Working on Automatic Tester, need help with preparing reagents

[Tariq-Shiwani]

Curious why you're worried about total color change amount? Can't the photo sensor pick up any change?

Yes it can, so far I've tried with TCS3472 color sensor which reads RGB.

my experience with this sensor is, it does not produce the same result everytime, infact if I read the value twice with a gap of 1s I get slightly different number (yes the device is complete isolating the cuvette to avoid environmental light to interfere) could be my programming error or most likely the LED flicker is causing this because I tried with PWM dimming to read the results with different brightness of light, I will have to test with full brightness too and with different methods.

if the color has a significant change then it should be fairly easy to distinguish, if there is a minor difference between two colors of major difference in values, a slight error in sensor readings would be mistaken for wrong results. I need to produce repeatable results with the same sample.

However I am yet to test it with a photodiode to see the results. And because we are getting multiple colors instead of linear increment in a single color we may need to use color sensor that reads RGB so we can tell the color difference too, it looks like the calibration curve is going to be pretty complex.

Will require testing with multiple wavelengths of LEDs also to see which one works best with which color.

But yes you are right the device should be able to accurately read the color values even if they are not clearly different to a human eye. Even if I am able to produce significant color difference with this reagent, this is not going to be the case with all the reagents.

 

[Tariq-Shiwani]

Hi

I am about to try using the method outlined in this paper for nitrate analysis: A simplified resorcinol method for direct spectrophotometric determination of nitrate in seawater, Jia-Zhong Zhang, Charles J. Fischer, Marine Chemistry,2006 . It is copyright but findable online. It differs from the commonly used methods in that it does not require a metallic reducing reagent and involves the simple addition of resourcinol and sulphuric acid. The downsides as published from autoanalyser perspective would seem to be the use of conc acid, maybe that use could be modified, plus that resourcinol solutions deteriorate with storage. On the face of it it looks attractive as a lab method in that it avoids cadmium column preparation, if it works. I also saw a nice way to make a cadmium column while researching nitrate, involves using Cd 1mm wire in about 1m of thin ptfe tubing, sounds easier than using filings …if the wire could be sourced or made, that’s the tricky bit.

A good read is the classic: A Handbook of Seawater Analysis, Strickland - scanned copy easily available online, lovely old book. I found an old copy in print, nice book to have.

Strikes me that nitrate may not be the easiest test to start with here .

Steve

Thanks, will look into those books too, please do let me know your outcome too.

Thank you

 

[WillpoleReefers]

Thanks, will look into those books too, please do let me know your outcome too.

Thank you

Will do. I prepared nitrate standards and am ready to go with it. Just so many demands on my time setting up a tank

 

[taricha]

(On Nitrate, a packed cadmium column sounds fun, but I can't actually imagine any scenario where the accuracy need for aquarium NO3 is so great as to justify messing with Cadmium toxicity.)


anyway...

Hmmm, a little improvement??

It looks like what you did here is take samples (equal volumes?) made up to different known alk (in dKH) and add a fixed amount of indicator (bromocresol green?) and a fixed amount of acid to each sample, so they go to different colors.
Do less acid to shift the color sweet spot down a bit - people won't care very much about the difference between 13 and 17 dkh. but many people would care more about differences close to 7dKH, and it looks like that may be where you are fully yellow and have no color variation left to play with.
What does your digital eyeball say in terms of raw data about the colors of those samples? If your digital eyeball is good enough you can certainly make a calibration curve from that color range. (If not, you may need to rework your digital eyeball - replace with an iorodeo colorimeter or similar.)

 

[Randy Holmes-Farley]

Curious why you're worried about total color change amount? Can't the photo sensor pick up any change?

No. Any photometer has a lower limit, and that may well limit your lower end of accurate detection. Also, background color in the water becomes more and more significant, leading to errors if it is not subtracted somehow.

