Comparing ICP Results & Accuracy

clifg

Well-Known Member
View Badges
Joined
Mar 28, 2013
Messages
981
Reaction score
1,099
Location
Belmont, MA
Rating - 0%
0   0   0
Thought I would throw this out there for discussion. Below is a table of a couple of ICP tests I've run as well as the N-DOC and a comparison to my own tests via Apex, Trident, Hanna, & Nyos. The first two columns were sampled within seconds of each other and I pulled the Apex & Trident data from the same time frame. The second two columns are from an earlier time when the N-DOC & Triton were taken within seconds of each other as well. Trident had been calibrated a day or two prior. My reactions:

-Alk - seems to match up pretty well between the three although triton only offers Carbonate Hardness through their N-DOC test ($$)
-Cal- All within 1 point of each other, pretty impressive for the Trident I would say.
-Mag- Trident is right in between the two tests
-Most of the trace and heavy metals are very similar between ATI & Triton, I don't have a way to manually test these
-ATI tests more parameters per kit but takes about a week longer to get the results
-ATI Tests Nitrate while you need the N-DOC to test for it from Triton, and you get a calculated number which doesn't seem correct to me. For the NYOS test I put in the closest color but I do feel like ATI has it right as the NYOS test was an inbetweener on the colors.
-One result from each of the ICPs was clearly inaccurate. For ATI it was silicone - reading at 2456 while the two Triton's have been 141 & 229. For Triton it was Phosphate at .06, when ATI read .17 and Hanna read .26. I've never clocked my phosphate below .16 on the Hanna and test almost every day, which is why I believe Triton was inaccurate this time. As well my first Triton came back .3, which was probably accurate at the time. I'm not pleased with either of the inaccuracies but I'd much rather silicon be inaccurate than phosphate, especially when one is clearly out of bounds and the other could be correct if I didn't have other means of data.
-I've attached the full PDFs to this thread and will update it with my next Triton (only one I currently have) as well as with my Aquabiomics results.
-Thanks for looking!

ICPResults.PNG



N-DOC URL https://www.triton-lab.de/en/showroom/n-doc/11055
 

Attachments

  • RSR XL200 - 04_29_2021 (B-lmBn5y) (1).pdf
    159.6 KB · Views: 61
  • RSR XL200 - 04_07_2021 (B-lnqnpQ) (1).pdf
    159.6 KB · Views: 61
  • Analyse117760.pdf
    129.8 KB · Views: 57

Dr. Dendrostein

Marine fish monthly
View Badges
Joined
Nov 8, 2017
Messages
9,581
Reaction score
20,789
Location
Fullerton, California
Rating - 0%
0   0   0
Thought I would throw this out there for discussion. Below is a table of a couple of ICP tests I've run as well as the N-DOC and a comparison to my own tests via Apex, Trident, Hanna, & Nyos. The first two columns were sampled within seconds of each other and I pulled the Apex & Trident data from the same time frame. The second two columns are from an earlier time when the N-DOC & Triton were taken within seconds of each other as well. Trident had been calibrated a day or two prior. My reactions:

-Alk - seems to match up pretty well between the three although triton only offers Carbonate Hardness through their N-DOC test ($$)
-Cal- All within 1 point of each other, pretty impressive for the Trident I would say.
-Mag- Trident is right in between the two tests
-Most of the trace and heavy metals are very similar between ATI & Triton, I don't have a way to manually test these
-ATI tests more parameters per kit but takes about a week longer to get the results
-ATI Tests Nitrate while you need the N-DOC to test for it from Triton, and you get a calculated number which doesn't seem correct to me. For the NYOS test I put in the closest color but I do feel like ATI has it right as the NYOS test was an inbetweener on the colors.
-One result from each of the ICPs was clearly inaccurate. For ATI it was silicone - reading at 2456 while the two Triton's have been 141 & 229. For Triton it was Phosphate at .06, when ATI read .17 and Hanna read .26. I've never clocked my phosphate below .16 on the Hanna and test almost every day, which is why I believe Triton was inaccurate this time. As well my first Triton came back .3, which was probably accurate at the time. I'm not pleased with either of the inaccuracies but I'd much rather silicon be inaccurate than phosphate, especially when one is clearly out of bounds and the other could be correct if I didn't have other means of data.
-I've attached the full PDFs to this thread and will update it with my next Triton (only one I currently have) as well as with my Aquabiomics results.
-Thanks for looking!

