ICP tests...now what?

[Rick Krejci]

Recently decided that I'd like to understand where my tank's trace elements are sitting since I've been running for almost 3 years with minimal water changes (maybe 10% every few months). and have been using Chaeto and now Algae Turf Scrubbing as primary nutrient exports.

I sent a sample to Triton. When I got it back, things looked pretty good except my Strontium was 50x the reference value(397ppm vs 8)! The magnesium also came out over 100 ppm higher than my Trident reads (1516 ppm vs 1395). The Trident has been very consistent, even through a few changes of reagents and re-calibration, so I have confidence in it. Ca came out spot on between Trident and Triton.

Well, a few weeks pass and I decide to try the ATI ICP for comparison, since it tests the RO water as well...trying to track the Strontium thing down. Hadn't done any water changes, and the only other thing changed was switching from Sodium Carb to Sodium Bicarb, which helped stabilized things a bit and stop my gravel clumping. I also dosed some Iron supplement since my Turf scrubber was struggling. Well, Strontium came back lower than baseline (1.31ppm), 1/300 of Triton's test value!!! Magnesium matched the Triden't value very well. Calcium came out 55ppm lower than what Trident measured this time.

There were various other large and small discrepancies between ATI and Triton that I really couldn't explain. Aluminum was 3.1 on Triton, 1 on ATI. Silicon was 158 on Trident vs. 327 on ATI. Iron went from not detected for Triton, 2.24 for ATI which actually makes sense since I specifically dosed some. I've attached my consolidation of the results in a spreadsheet for anyone interested.
But generally, I'm disappointed in some of the major inconsistencies between the 2 ICP tests. My Strontium didn't go from ridiculously high 397 down to 1.3 in a few weeks, without water changes (and yes, I checked the units were the same). My Silicon didn't go from fairly normal to double with no additives other than 2 part and no water changes. My Magnesium didn't go down by 130ppm, with the Trident reading being fairly constant. How can I be confident to dose something if I don't believe it. I'm certainly not going to chase these minor elements.

It did give me some confidence to help me decide to dose Brightwell's Chaeto Gro (most of the elements is provides were on the low side), and I noticed a positive change to my turf scrubber color and production in a week, whereas before just iron supplements didn't help much.

I'll just listen to my tank and save the high money the ICP tests cost in the future.

 

[Lasse]

In a Triton test - there are 2 tubes. They only analyse one of them and keep the other as backup. When I get such unexpected reading of something like you did with the strontium concentration - I´m always ask for an analysis of the second tube.

Sincerely Lasse

 

[Rick Krejci]

In a Triton test - there are 2 tubes. They only analyse one of them and keep the other as backup. When I get such unexpected reading of something like you did with the strontium concentration - I´m always ask for an analysis of the second tube.

Sincerely Lasse

Thanks...I didn't know that was an option. Hope it's not too late.

 

[Dana Riddle]

Fauna Marin performs ICP analyses for $19.99 when buying a 2-pack. Less if you buy more than 2. Just got mine back and there weren't any surprises.
The samples are shipped to Germany (from Florida) and I got the report in 9 days.

 

[Rick Krejci]

May wait a few moths of dosing the Chaeto Gro and send a test in to get an idea if I'm over or under doing it.

 

[Dan_P]

Recently decided that I'd like to understand where my tank's trace elements are sitting since I've been running for almost 3 years with minimal water changes (maybe 10% every few months). and have been using Chaeto and now Algae Turf Scrubbing as primary nutrient exports.

I sent a sample to Triton. When I got it back, things looked pretty good except my Strontium was 50x the reference value(397ppm vs 8)! The magnesium also came out over 100 ppm higher than my Trident reads (1516 ppm vs 1395). The Trident has been very consistent, even through a few changes of reagents and re-calibration, so I have confidence in it. Ca came out spot on between Trident and Triton.

Well, a few weeks pass and I decide to try the ATI ICP for comparison, since it tests the RO water as well...trying to track the Strontium thing down. Hadn't done any water changes, and the only other thing changed was switching from Sodium Carb to Sodium Bicarb, which helped stabilized things a bit and stop my gravel clumping. I also dosed some Iron supplement since my Turf scrubber was struggling. Well, Strontium came back lower than baseline (1.31ppm), 1/300 of Triton's test value!!! Magnesium matched the Triden't value very well. Calcium came out 55ppm lower than what Trident measured this time.

There were various other large and small discrepancies between ATI and Triton that I really couldn't explain. Aluminum was 3.1 on Triton, 1 on ATI. Silicon was 158 on Trident vs. 327 on ATI. Iron went from not detected for Triton, 2.24 for ATI which actually makes sense since I specifically dosed some. I've attached my consolidation of the results in a spreadsheet for anyone interested.
But generally, I'm disappointed in some of the major inconsistencies between the 2 ICP tests. My Strontium didn't go from ridiculously high 397 down to 1.3 in a few weeks, without water changes (and yes, I checked the units were the same). My Silicon didn't go from fairly normal to double with no additives other than 2 part and no water changes. My Magnesium didn't go down by 130ppm, with the Trident reading being fairly constant. How can I be confident to dose something if I don't believe it. I'm certainly not going to chase these minor elements.

It did give me some confidence to help me decide to dose Brightwell's Chaeto Gro (most of the elements is provides were on the low side), and I noticed a positive change to my turf scrubber color and production in a week, whereas before just iron supplements didn't help much.

I'll just listen to my tank and save the high money the ICP tests cost in the future.

Unless the ICP companies are providing 95% confidence intervals for their analyses, I am not exactly sure why you would trust the results.