 

[Tariq-Shiwani]

(On Nitrate, a packed cadmium column sounds fun, but I can't actually imagine any scenario where the accuracy need for aquarium NO3 is so great as to justify messing with Cadmium toxicity.)


anyway...


It looks like what you did here is take samples (equal volumes?) made up to different known alk (in dKH) and add a fixed amount of indicator (bromocresol green?) and a fixed amount of acid to each sample, so they go to different colors.
Do less acid to shift the color sweet spot down a bit - people won't care very much about the difference between 13 and 17 dkh. but many people would care more about differences close to 7dKH, and it looks like that may be where you are fully yellow and have no color variation left to play with.
What does your digital eyeball say in terms of raw data about the colors of those samples? If your digital eyeball is good enough you can certainly make a calibration curve from that color range. (If not, you may need to rework your digital eyeball - replace with an iorodeo colorimeter or similar.)

Nitrate: I will be using copper to reduce nitrate to nitrite (also planning to try zinc) and then the griess reagent will do its job, however I still have to figure out the desired ratio of sulfanelamide and N-(1-naphthyl)ethylenediamine dihydrochloride.

Alkalinity: yes I took 5 samples of known alkalinity (measured by Hanna) and then added my reagent to it.

Solution A:
Bromothymol blue 0.04% (mixed 0.1gm in 250ml of RODI)
Solution B:
5% Phosphoric Acid diluted further by taking 1ml acid and mixing with 100ml of water

Reagent:
Mixed A and B at 6:4 ratio

Testing:
took 4ml of each sample added 0.4ml of reagent to each of them and got these results

sorry about providing quantities in ml and gm but this is what we use here, always have problems with converting them to imperial system

yes you are right people wont care much about difference between 13 and 17, I have to narrow down the range and try to produce color intensity under 13.


now the digital eyeball (I liked the name btw ) well its a color sensor TCS3472 which appears to be reading different values even if the cuvette is not removed and I keep on reading the color every x second, the device is covered and there is no stray light making into it, so cuvette, color sensor and the led all are isolated from environmental light.

I will be trying again with a photodiode to see how it is coming along, we may not need to identify the color in this scenario maybe just a grayscale gradient will work, so still doing my research on that part.

 

[WillpoleReefers]

(On Nitrate, a packed cadmium column sounds fun, but I can't actually imagine any scenario where the accuracy need for aquarium NO3 is so great as to justify messing with Cadmium toxicity.)

Certainly a consideration when I chose to investigate a different method.

On scenarios where high accuracy is valuable - I guess not in totally stable tank with stable dosing requirements, feeding whatever - but accuracy beyond test kits to me is incredibly valuable in a situation where the parameter is changing or you wish to change it. Resolution allows you to see changes over a day rather than a week or more?

Personal experience so far for accurate methods is using a burette- based alkalinity titration. I can see changes by the day if dosing is changed, might take a lot longer to draw conclusions with a basic test kit. If I chose to try and modify my nitrate level I am sure accurate analysis would be useful.

Steve

 

[taricha]

I will be trying again with a photodiode to see how it is coming along, we may not need to identify the color in this scenario maybe just a grayscale gradient will work, so still doing my research on that part.

this is closer to how colorimeters like hanna do it. The detector part is stupidly simple - relative light intensity only, the selectivity is which wavelengths of LED light get shined through the sample.
So the detector doesn't have to think about color, just amount of light absorbed, and the choice of LED governs the wavelength at which you make that measurement of blocked light.
If you shine white light through, then your detector has to work harder/smarter to figure out what's going on.