ICPResults.PNG



N-DOC URL https://www.triton-lab.de/en/showroom/n-doc/11055
Now the ICP test I do it for s$#@ and giggles. This is my opinion no one else's but most corals are similar to the canary in the mine. They tell you a lot, way before you test that water or go get your test kits and test. Right now I have an ICP 1999 kit and ready to send it off just for.....

Hopefully I didn't offend anyone, don't mean to. Doesn't hurt to know what's in the water, and don't forget don't drink the water.
 
OP
OP
C

clifg

Well-Known Member
View Badges
Joined
Mar 28, 2013
Messages
981
Reaction score
1,099
Location
Belmont, MA
Rating - 0%
0   0   0
Now the ICP test I do it for s$#@ and giggles. This is my opinion no one else's but most corals are similar to the canary in the mine. They tell you a lot, way before you test that water or go get your test kits and test. Right now I have an ICP 1999 kit and ready to send it off just for.....

Hopefully I didn't offend anyone, don't mean to. Doesn't hurt to know what's in the water, and don't forget don't drink the water.
I'm coming to that same conclusion. This all started when one of my Acan's looked ticked off. I tested phosphate and it was at .5 so I sent in that first ICP. I had it down to .3 via water changes and the Acan looked happy again. At .16 now and everything still looks happy. Might continue to do ICPs every quarter or so because what I think it's really good for is making sure nothing is leeching Aluminum or rusting away in the tank. Other than that I think the eye test is the number one test we can do.
 

Dan_P

5000 Club Member
View Badges
Joined
Sep 21, 2018
Messages
6,573
Reaction score
7,031
Rating - 0%
0   0   0
@Rick Mathew this seems like a good place to share your findings of multiple vendor results.

My opinion hasn’t change. Vendors who sell you information without telling you about the accuracy and precision of their tests are selling almost useless information.
 
OP
OP
C

clifg

Well-Known Member
View Badges
Joined
Mar 28, 2013
Messages
981
Reaction score
1,099
Location
Belmont, MA
Rating - 0%
0   0   0
@Rick Mathew this seems like a good place to share your findings of multiple vendor results.

My opinion hasn’t change. Vendors who sell you information without telling you about the accuracy and precision of their tests are selling almost useless information.
Honestly hadn't even looked into that. They've become so popular that I assumed those would've been published somewhere. Are you referring to one company or another, or are all ICP tests like that?
 

Dan_P

5000 Club Member
View Badges
Joined
Sep 21, 2018
Messages
6,573
Reaction score
7,031
Rating - 0%
0   0   0
Honestly hadn't even looked into that. They've become so popular that I assumed those would've been published somewhere. Are you referring to one company or another, or are all ICP tests like that?
Triton published something years ago about the accuracy of their method, but that is not the point. You need to know how well they are doing on everyone of your samples but they cannot tell you.

No vendor supplies you with the confidence interval for the results that they sell you. Another bit of information that is not shared with you by any vendor is that water samples can change during shipment. When your results vary from test to test you have no idea whether it is because of variation in their method, changes during shipping or changes in your water.
 
OP
OP
C

clifg

Well-Known Member
View Badges
Joined
Mar 28, 2013
Messages
981
Reaction score
1,099
Location
Belmont, MA
Rating - 0%
0   0   0
Triton published something years ago about the accuracy of their method, but that is not the point. You need to know how well they are doing on everyone of your samples but they cannot tell you.

No vendor supplies you with the confidence interval for the results that they sell you. Another bit of information that is not shared with you by any vendor is that water samples can change during shipment. When your results vary from test to test you have no idea whether it is because of variation in their method, changes during shipping or changes in your water.

Right. I was doing the microbe sample from Aquabiomics for ***** and giggles and they have you put the sample through some sort of stabilizer, which had me thinking about the ICPs I had done. Surely temperature and length of sitting around can have some effect on the water sample.

Do you do anything to test for contaminants such as lead, tin, or silicon?
 

Dan_P

5000 Club Member
View Badges
Joined
Sep 21, 2018
Messages
6,573
Reaction score
7,031
Rating - 0%
0   0   0
Right. I was doing the microbe sample from Aquabiomics for ***** and giggles and they have you put the sample through some sort of stabilizer, which had me thinking about the ICPs I had done. Surely temperature and length of sitting around can have some effect on the water sample.

Do you do anything to test for contaminants such as lead, tin, or silicon?
I do not test for contaminants, but definitely would if ICP vendors used validated methods. Since their results cannot be trusted, I avoid giving myself things to worry about that might not be real.
 

