HOBBY GRADE TEST KITS CAN OUTPERFORM ICP MEASUREMENTS…REALLY??

[BeanAnimal]

Not sure why you put (nonsense). I agree with most of what you said, but I’m unsure what you meant about nonsense. Anyway, ICP-MS will find those ultra trace elements that OES will not. I’ve seen that over and over again. The sensitivity is good. 0.05 ug/L or better for most elements. Uranium is probably the most sensitive.

Looking at the chart - it "found" those trace elements, but the 4 (identicle? samples) hey are all over the chart. So the data really makes no sense when compared to the other brands that return Zero.

The presence of data does not mean the data is valid or useful.

 

[Reefahholic]

Reefahaloc and the moonshiners certainly trust the data return and their system results - I am just not sure if I do (back to that correlation/attribution thing).

We trust OCEAMO’s ICP-MS. I do not trust ATI-OES or any OES for that matter with anything under a few ug/L. For elements above 2-3 ug/L it is trustworthy, but even around 3 ug/L I still feel much better with MS.

I’ve posted this pic several times, but this illustrates my point. This was when I first transitioned to MS. Keep in my I was not doing water changes and we dosed dailies just to keep these elements present w/o them becoming detectable on ATI-OES because they would have been overdosed at that point. Note low values that ICP-MS picked up, and how ATI was unable to detect them the previous month. We’re taking specifically about the trace metals that have target ranges below 1 ug/L. You’ll see Triton is the same. Just another OES.

 

[BeanAnimal]

We trust OCEAMO’s ICP-MS. I do not trust ATI-OES or any OES for that matter with anything under a few ug/L. For elements above 2-3 ug/L it is trustworthy, but even around 3 ug/L I still feel much better with MS.

Verified against what? I think that is part of the hangup here.

 

[Reefahholic]

Looking at the chart - it "found" those trace elements, but the 4 (identicle? samples) hey are all over the chart. So the data really makes no sense when compared to the other brands that return Zero.

The presence of data does not mean the data is valid or useful.

Yes, I trust ICP-A about as much as I trust the CEO of Vibrant.

I agree with your last statement if it pertains to labs using OES, and ultra trace elements are being detected. Like you said…the presence of data doesn’t mean the data is valid. When the machine is at the LLOD, that’s it. Everything close to that or under is useless. It’s not reliable or trustworthy.

 

[areefer01]

Yep, all zero just like I mentioned. It’s either very low or depleted. My guess is that it’s in there, but probably between about .05 to 0.2 ug/L.

Yeah not arguing I was curious is all as I was reading through the updates this morning.

 

[Reefahholic]

Verified against what? I think that is part of the hangup here.

ICP-MS is much more sensitive than OES. That’s a fact. You can look it up on any certified lab’s website. Christoph knows the LLOD of his machines. He has tested and verified against known sample concentrations. He told me that typical detection limits for most elements are 0.05 ug/L or better. I’ve seen it many times when comparing any other labs OES to his MS. I’ve seen my dose’s come out as expected, and when I increase a dose, it acts as expected on the next ICP and trending.

 

[Reefahholic]

Yeah not arguing I was curious is all as I was reading through the updates this morning.

What would be a excellent example for these guys here, is if you send an OCEAMO ICP-MS in the next few days. Let’s compare the data from that MS to the data from your last OES. I will bet it picks up those Ultra Trace if you’re actively dosing AFR and your demand/bioload in the system isn’t crazy high. If that was the case, they could be truly depleted, but I doubt it. I’d really need to see a picture of your system to give a final guesstimate.

 

[areefer01]

What would be a excellent example for these guys here, is if you send an OCEAMO ICP-MS in the next few days. Let’s compare the data from that MS to the data from your last OES. I will bet it picks up those Ultra Trace if you’re actively dosing AFR and your demand/bioload in the system isn’t crazy high. If that was the case, they could be truly depleted, but I doubt it. I’d really need to see a picture of your system to give a final guesstimate.

I have sent in a couple to Oceamo I believe. Let me check. I did not send them at the same time nor was it to compare who is accurate. I was curious of the service provided by Oceamo and the personal feedback from them.