 

[Tariq-Shiwani]

Nitrate

Not completely done yet, but getting closer

 

[Tariq-Shiwani]

ok so here is the explanation on this:

Solution A:
10gm Sulphanilamide mixed in 100ml of HCI then made up to 1L

Solution B:
0.5gm N-(1 -naphthyl)-ethylenediamine dihydrochloride mixed in 1L DI water
these two solutions are supposed to test NO2
took 4ml of water sample, added 1 drop of solution A, stirred for 1 minute and then added 1 drop of solution B, the sample turned a little pink, not sure why because the reagent is supposed to test NO2, so assuming there is some nitrite present in my tank water, I never tested NO2 and the tank is more than a year old. anyway, dipped a toothpick in zinc dust and dipped it into the sample reduction process occured almost instantly and assuming NO3 was reduced to NO2 resulting in developing much more enhanced color, to verify if it reduced NO3 to NO2 I added a few drops of Potassium Nitrate to the sample and it turned more intensed pink, so its proved that it is working with NO3.

now the Proplem
as mentioned earlier I am planning to make an automated tester, I wont be able to add any powders to sample, I tried following:

Attempt 1:
mixed zinc dust in water (left it overnight to completely dissolve it)
added few drops to sample to see if it reacts, no luck

Attempt 2:
mixed zinc with solution A (took few hours)
use that instead of regular solution A, no luck

Attempt 3: (edited)
mixed zinc dust in HCl, it dissolved instantly, added a few drops of it after adding solution A and B but didn't react with sample, assuming the concentrated HCI altered the properties of zinc.

Attempt 4:
mixed with solution B (took few hours but not completely dissolved, there is a residue at the bottom of the beaker)
when used this instead of regular B didn't have any impact on NO3 reduction, when swirled the beaker and used that solution it did work.

now the Quesion is:
how do I completely dissolve this zinc dust into solution without compromizing the solutions performance?

Note for later determination:
if you keep adding more zinc the solution starts to loose the pink color and turns transparent(not really transparent as the zinc makes the sample dirty so there is some grayish shade because of zinc, but if you keep it for longer period the zinc settles down at the bottom and the sample starts to look transparent.


@Randy or any chemist here please help.

 

[taricha]

now the Quesion is:
how do I completely dissolve this zinc dust into solution without compromizing the solutions performance?

yeah. this is the problem - you don't.
To reduce the NO3->NO2 and thus detect it, you need zinc in metal form. Not dissolved in solution. You've demonstrated pretty well what I'd expect. That when you actually dissolve zinc powder it stops working.
Every hobby test (I'm aware of) has zinc metal powder in it - if you look hard enough.
API is the only formulation I know of that leaves the zinc metal powder suspended in liquid - you have to shake the liquid vigorously before you dispense drops.

Like I said, Nitrate test is so filled with complications

 

[Tariq-Shiwani]

yeah. this is the problem - you don't.
To reduce the NO3->NO2 and thus detect it, you need zinc in metal form. Not dissolved in solution. You've demonstrated pretty well what I'd expect. That when you actually dissolve zinc powder it stops working.
Every hobby test (I'm aware of) has zinc metal powder in it - if you look hard enough.
API is the only formulation I know of that leaves the zinc metal powder suspended in liquid - you have to shake the liquid vigorously before you dispense drops.

Like I said, Nitrate test is so filled with complications

And as I said I would prefer to do the complicated fist so the rest will become easy.

I will try sulfanilic acid instead of sulfanilamide because it also acts as reductase for NO3 and is available in liquid form


Anyway I've done my research on Ca / Mg and figured that they are determined together by computing total hardness first and then calcium hardness and then subtract calcium hardness from total harness to determine mag.

And the ingredients used in the tests are:
EDTA (Chelating Agent)
EBT (Total Hardness Indicator)
Murexide (Calcium Indicator)
Sodium Hydroxide (Buffer)

Now I understand, why trident doesn't allow you to test cal or mag or alk separately, specially when testing cal and mag it must test alk too.

I am assuming, to keep the reagent bottles as minimum as possible they have probably put each indicator of hardness in reagent B and C along with the buffer and EDTA in reagent A along with pH indicator, which probably act as acid for alk test too while helping in titration for Cal and Mag.


I wish I had studied chemistry a little bit, fortunately or unfortunately I was a commerce student

 

[WillpoleReefers]

And as I said I would prefer to do the complicated fist so the rest will become easy.

