spectrophotometer despite the fact that the not-too-faint color of my samples did not correspond well with the low absorbance observed by the machine, which would usually mean most nanoparticles have filtered out, leaving a faintly colored, or even clear, solution. I illuminated a small portion of the blank sample with 455 nm blue light and stared at its fluorescence. It also appeared as bright as a normal, concentrated sample. I was baffled but didn’t think for a moment that the UV-Vis could be questioned.

Tolerating a strong sense of defeat I wrapped up yesterday’s experiments and put new polymer solutions on stir just so that I could have a completely fresh start the next day.

So this morning I came in and prepared a sample using the same procedure and recipe. The UV-Vis spectrum of the dye gave an absorbance of 0.043 at 559 nm. I believed in the machine again. Low absorbance (<0.1) again for the nanoparticle peak at around 467 nm (VL 64).

VL 65 is a blank sample with a newly prepared 20 ppm pre-nano solution. The absorbance at 467 nm was half that that of VL 63, the 40 ppm blank sample prepared yesterday.

This time I started to consider drastically changing my way of approaching this issue. I used the fluorescence spectrophotometer to examine my blank samples. It turned out that they both exhibit a level of fluorescence intensity comparable to a concentrated sample, contradicting the UV-Vis spectra that should indirectly measure concentration by Beers Law.

While I was napping during lunch break I thought about a plan to challenge the UV-Vis spectrophotometer, not knowing that the department had another, similar machine. The plan was to inspect an APESO-in-THF sample that will have been prepared to a high enough concentration that I can definitely tell is above 0.1 in absorbance. After that I would use some standard, blue, copper (II) sulfate solution whose concentration I know.

So I followed my plan this afternoon. I made the APESO very concentrated and put it in the UV-Vis. The absorbance at 560 nm was still below 0.1. At this point I was convinced that the machine was wrong. I talked to Sarah about the problem, the first time I had ever expressed my suspicion except for the occasional, cursing-like murmurs that I spitted out yesterday. I suggested my plan of calibrating it with standard copper solutions but she suggested that I talk to Wustholz first. I did that and Wustholz advised me to use the departmental spectrometer to confirm the issue. I did that and after some struggle with the departmental machine I succeeded in proving my judgement. The actual absorbance at 560 nm as was reported by the departmental machine was around 0.3 while our machine gave a low reading of 0.07.

Wustholz contacted Harbron and Harbron placed an order for a new lamp through Ms. Mulholland. Later Molloy came commissioned by Ms. Mulholland to provide more evidence for the potential replacement and to research providers for the purchase. He came and checked the Align application on the UV-Vis computer to find out about whether the problem was really about the lamp. It turned out that the lamp currently installed in the machine had only 0.003% of its lifetime left, as was evident from the extremely weak signal that the lamp was registering on the detector. Indeed the lamp hasn’t been replaced since the machine came 12 years ago.


  1. There is nothing worse than a research delay due to a broken machine. It’s the most annoying thing. Last summer, a necessary machine was broken half the summer and it put me painfully behind on my work. I had nothing I could work on that didn’t require the machine and I basically just twiddled my thumbs for a while. It’s such a waste of time. [Except for the time doing literature research, that was actually valuable.]