To be honest I don't know why I am still sending samples in other than I like the data and it is easy to visualize in charts. I wish the eDNA results are in a similar presentable format but that is another story. I guess what I'm saying is that I have a history of data from ATI so I forego a few lattes and send in a new sample about every 3 months For me it is about trends and an outside view in that I can spot check with my Trident and manual tests that are quick to use (hanna nitrate, phosphate, and alk).

Since you asked I will login to Oceamo later and see if I have anything close to sample dates.

 

[taricha]

How likely is it for the person that's preparing the reports for customers to make assumptions that certain elevated elements are unlikely and then decide to remove those from the report?

I would have expected that, yes.
But @Rick Mathew, @Dan_P and I have sent in some kinda weird samples to ICP testing companies, and there's never been any indication that they manually rejected the sample. It's always seems like they measured it and reported what the machine gave. Sometimes well measured sometimes poorly, but seemed to be measured each time.

I thought you guys sent a MS with ICP-A?

Yes. This was from a data set involved in an earlier study. Not the current one.
You were talking about different tech specs of OES machines, and my comment was I've never seen any data that made it look like the specs of the machine were the controlling factor in accurate performance, but rather outside factors were more important.

If you know of some prediction where we could see a difference in data based on the different tech specs of the OES machines "green" vs "blue" or FM vs ATI, is be interested in hunting for the data to back it up.

Looking at the chart - it "found" those trace elements, but the 4 (identicle? samples) hey are all over the chart. So the data really makes no sense when compared to the other brands that return Zero.

Exactly. None of those elements were spiked to detectable amounts, so those four samples are identical regarding those elements. Note that ATI only reported two elements and their values were fairly consistent. Whereas ICP-A reported many elements fluctuating from 0 to several ppb from one replicate to the next. Hence "nonsense" - shorthand for reporting numbers that are well smaller than the measurable uncertainty - so could be zero and ought to have been reported as such.

 

[jda]

I don't believe DSR uses ICP testing but maybe it has changed.

DSR existed for a few decades before ICP was a thing. Saying that they are ICP based is a revisionist history exaggeration of some sort. You can make your own supplements, so purity is on you. It also works outside of Europe since Geofencing of supplements to limit their effectiveness was not a thing when it started.... and, still is not a thing. The few that I know that use it still do it the way that they always did and also have no use for ICP although it could be an novelty for some.

 

[jda]

Since we talked about truth and facts and stuff, can we list some out?

Please correct me if I get any of these wrong.

Facts:

• ICP testing companies do titration test for alk, calcium and magnesium
• ICP testing companies do color changing test for po4 and no3
• ICP testing companies do math to calculate total phosphorous from the po4 value that they get above, or use a total phosphorous color changing test kit
• The three above do not matter if you send in OES or MS
• Plasma ONLY can detect elemental levels of things and not compounds
• Plasma ONLY can count elemental atoms and does not know any form that anything is in
• Nobody has yet to produce any actual scientific facts or research saying how elemental atom counts are useful for dosing amounts
• OES is better with small amounts of some elements
• MS is better with small amounts of others
• MS perhaps is better with a wider range of elements
• There is no collaboration between ICP companies on standards or best practices - Chrisoph has said as much so I am assuming that this is still fact

What did I mess up? What did I miss?

 

[areefer01]

DSR existed for a few decades before ICP was a thing. Saying that they are ICP based is a revisionist history exaggeration of some sort. You can make your own supplements, so purity is on you. It also works outside of Europe since Geofencing of supplements to limit their effectiveness was not a thing when it started.... and, still is not a thing. The few that I know that use it still do it the way that they always did and also have no use for ICP although it could be an novelty for some.

Glenn is on another level for sure and an amazing hobbyist. I was only replying that I didn't think he was ICP driven based on my understanding.

 

[BeanAnimal]

Exactly. None of those elements were spiked to detectable amounts, so those four samples are identical regarding those elements. Note that ATI only reported two elements and their values were fairly consistent. Whereas ICP-A reported many elements fluctuating from 0 to several ppb from one replicate to the next. Hence "nonsense" - shorthand for reporting numbers that are well smaller than the measurable uncertainty - so could be zero and ought to have been reported as such.