I will try sulfanilic acid instead of sulfanilamide because it also acts as reductase for NO3 and is available in liquid form


Anyway I've done my research on Ca / Mg and figured that they are determined together by computing total hardness first and then calcium hardness and then subtract calcium hardness from total harness to determine mag.

And the ingredients used in the tests are:
EDTA (Chelating Agent)
EBT (Total Hardness Indicator)
Murexide (Calcium Indicator)
Sodium Hydroxide (Buffer)

Now I understand, why trident doesn't allow you to test cal or mag or alk separately, specially when testing cal and mag it must test alk too.

I am assuming, to keep the reagent bottles as minimum as possible they have probably put each indicator of hardness in reagent B and C along with the buffer and EDTA in reagent A along with pH indicator, which probably act as acid for alk test too while helping in titration for Cal and Mag.


I wish I had studied chemistry a little bit, fortunately or unfortunately I was a commerce student

I am using murexide for Ca titration , one thing to be aware of is that once made solutions deteriorate. The standard way to use the indicator is as a powder ground up with sodium chloride to dilute it. I use it as 1% in NaCl, adding a tiny scoop (size as per typical test kit scoops) before each titration. My recipe to measure Ca (plus Sr) is 6ml 1M NaOH plus 10ml seawater sample, all diluted to 50ml with RODI. I titrate with 0.01M disodium EDTA. 400 ppm Ca = 10ml EDTA. (ie vol EDTA x 40). Strontium error averages approx 4ppm with this method.

Mg is then done by combined Ca/Mg titration using Eric chrome black and a pH 10 ammonia/ammonium chloride buffer instead of NaOH and 0.03M Na2EDTA. Again the salt diluted indicator is stored dry.

No idea how long these indicators last in solution

Steve

 

[monfilsi]

@Tariq-Shiwani What a fun and ambitious project! Can't help much on the chemistry side, but I can offer some assistance on the electrical design side and optics side. I am an electrical engineer who specializes in optics. What LED are you using? The spectrum of the led will make a significant difference to the colors your sensor will observe. Low noise transimpedance amplifiers after the photodiode will help with sensitivity and have you measured the rise time of your photodiode to know when your source is stable?

 

[Tariq-Shiwani]

@Tariq-Shiwani What a fun and ambitious project! Can't help much on the chemistry side, but I can offer some assistance on the electrical design side and optics side. I am an electrical engineer who specializes in optics. What LED are you using? The spectrum of the led will make a significant difference to the colors your sensor will observe. Low noise transimpedance amplifiers after the photodiode will help with sensitivity and have you measured the rise time of your photodiode to know when your source is stable?

Thanks @monfilsi


I've only tried it with the TCS3472 sensor, didn't use it's built in LED as I wanted the light to pass through the cuvette, so I installed an LED on the opposite side.

So far tried with 540nm and 6000k

The problem is that I was not getting consistent results, I have done many different experiments with different colors and even the intensity of LED, the device was on continuous read i.e. after every integration interval but the numbers gradually start to go up, there are several different scenarios, I am planning to create a separate thread for the device where I will also post some pictures, circuit diagrams and all the related information, that would probably better place to discuss the device in detail.


As you can see I've become a chemist these days my inner electronic engineer is sleep

Wanted to be done with reagent part so I can then focus on the device side.

And thank you so much for offering your support.

 

[Tariq-Shiwani]

I am using murexide for Ca titration , one thing to be aware of is that once made solutions deteriorate. The standard way to use the indicator is as a powder ground up with sodium chloride to dilute it. I use it as 1% in NaCl, adding a tiny scoop (size as per typical test kit scoops) before each titration. My recipe to measure Ca (plus Sr) is 6ml 1M NaOH plus 10ml seawater sample, all diluted to 50ml with RODI. I titrate with 0.01M disodium EDTA. 400 ppm Ca = 10ml EDTA. (ie vol EDTA x 40). Strontium error averages approx 4ppm with this method.