Rick Mathew

Valuable Member
View Badges
Joined
Mar 21, 2016
Messages
1,472
Reaction score
4,735
Location
North Central Florida
Rating - 0%
0   0   0
As Dan_P has very well communicated the results from an ICP test that does not include a confidence interval are of little value...One has no way of know what the measurement error...all you have is a single data point for each element evaluated...I have done several comparison as you have done with similar outcomes...and I always asks myself..."Which one is correct"....I have attached two of the ones I did...I also noted in you results the same issue I had in the phosphate measurement results...Your measurement was significantly higher using the Hanna method vs the ICP results...you might be interested in having a look at this article...It explains what is happening...At least my take on it

https://www.reef2reef.com/ams/sample-storage-and-its-impact-on-measurement-results-part-3.800/

ICP RESULTS CHART CORRECTION-page-001.jpg
ICP TEST COMPARISON WITH CHLORINATED SAMPLES WO VENDOR NAMES-page-001.jpg
 
OP
OP
C

clifg

Well-Known Member
View Badges
Joined
Mar 28, 2013
Messages
981
Reaction score
1,099
Location
Belmont, MA
Rating - 0%
0   0   0
As Dan_P has very well communicated the results from an ICP test that does not include a confidence interval are of little value...One has no way of know what the measurement error...all you have is a single data point for each element evaluated...I have done several comparison as you have done with similar outcomes...and I always asks myself..."Which one is correct"....I have attached two of the ones I did...I also noted in you results the same issue I had in the phosphate measurement results...Your measurement was significantly higher using the Hanna method vs the ICP results...you might be interested in having a look at this article...It explains what is happening...At least my take on it

https://www.reef2reef.com/ams/sample-storage-and-its-impact-on-measurement-results-part-3.800/

ICP RESULTS CHART CORRECTION-page-001.jpg
ICP TEST COMPARISON WITH CHLORINATED SAMPLES WO VENDOR NAMES-page-001.jpg
Thank you! Very interesting read on the chart and the article. I'm inferring a couple of things from your posts but please correct me if I'm wrong:

-If an ICP lab were to include some sort of pre-measured chlorine tablet & confidence interval it could be more accurate
-BUT there is still some degradation especially with Iron & other elements that can be used up by our biologically active water or other chemical reactions.
-Also a "but," some elements are negatively impacted by being stabilized
-Relying on ICP for notification solely of contaminants is because of degradation, i.e. Iron might be in your tank in a vast quantity but get "used up" as well as degrade far before it is processed.

What do you make of vendor 1 having the same PO4 for both the standard & chlorinated samples?
 

Rick Mathew

Valuable Member
View Badges
Joined
Mar 21, 2016
Messages
1,472
Reaction score
4,735
Location
North Central Florida
Rating - 0%
0   0   0
Thank you! Very interesting read on the chart and the article. I'm inferring a couple of things from your posts but please correct me if I'm wrong:

-If an ICP lab were to include some sort of pre-measured chlorine tablet & confidence interval it could be more accurate
-BUT there is still some degradation especially with Iron & other elements that can be used up by our biologically active water or other chemical reactions.
-Also a "but," some elements are negatively impacted by being stabilized
-Relying on ICP for notification solely of contaminants is because of degradation, i.e. Iron might be in your tank in a vast quantity but get "used up" as well as degrade far before it is processed.

What do you make of vendor 1 having the same PO4 for both the standard & chlorinated samples?

The main takeaway for me is that the sample we take for our tanks is biologically active. Any element or compound that could be consumed or generated by this biological activity can and most likely will happen. The incorporation of something (Cl) to hinder or stop this activity will stabilize the sample....It also has the potential of interfering with other as you point out....Discovering which ones would be a big challenge given that the experiment would require multiple samples to be evaluated by the vendor(s) and without a clear understanding of the error of each test one would not be able to adequately evaluate the outcome...Catch 22!!:oops:

The main purpose of my experiments was to help me understand what might be going on with the PO4 measurement...Why was my (and others...Including yours) Hanna measurement consistently higher the ICP (specifically Triton, but also others) results...It was not my intent to create a new sample handling procedure...That is way outside of my expertise... I wanted to understand and share my findings with others for review and comment.

Yes...Vendor # 1...I am not at all sure what to make of the results...It might be that their procedure is more robust than other....I will say that in other evaluations regarding PO4 measurement this vendor reports the same value exact in different samples that other ICP vendors as well as myself measure as being different...Example in the first chart ICP-3 is Vendor # 1 in the second chart...Again without a clear reference point not sure how to interpret this

Just one other note. Water sample gathering and handling is a science unto itself. The manor in which we handle our samples for ICP analysis is a long way from "standard" practice...If you are interested here is a link ... https://apps.who.int/iris/handle/10665/41851 ....

