[curiouscat]

I was having some suspicion that our lab was messing up the COD analyses so I had them prepare & report CODs for standard solutions of Methanol. (chosen only because it was convenient; I don't see a consensus reagent to standardize COD analyses)

One solution was 4800 ppm MeOH aq. They reported a COD of ~25,000 ppm. To me that's bizzarely high.

Just wanted to see if anyone can confirm my calculations. According to my calculations 4800 ppm MeOH ought to lead to a theoretical COD of 7200 ppm. In practice if all of it isn't oxidized then I might expect the experimental value to be lower but never higher. Right?

[Archer]

Can you provide your protocols for the COD method that you use and include the calculations?

We routinely use the following method and use premade standards

www.fergusonscientific.co.uk/Reagecon%20COD.pdf

[curiouscat]

Can you provide your protocols for the COD method that you use and include the calculations?

Ok, I will send. Basically we go via the Ferrous Ammonium Sulphate titration of acidified Pot. Dichromate route. Ag as catalyst for oxidation & Hg salts to suppress Chlorine interference.

[curiouscat]

Although my core question is protocol independent: What's the theoretical COD from some known amount of MeOH. I can imagine different protocols yield some lower COD value depending on how completely they can oxidize the organics. But I cannot imagine any protocol exceeding the theoritical demand.

As an aside, any idea what reagent is in your premade standards?

[Archer]

I shall check what is in the standards today.

I will review our method and see if I get the same figures from methanol as you got.

What was the blank reading?

If you are routinely running an analytical test, as part of your QC your analysts should be using a certified reference material which is separate from any certified calibration standards you are using (i.e. separated upon delivery) at the beginning, middle and end of a batch of tests.

[curiouscat]

If you are routinely running an analytical test, as part of your QC your analysts should be using a certified reference material which is separate from any certified calibration standards you are using (i.e. separated upon delivery) at the beginning, middle and end of a batch of tests.

I agree. But our ground reality is a bit different & messier. I wish it wasn't but that's that. 

[Archer]

Our QC standard is potassium hydrogen phthalate, I though this was a premade certified solution but it is the dichromate / silver which is pre-made.

Anyways the potassium hydrogen phthalate is dried at 120°C for 2 hours and then cooled in a dessicator and made up to 340 ppm , this should provide 400ppm O₂.

I can't engage my brain into the calculation at the moment but I will when I have a quiet moment later on.

[curiouscat]

Anyways the potassium hydrogen phthalate is dried at 120°C for 2 hours and then cooled in a dessicator and made up to 340 ppm , this should provide 400ppm O₂.

Makes sense.

2 C8H5KO4 + 15 O2 = 16 CO2 + 4 H2O + 2 KOH

MW = 204

Hence 1 ppm KHP = 1.176 ppm Oxygen.

[Archer]

Makes sense.

2 C8H5KO4 + 15 O2 = 16 CO2 + 4 H2O + 2 KOH

MW = 204

Hence 1 ppm KHP = 1.176 ppm Oxygen.

I am not doing maths today beause my head hurts so I'll take your word on that

This is taken from a standard method used by many accredited laboratories in the UK so give your analyst's some of that and see how they get on.

I agree. But our ground reality is a bit different & messier. I wish it wasn't but that's that. 

I used to have to have this issue when I worked in an unaccredited lab, if I got an anlmylous result I resampled and either ran it again or got a colleague to do it. Although the paperwork that goes with accreditation is extremely tedious, when you are presented with an unusual result you have much more confidence in it.

[DrCMS]

I agree that 4800ppm methanol should be 7200mg/L COD.

But I disagree with Archer that I'd have more faith in an accredited lab not giving a wrong answer. 

My experience of every QC lab I've ever dealt with in any capacity is that they very rarely 2nd guess their own work.  Analysts trust the numbers they get even if a simple calculation can prove that number to be impossible.   I can not count the number of times with internal and external QC lab data that I have had to take the time to pick that data apart to show that at least one result must be wrong.

[curiouscat]

I agree that 4800ppm methanol should be 7200mg/L COD.

Thanks DrCMS! That gives me confidence.

But I disagree with Archer that I'd have more faith in an accredited lab not giving a wrong answer. 

My experience of every QC lab I've ever dealt with in any capacity is that they very rarely 2nd guess their own work.  Analysts trust the numbers they get even if a simple calculation can prove that number to be impossible.   I can not count the number of times with internal and external QC lab data that I have had to take the time to pick that data apart to show that at least one result must be wrong.

I don't have much experience to comment but my overall opinion of (industrial) QC guys I've dealt with is quite low. Many I've met are script monkeys. Not knowing why they are doing something & blindly adhering to a system. That can be good on routine matters but not always. It's especially bad when a non-routine analysis comes their way.

In my grouchy way I attribute a part of it to the whole alter of ISO / validation worship. People have come to respect the system more than the result itself. Sometime they go through this fascinating routine of to-the-book paperwork etc. and magically expect the answers to be right.

On many a bad day I've wished I could just have an old school chemist who knew what he was doing, screw all the paperwork and fluff and ISO inspectors and gleaming certificates.

Frankly, the ISO-9000 etc. has diluted their utility to an extent where when someone tells me they are accredited I no longer care much.

/rant

[Archer]

I don't have much experience to comment but my overall opinion of (industrial) QC guys I've dealt with is quite low. Many I've met are script monkeys. Not knowing why they are doing something & blindly adhering to a system. That can be good on routine matters but not always. It's especially bad when a non-routine analysis comes their way.

In my grouchy way I attribute a part of it to the whole alter of ISO / validation worship. People have come to respect the system more than the result itself. Sometime they go through this fascinating routine of to-the-book paperwork etc. and magically expect the answers to be right.

On many a bad day I've wished I could just have an old school chemist who knew what he was doing, screw all the paperwork and fluff and ISO inspectors and gleaming certificates.

Frankly, the ISO-9000 etc. has diluted their utility to an extent where when someone tells me they are accredited I no longer care much.

/rant

I've been an analyst and a process chemist and I agree totally but it's bad practices which made ISO what it is.

I have seen some shoddy practices "that'll do" and so on.

However what ISO has done is removed any kind of independent thinking and improvisation when a problem occurs which does not fit the written protocol.

I have worked with labs which have never calibrated their balances or don't check their pipettes. Ok the results won't be massively imprecise but when your calculations are dependent on precision and accuracy for a standardised extract  it can get very frustrating when you can't be confident with the results.

I guess it's not hard to pull the wool over the eyes of inspectors.

[Archer]

I agree that 4800ppm methanol should be 7200mg/L COD.

But I disagree with Archer that I'd have more faith in an accredited lab not giving a wrong answer. 

My experience of every QC lab I've ever dealt with in any capacity is that they very rarely 2nd guess their own work.  Analysts trust the numbers they get even if a simple calculation can prove that number to be impossible.   I can not count the number of times with internal and external QC lab data that I have had to take the time to pick that data apart to show that at least one result must be wrong.

Fair point, I am just lucky to be working with a really good team at the moment. Thinking about it this is about the individuals rather than the accreditation.

[curiouscat]

Thinking about it this is about the individuals rather than the accreditation.

Precisely how I feel. I've bumped into the most  reliable & wise analytical chemists in the most unlikely of settings.