 

[Rick Krejci]

Unless the ICP companies are providing 95% confidence intervals for their analyses, I am not exactly sure why you would trust the results.

I guess that was my expensive lesson for the month...

 

[Rick Krejci]

Update...based on Lasse's suggestion, I wrote to Triton to get a re-test. Within 24 hours, they had an updated analysis. I'm assuming it wasn't an actual re-test of the water sample, since every single parameter except Strontium was 100% identical and even that was an exact factor of 1000 off...I'd think there would be a some slight variance in at least one measurement. So I'm assuming a data entry error from the original results. The units were the same, and the graph went from being pegged right too high at 397 mg/l to being on the low/left side at 0.397 mg/l and recommendations went from water changes to reduce Strontium to dosing to increase it.

I'll chalk this one up to human error...it happens. I'll just know if I ever do this again to re-confirm stray measurements like this.

 

[Dan_P]

Update...based on Lasse's suggestion, I wrote to Triton to get a re-test. Within 24 hours, they had an updated analysis. I'm assuming it wasn't an actual re-test of the water sample, since every single parameter except Strontium was 100% identical and even that was an exact factor of 1000 off...I'd think there would be a some slight variance in at least one measurement. So I'm assuming a data entry error from the original results. The units were the same, and the graph went from being pegged right too high at 397 mg/l to being on the low/left side at 0.397 mg/l and recommendations went from water changes to reduce Strontium to dosing to increase it.

I'll chalk this one up to human error...it happens. I'll just know if I ever do this again to re-confirm stray measurements like this.

You are being too lenient! If you had not flagged their strange result, they would have been just as happy with what they gave you. Human error, or poor quality?

What they most likely did was to relook at the recorded spectrum for you sample and reinterpret the strontium signal. That is why all the other results did not change. The software that read the peak height made a “mistake” and the human adjusted the value, but only after you mentioned it. I wonder how often they deliver spurious data?

Your experience highlights why anyone working with these ICP companies must use caution when taking action based on the data they supply. The results might just be off or they could be erroneous.

 

[Rick Krejci]

I agree, Dan. My expectations of the ICP tests (Triton and ATI) were not met. For the amount they charge, I would expect better, but the both had enough discrepancies that I don't know really that much more for sure than I did $90 ago.

I hadn't heard many people having issues, but unless you submit 2 tests in relatively close time proximity, you don't know what you don't know.

 

[Dan_P]

I agree, Dan. My expectations of the ICP tests (Triton and ATI) were not met. For the amount they charge, I would expect better, but the both had enough discrepancies that I don't know really that much more for sure than I did $90 ago.

I hadn't heard many people having issues, but unless you submit 2 tests in relatively close time proximity, you don't know what you don't know.

The ICP companies are getting away with poor quality analytical services for precisely that reason: no one is talking about it. The reason quality is not a hot topic Is 1) very few people understand the ICP technique, 2) have very little background in analytical chemistry, 3) we humans tend to want to believe in a good story, and 4) it takes money to understand how well the vendor is doing. So yeah, how would we as a group even have an inkling of anything being wrong? A tip off might have been that the vendors were not bragging about the small size of their measurement standard deviations.

 

[robbyg]

You are being too lenient! If you had not flagged their strange result, they would have been just as happy with what they gave you. Human error, or poor quality?

What they most likely did was to relook at the recorded spectrum for you sample and reinterpret the strontium signal. That is why all the other results did not change. The software that read the peak height made a “mistake” and the human adjusted the value, but only after you mentioned it. I wonder how often they deliver spurious data?

Your experience highlights why anyone working with these ICP companies must use caution when taking action based on the data they supply. The results might just be off or they could be erroneous.

I agree with you also. Who is certifying the ICP companies to make sure that are producing accurate results?
I would love it if someone like @Randy Holmes-Farley made up a specific batch of SW from scratch and by scratch I mean starting with distilled water and then adding in a specific amount of Sodium, Ca, Mg and all the trace elements as best he can until he has a gallon of water with as many ICP tested elements in it. Then get a few guys to send them out to various companies and lets see which one of them is providing us with the best results. I would donate some money to such a project as it would save me a lot of money and grief later on.

 

[Randy Holmes-Farley Verified Community Expert]

Major and minor ion standard would be easy to evaluate by buying a true seawater standard (like from Osil). Trace elements would be a major undertaking to verify as a standard.

 

[Rick Krejci]

For ATI at least, they supposedly run a control sample every 3rd test to ensure calibration, but no published confidence intervals.

Follow up to my inquiries for Triton and ATI if they could re-run a sample:

As I said before, Triton didn't re-run it, but just fixed a clerical error. I told them I was a bit disappointed and that I felt compelled to send in for another test due to their error, and they offered to send another test kit. I thought that was fair.

ATI indicated that they no longer had the sample at all (less than a week after the test). I mean they have you send in 3 water samples plus an RO sample, odd that they wouldn't retain at least one for a few weeks in case there were legitimate questions.

 

[Rick Krejci]

Final update...When I questioned ATI on their retention of samples, the response was

"We store backup samples for one week, but I could not find your samples in the box that belongs to the day your samples were uploaded. The data we´ve obtained from the analysis of your samples were plausible, therefore I do not have a doubt that the results are correct. "

So much for sending in 3 samples

They did offer to do a re-test if I buy my own vials and send them to their lab in Germany on my own nickel.

 

[h2so4hurts]

They did offer to do a re-test if I buy my own vials and send them to their lab in Germany on my own nickel.

Super generous of them. That'd be enough for me to never use them again.

I think a lot of people do talk about the voodoo of ICP testing. I'm not sure you can trust their results especially if they're not running the requisite standards for all of their analytes.