Mg is then done by combined Ca/Mg titration using Eric chrome black and a pH 10 ammonia/ammonium chloride buffer instead of NaOH and 0.03M Na2EDTA. Again the salt diluted indicator is stored dry.

No idea how long these indicators last in solution

Steve

Thanks @WillpoleReefers I will test with your method too when working on Ca/Mg

Is there a way to add some preservatives to all the reagents without compromising their original role?

 

[Randy Holmes-Farley]

Thanks @WillpoleReefers I will test with your method too when working on Ca/Mg

Is there a way to add some preservatives to all the reagents without compromising their original role?

What do you want to preserve them from?

If they are dry, air oxidation is the main issue. Sealing them well is the best way. Hach uses metal foil for this purpose. Moisture also facilitates the oxidation. In a few situations, changing physical form might be an issue, if a more soluble crystal form slowly converted to a less soluble one. Moisture sometimes facilitates that conversion. Pharmaceuticals sometimes include oxygen or moisture absorbing devices in prescription bottles for these reasons..

Some liquid reagents might be low enough in concentration and have other characteristics that may allow microbial growth. How to deal with that would be very case by case.

In a very few instances, light (typically low wavelength UV or blue) can cause breakdown of chemicals. Dyes might slowly break down when absorbing their specific wavelengths since that can put them into an excited electronic state that may be more prone to degradation. .

 

[Tariq-Shiwani]

Some liquid reagents might be low enough in concentration and have other characteristics that may allow microbial growth. How to deal with that would be very case by case.

Yeah I might need to deal with that to extend the shelf life once the reagent is prepared.

Once I am done preparing my reagents, this will be the next task

 

[taricha]

I will try sulfanilic acid instead of sulfanilamide because it also acts as reductase for NO3 and is available in liquid form

might be a little confusion here.
sulfanilic acid is used as a reagent in lab tests (and some hobby kits) to react with nitrite - that is looking for nitrate reduction - that is to see if the biological sample has reduced nitrate to nitrite.
https://www.austincc.edu/microbugz/nitrate_reduction.php

So if it turns red, then there is nitrite - NO2 present, and if no red color - then the nitrate - NO3 in the culture broth was NOT reduced to nitrite - and thus sulfanilic acid didn't do anything to the nitrate.

If it gives no color with the sulfanilic acid, then a bit of zinc powder is added and it goes red. Proving that there was a bunch of nitrate, and the microbes in the sample didn't turn any into nitrite - NO2.

One way others have gotten around the difficulty of dispensing solids is to do suspensions in liquids, and use magnetic stirrers to spin up the suspension before they dispense the liquid with their automated system.

Seems easier than trying to find/invent an all-liquid, sensitive, quantitative, saltwater nitrate test.

 

[Tariq-Shiwani]

might be a little confusion here.
sulfanilic acid is used as a reagent in lab tests (and some hobby kits) to react with nitrite - that is looking for nitrate reduction - that is to see if the biological sample has reduced nitrate to nitrite.
https://www.austincc.edu/microbugz/nitrate_reduction.php

So if it turns red, then there is nitrite - NO2 present, and if no red color - then the nitrate - NO3 in the culture broth was NOT reduced to nitrite - and thus sulfanilic acid didn't do anything to the nitrate.

If it gives no color with the sulfanilic acid, then a bit of zinc powder is added and it goes red. Proving that there was a bunch of nitrate, and the microbes in the sample didn't turn any into nitrite - NO2.

One way others have gotten around the difficulty of dispensing solids is to do suspensions in liquids, and use magnetic stirrers to spin up the solution before they do the automated liquid suspension.

Seems easier than trying to find/invent an all-liquid, sensitive, quantitative, saltwater nitrate test.

Yeah I understand the logic of stirring the reagent before dispensing. But that would require adding a stirrer under the reagent container.

I am wondering how mastertronic or reefMaster SPA of kamoer is doing it.