This is the most basic question that should be answered by @arometrix
Side note:
I was just reading my old messages from arometrix and when i asked when the ultra sensor would be released they said “it will be released when we are confident with its performance.”
I dont get the impression you guys are very confident in its performance at this time so i really question decision to release this to the public.
The fraction finder does not work correctly for indicating when to change spd fractions. Cannabinoids inversely correlate to fluorescent signal during distillation. On the other hand Pigment directly correlate to the fluorescent response.
The only reason why the fraction finder MIGHT seem to work SOMETIMES is a phenomenon called turbidity. Turbidity is essentially the inability for light to enter or exit a sample.
Turbidity causes the reading on the fraction finder to go down when too much pigment is present. Essentially the pigments block all of the available light and the reading is none. This is why so much research done on fluorescence are done on DILUTE samples.
When pigments are low (successful processing usually) the reading on the fraction finder indicates that the processing was unsuccessful.
Arometrix got lucky that turbidity lowers and raises signal at almost kind of the right times. It’s dumb luck at best and negligence at worst. Pair that with @spdking and here we are
Not sure if it’s been mentioned, but when using UV light, keep in mind that glass will block UV transmission. Perhaps quartz test vials would help the UV light actually make it to the sample, leading to increased fluorescence of the sample.
You’re right that glass can block UVa. However a lot of glass still transmits something like 80% of the 365nm light and around 80% of the visible light too.
Glass/plastic is really only viable for UVa/VIS 320 nm+
Quartz is optimal for both UV/VIS at 190/200nm+.
If you’re using glass for higher energy emission, it could have a small effect on results. But nothing too significant, just a bit of interference.
I wonder about the geometry of the glass, since this is meant to be strapped to a cylindrical borosilicate tube. Will some of the excitation source refract into the detector?
This is something I’d definitely look into. It does absorb strongly in the UV region the further down you go. So if they’re at 330-350nm+, it shouldn’t really be that big of a deal.
However, my knowledge on glass/quartz absorbance is primarily limited to the use of standard 10mm cuvettes with uniform geometry and pathlength in a UV/VIS spectrophotometer. VERY different situation here. This difference of quartz/glass could be more significant than you would originally think, depending on several factors.
Yep, that IS the technological hurdle for using photons that are waving faster to induce a stronger response. 
Cuvettes are square? Fluorescence spectroscopy needs the excitation light at 90° I believe
The excitation source is active during the measurement so the 90 degree angle is meant to avoid crosstalk between the fluorescent emission and the excitation source.
Seems like the only answer.
source is being pulsed.
We think this dialogue is productive, and asking to see our initial data makes sense. A quick bit on context:
The Arometrix technology is an in situ process monitoring tool, not an analytical benchtop tool. The purpose of this technology is to make process decisions that help improve the manufacturing process. It’s optimal implementation involves putting the unit in line, and empirically observing what is “a good run” versus a “bad run”. Comparing the technology to an HPLC result is interesting, but not entirely proper, as HPLC is an analytical tool, while the Arometrix device is used to give operators information during processing to improve outcomes, i.e. to be able to make more “good runs”. . This is achieved by the operator by adjusting the process parameters in order to see in real time the same patterns and signatures that characterize the “good run”. One of the reasons we chose to use the term “arbitrary units” on our y axis is to highlight the fact that what we are detecting and showing are really trends and patterns rather than giving quantitative cannabinoid measurement of concentrations.
Extraction and distillation process involve a complex mixture of organic compounds. While our research suggests that some cannabinoids fluoresce, so do many other compounds in the distillate and extract. Some are there all the time, but some of them may appear depending on the raw material or other previous processing operations. There are byproducts, fats/lipids, oils, and pigment molecules unrelated to cannabinoids that may fluoresce. Chlorophyll is an example of one of these. Further, fluid clarity and turbidity of the mixture matter… how much signal can actually get through, how much scattering there is and general “cleanliness” of the optical path. Marmoris described this well in his post. Another interesting characteristic we see is that chlorophyll gives a stronger response than other molecules that we typically see in distillation and extraction mixtures.
From our paper, below is the summary of what we tested:
https://arometrix.com/wp-content/uploads/2020/03/In-Situ-Fluorescence-Spectroscopy.pdf
| Chemical | Concentration | Excitation | |
|---|---|---|---|
| Test A | D8 | 1mg/ml | 365 |
| Test B | D9 | 1mg/ml | 365 |
| Test C | CBD | 1mg/ml | 340nm |
| Test D | CBDA | 10mg/ml | 365 |
The paper goes into more detail about specific test standards used and the apparatus we used to test. The test apparatus used identical components to the Fraction Finder product with a few modifications to isolate variables, reduce sample size, and improve signal through optimized optical geometry. We’ve repeated these tests with standard samples over the years, and most recently in April of 2021 with similar results.
There is a big leap that happens from laboratory results to a commercial product. In order to make a product more rugged, buildable and repeatable, there are improvements we’ve made along the way that may have actually improved the response, or hurt the response of the unit in terms of minimum detectable limits. We try to mitigate this through extensive design and testing, but our laboratory is not the field. We find that the Arometrix tech works really well as a process control tool for short path, wiped film and primary extraction where concentrations are relatively high even though the matrix is complex. It works less well for secondary extractions (getting the last bit out of previously run material) and Chromatography, especially when you want to differentiate between CBD and THC.
If it’s not working for you, please call us. Recently we’ve had a few instances where the thing just stopped working, and it ended up being a simple fix. Most of the time it is a warranty repair.
In summary, our tests on HPLC standards have routinely indicated that the Arometrix system detects cannabinoids. More laboratory work is definitely valuable as our internal analytical capabilities are limited. This may be interesting to your process or not, as another process indicator may be more useful.
Talk to us about what you’d like to do and we’ll give you our experience on what we’ve seen and if this technology may be able to help you.
It seems like what you’re saying is the matrix goes in and squiggly lines come out. The squiggly lines aren’t intended to provide quantitative data, but rather to enable your operator to differentiate ‘good runs’ from ‘bad runs’ by changing process parameters so you figure out yourself which squiggles go with the ‘good runs’ and which with the ‘bad runs’.
In other words you know the output changes with the input just not why, and you are convinced you can use it to make better disty but we don’t know exactly how. But if it’s not working for us, please reach out?
Does this mean you used a square quartz cuvette instead of cylindrical borosilicate?
I promise I’ll read the paper tomorrow when I’m supposed to be working.
So this is what you’re giving as a definitive answer. Now the complaint people are making is that they are not getting any signal when they do that. Why do you think it’s not working?
He didn’t say ‘always detects’ and he specifically said plural cannabinoids (i.e. not specific cannabinoid detection)
Should have just put this in the Ad.
Yes, I think it’s fair to say nobody expects it to differentiate between different cannabinoids. But why is it failing to generate any signal at all on isolate concentrations that are supposedly within it’s detection range?
So I’m digging through the paper. The only work done in-situ (in a short path) was using Excitation Emission Microscopy, not using the fraction finder.
The work using the fraction finder was done in a blacked out square cuvette.
I’m honestly not surprised people are having problems. I’m not sure Arometrix knows about the inner filter effect (IFE) or quenching of fluorophores at even moderate concentrations.
[Bonus points for “Peer Reviewed Article” plastered on every page. lol
