Question: Can DIY spectrometers be useful for pH determination and/or dissolved organic carbon of a water solution?

alejobonifacio is asking a question about spectrometry
Follow this topic

by alejobonifacio | March 02, 2022 15:12 | #30090


These questions have to do with certain things that I would like to achieve with DIY spectrometers. The first thing that I want to achieve is to determine the pH of a solution with purple cabbage juice. There is a research paper that tells that the pH could be accurately measured at 533 nm with absorbance values for pH values between 2 and 11, which is a wide range. I was doing some tests in the spectral workbench and noticed that for each wavelength there are 3 values, red, green and blue. Are there some ways to change these values to absorbance measurements? because, in the paper use this unit measurement.

The other thing that I would like to achieve is to determine the dissolved organic carbon (DOC) of a water sample also with a spectrometer. This is a good tool for community science because you need no reagents. The only thing that I don't know if it is possible with DIY spectrometer is to measure in the UV wavelength, because the DOC is measured at 254 nm.

I hope someone have some time to answer these questions.

Cheers!



19 Comments

This seems like a great application and very feasible. Indeed the color/sensitivity range of anthocyanin in cabbage juice is really huge. 533 is in the green range, right? I'm a little skeptical that just absorbance in one color is the best way to go; it sounds like they just had a spectrometer, and traditional scientific methodology is to find one wavelength and try to build knowledge within it. That's partially just how old spectrometers are built rather than it being the ideal approach. Cabbage juice changes color across a wide range of colors, so my intuition is that actually measuring the hue would be a better match. The danger with just one color is that we lose sensitivity when the hue change is occurring far away from the chosen wavelength -- maybe the dynamic range is worse at the extremes. That might kind of explain why they chose a color roughly in the middle of the range, to try to have no hues too far away from the measurement wavelength.

So that makes me think - what about just hue measurement using a controlled angle/color/intensity of light and a camera? Or just use the spectrometer and measure the relative hue shift by looking across the whole spectrum? Could you find the "median" hue of a spectrum? I could imagine a standard equation for this.

I was doing some tests in the spectral workbench and noticed that for each wavelength there are 3 values, red, green and blue. Are there some ways to change these values to absorbance measurements? because, in the paper use this unit measurement.

The way to do this is to develop a scale through testing. So pre-mix some pH solutions bracketing your ideal range - from 2 to 11, maybe... 5 different values? And scan them, then to copy the paper, you'd look at the intensity at 533. You should be able to establish a scale, and it may not be linear. But since the spectrometer is not measuring absolute units of brightness, you need to use this technique to scale your data to a set of reference values.

The other thing that I would like to achieve is to determine the dissolved organic carbon (DOC) of a water sample also with a spectrometer. This is a good tool for community science because you need no reagents. The only thing that I don’t know if it is possible with DIY spectrometer is to measure in the UV wavelength, because the DOC is measured at 254 nm (Evaluation of Specific Ultraviolet Absorbance as an Indicator of the Chemical Composition and Reactivity of Dissolved Organic Carbon | Environmental Science & Technology (acs.org)).

254 is really low. LEDs we've bought have gone only as low as 360. I have found some "UVC" bulbs for disinfecting stuff down lower. Just looking now, it seems like COVID has helped us, both LED and quartz bulbs:

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Hi @warren! maybe you have something to say about this... 😃

Reply to this comment...


Sorry, but I only have access to the abstract. The wavelength mentioned there is 633 nm. Is the correct wavelength 633 nm or 533 nm?

Regards

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


I got one of the 254 nm UVC lamps that might be usable for an uv/vis instrument as a gift . It's advertised as for sterilization of surfaces. The one purchased was" fbfl for portable uv". It's a mercury lamp in sheep's wool. It's powered by three "AA" batteries. And the mercury lamp is the smallest I've ever seen. There are many precautions that need to be taken with UVC light, including eye protection. Mercury disposal(as well as mercury breakage) will also be an issue. The led version out there might be better. But still, be careful using any of them.

Definitely these can be dangerous for the eyes. The LEDs seem nice because they are a little less 1800s feeling than those tiny tiny mercury bulbs!! Energy usage and probably consistency of brightness as well, i'd guess? Though I don't know mercury lines in that range, do you think there are advantages to spectral narrowness too? LEDs are usually not as narrow range, but I know very little about deep UV LEDs.

I'm very curious about lead detection with deep UV. A long time ago I used a toothbrush sterilizer to try to see lead residue on a bullet casing and it didn't glow at all. But I do see the references (https://libanswers.cmog.org/faq/143932) to "icy blue" fluorescence in glass due to lead. I'll post a question about that too...

OK here i posted it separately: https://publiclab.org/questions/warren/03-02-2022/using-uvc-or-deep-uv-light-to-detect-lead-by-fluorescence

Is this a question? Click here to post it to the Questions page.


Reply to this comment...


On battery life- it wasn't very good.
Mercury (as in mercury lamps) have a lot of lines that go all through the uv/vis spectrum. For a while(as in late 1970s), it was used as a source for hplc(high pressure liquid chromatography) uv/vis detectors. Trouble is, you have to filter out the ones that aren't needed. And there are a lot.