I am not going to entertain trying to explain what is happening here (it does not matter). But we see this a lot with any type of data (both in the digital and analog domains, but more in the digital). It can be the result of floating point errors, compounded rounding errors, lack of data being replaced by statistical noise, filter noise, etc. While those ppb results may be in the theoretical significant sensitivity of the machine, the results returned appear to be either in error or a casualty of the aforementioned significant or noise problems somewhere in the process.

While I don't know (or care to know) the details of how either EOS OR MS results are interpreted, I do know that there is a lot of "fitting" that has to be done with regard to how the results are interpreted and scaled, etc.

We say that we know what the different layers of the earth from crust to core are comprised of because we measure wave propagation through the earth from different points on the surface. We don't actually know, but we interpret what we do know about wave propagation, material density and composition, boundary layers, temperature, etc. and apply it to those wave patterns and paint a probable picture of what we can't see. I take ICP testing as closer to this than to actually counting a finite number of atoms or molecules of each thing being tested for.

 

[MnFish1]

Yes, this is true on average. Vendor-to-vendor variations tend to be larger than single-vendor-over-time variations.
And variations between vendors in the absolute values of elements are generally larger than the variations between vendors on their assessments of the trends of those elements.
What I mean is that you probably get better data if you ask ICP about trends of your elements than ask about the raw levels of those same elements.

An interesting question would be do the different ICP companies have different 'normal ranges - (i.e. if element xxx is between 1 and 3 - its ok on company A's test, but on Company B's test ok for element xxx may be between 2 and 4). This could account for different variations between vendors.

If one looks purely at medical blood tests, the 'normal range' is usually +- 2 standard deviations - which implies if an MD orders 100 tests on 1000 'proven normal (healthy) people, about 2.5 percent will have values over the normal range, and about 2.5 percent will be below the normal range - which have nothing to do with the health of the person.

 

[Randy Holmes-Farley]

An interesting question would be do the different ICP companies have different 'normal ranges - (i.e. if element xxx is between 1 and 3 - its ok on company A's test, but on Company B's test ok for element xxx may be between 2 and 4). This could account for different variations between vendors.

If one looks purely at medical blood tests, the 'normal range' is usually +- 2 standard deviations - which implies if an MD orders 100 tests on 1000 'proven normal (healthy) people, about 2.5 percent will have values over the normal range, and about 2.5 percent will be below the normal range - which have nothing to do with the health of the person.

IMO, it would be misleading to quote a normal or optimal or subpar range for many trace metals. Icp of any sort just does not provide sufficient information to go from an academic exercise (knowing the concentration of the element) to the thing that matter: bioavailability to organisms.

This discrepancy has real world issues. Back years ago when Ron Shimek was on a rampage about copper, I tried to explain to him that 10 ppb raw ionic copper added to artificial seawater was a very different story than 10 ppb present in a reef tank where it is complexed by organics. When tested on sea urchin larvae, he did find the former was quite toxic and the latter was not.

 

[MnFish1]

IMO, it would be misleading to quote a normal or optimal or subpar range for many trace metals. Icp of any sort just does not provide sufficient information to go from an academic exercise (knowing the concentration of the element) to the thing that matter: bioavailability to organisms.

This discrepancy has real world issues. Back years ago when Ron Shimek was on a rampage about copper, I tried to explain to him that 10 ppb raw ionic copper added to artificial seawater was a very different story than 10 ppb present in a reef tank where it is complexed by organics. When tested on sea urchin larvae, he did find the former was quite toxic and the latter was not.

I agree 100% - when ICP companies are coloring element x as 'green', 'yellow' or 'red' based on (I assume a certain range that they have developed - whether its internal - or whether its based on Sea water - I don't know) aren't they doing just what you suggest they should not be doing?

 

[jda]

...aren't they doing just what you suggest they should not be doing?

Of course they are... or even worse to sell a supplement in some cases.

 

[BeanAnimal]

Back years ago when Ron Shimek was on a rampage about copper,

Ron on a rampage? Nah.. don't recall him being that passionate

 

[areefer01]

Since it was asked earlier or talked about I have an ATI and Oceamo somewhat close as it relates to sample date. I've never compared them before but this is back in February of 2021.

Just threw them into excel real quick for more noise and fun.

 

[MnFish1]

It comes down to the ranges each labs use - I would not for example change from lab A one month - and lab B the second month