Reply to this comment...


Going back over this. We did a lot of work with total organic carbon (TOC). According to Dr. Paul Whitehead at egalabwater.com (the blog title is "let's talk about lab water" and the specific article is "TOC and it's measurement" ), TOC equals DOC plus NDOC(non-dissolved organic carbon). There are a number of other interesting things there, too. So the base instrument is a TOC. With the TOC instrument we used, reagents could be required, depending on purity of the water. Dirty water would need reagents. Relatively clean water, probably not. At the time, the instrument was relatively expensive (about $20k with autosampler). Hopefully, the price has dropped. Calibration and reagent changing was also moderately expensive.

Reply to this comment...


Previous methods could be somewhat hazardous. To give an example, ammonia related chemicals often work their way into drug compounds. It can be for the stoppers, the purified water, the drug vehicle itself, or many other places. For a long time, the routine test for ammonia nitrogen used Nesslers reagent. The problem- Nesslers reagent is mercury based. So the test for ammonia nitrogen would generate a bunch of mercury - which had all kinds of environmental issues.

Reply to this comment...


If you want an overview of TOC, Wikipedia is a good start. The current article seems pretty thorough ( dated 19 July 2022 by 75.143.92.215). Any carbon compound in solution are oxidized to carbon dioxide by several possible methods. The one that is most lab friendly is UV oxidation(often with a 254 nm. lamp) , usually with persulfate oxidizer, but sometimes with straight dissolved oxygen.

Reply to this comment...


The CO2 produced can be measured several ways. Each method has advantages and disadvantages and depends on the final use of the instrument. For example, if using the instrument in a medical device/pharmaceutical field, the vendor should verify the instrument meets the requirements of the "USP/EP/BP" and so forth.

Reply to this comment...


USP/BP/EP means United States Pharmacopoeia,BP is British Pharmacopoeia, and EP is European Pharmacopoeia. Sections of the USP are divided into monographs. The monograph of interest is <1231> Water for pharmaceutical purposes.

Reply to this comment...


Getting a hold of the USP used to be difficult and expensive. While current USP may still be expensive, here is a source for older versions. "Pharmacopeia Online" is the place to go. Put in USP31 and then search for <1231>. This is still a very large monograph. Go to the subtitle "Chemical Considerations" for explanations of why chemical tests were dropped and TOC and conductivity was added. As a subnote, the increased sensitivity of TOC has shown some extractables from Objective-C noted in this paragraph.

Error here. It should not be objective c, as noted above. It should be some plastics are showing higher extractables than bulk water. Something to keep an eye on, depending on the drug being used with the water.


Reply to this comment...


As a refresher, read USP 31, <1231>"Water for Pharmaceutical Purposes". This is available, for free, from Pharmacopeia online, as mentioned earlier. The part we are after is where it mentions "Types of Water". There you will find frequent mention of Bacterial Endotoxins <85>.

Here's where life gets a bit interesting. There are substances called pyrogens. These cause a fever, sometimes severe, even if the carrier bacteria is dead. The issue for the medical industry- the test says the product is sterile, but it still may pose a medical issue. The best way to test for pyrogens is L.A.L. (Limulus amebocyte lysate). This uses the blood from a horseshoe crab (see Wikipedia article titled "Limulus amebocyte lysate" dated 13 September 2022 for more details). A subset of pyrogens are Bacterial Endotoxins<85>.


Reply to this comment...


The Wikipedia article is an interesting read. The initial data said horseshoe crab mortality from L.A.L. testing should be about 3%. However, recent studies have shown the crab mortality value to be 15% or even 30% (read the Wikipedia article cited above). After studying an alternative (recombinant factor C), the USP decided to stay with Horseshoe crab LAL.

Reply to this comment...


"Pharmacopeia online" is at " uspbpep.com".

Reply to this comment...


There are variations on the method, but check out the USP if using for medical applications. The oldest is a straight " gel clot". The sample is compared to standards that look like a faint white clot to a severe white clot. The turbidimetric was better in quantitation, but both the gel clot and turbidimetric took a fair amount of time to run. Maybe the newer methods are better. Never observed the chromogenic method. Here is a quick presentation by Tim Sandle of Bio Products Labs to get you started :https://www.researchgate.net/publication/299507542_LAL_Choice_of_Test_Method

Reply to this comment...


Now, to change the topic a little bit. What happens once all this testing gets going? In industry, a financial auditor appears and says " you are spending too much on testing". So the usual approach is to check the data. Do you have failures? If the answer is no, frequently the company will do something like skip lot testing. in skip lot testing, only one in say 20 lots is tested. This allows fewer tests to be done and ( theoretically) doesn't alter quality. A few months later, the next step is taken. Only one in 100 lots will be tested. And so on and so on. This is often accompanied by loss of personnel, etc. When there is a failure, trying to nail down the exact lot of failure is difficult. Because the samples tested are so few, large recalls must be done. The environmental work locally functioned in the same manner, with tests often done at long intervals. A screening test might pick up problems earlier.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Log in to comment