Rick
 
OP
OP
C

clifg

Well-Known Member
View Badges
Joined
Mar 28, 2013
Messages
981
Reaction score
1,099
Location
Belmont, MA
Rating - 0%
0   0   0
The main takeaway for me is that the sample we take for our tanks is biologically active. Any element or compound that could be consumed or generated by this biological activity can and most likely will happen. The incorporation of something (Cl) to hinder or stop this activity will stabilize the sample....It also has the potential of interfering with other as you point out....Discovering which ones would be a big challenge given that the experiment would require multiple samples to be evaluated by the vendor(s) and without a clear understanding of the error of each test one would not be able to adequately evaluate the outcome...Catch 22!!:oops:

The main purpose of my experiments was to help me understand what might be going on with the PO4 measurement...Why was my (and others...Including yours) Hanna measurement consistently higher the ICP (specifically Triton, but also others) results...It was not my intent to create a new sample handling procedure...That is way outside of my expertise... I wanted to understand and share my findings with others for review and comment.

Yes...Vendor # 1...I am not at all sure what to make of the results...It might be that their procedure is more robust than other....I will say that in other evaluations regarding PO4 measurement this vendor reports the same value exact in different samples that other ICP vendors as well as myself measure as being different...Example in the first chart ICP-3 is Vendor # 1 in the second chart...Again without a clear reference point not sure how to interpret this

Just one other note. Water sample gathering and handling is a science unto itself. The manor in which we handle our samples for ICP analysis is a long way from "standard" practice...If you are interested here is a link ... https://apps.who.int/iris/handle/10665/41851 ....

Rick
Thanks Rick! Very interesting stuff. I've been reading some of your other articles on accuracy in testing and decided to invest in a pipette etc. that you recommended rather than relying on drops (which had always bugged me anyway). Keep up the great work.
 

Rick Mathew

Valuable Member
View Badges
Joined
Mar 21, 2016
Messages
1,472
Reaction score
4,735
Location
North Central Florida
Rating - 0%
0   0   0
Thanks Rick! Very interesting stuff. I've been reading some of your other articles on accuracy in testing and decided to invest in a pipette etc. that you recommended rather than relying on drops (which had always bugged me anyway). Keep up the great work.
You are most welcome...glad I could be some help.

Rick
 

taricha

5000 Club Member
View Badges
Joined
May 22, 2016
Messages
6,484
Reaction score
9,996
Rating - 0%
0   0   0
-One result from each of the ICPs was clearly inaccurate. For ATI it was silicone - reading at 2456 while the two Triton's have been 141 & 229.
That Si result is weird. Could you confirm which units ATI and Triton are reporting in (ppb for both)?
 

taricha

5000 Club Member
View Badges
Joined
May 22, 2016
Messages
6,484
Reaction score
9,996
Rating - 0%
0   0   0
Both are ug/l
Thanks.
So 2.5ppm Si on ATI and 0.22ppm on Triton. (and 0.15ppm later)
That Si (if real) is enough to interfere with hanna ULR PO4 reading, making it show higher than actual. Details in this thread.
https://www.reef2reef.com/threads/high-silica-interference-with-po4-measurement.777022/


But I don't think it's real. I don't see how Triton could miss 2+ppm Si. I can come up with ways to show fluky high values of Si (vaporize a clump of diatoms, or a silica sand grain in the ICP) but I can't come up with a way to miss 2.5ppm if it were actually in the water. It's also much too large a discrepancy to be explained by bioactivity of the samples (though @Rick Mathew work above found that a tenth or so Si does get lost in bioactive samples sent to ICP).
 
OP
OP
C

clifg

Well-Known Member
View Badges
Joined
Mar 28, 2013
Messages
981
Reaction score
1,099
Location
Belmont, MA
Rating - 0%
0   0   0
Thanks.
So 2.5ppm Si on ATI and 0.22ppm on Triton. (and 0.15ppm later)
That Si (if real) is enough to interfere with hanna ULR PO4 reading, making it show higher than actual. Details in this thread.
https://www.reef2reef.com/threads/high-silica-interference-with-po4-measurement.777022/


But I don't think it's real. I don't see how Triton could miss 2+ppm Si. I can come up with ways to show fluky high values of Si (vaporize a clump of diatoms, or a silica sand grain in the ICP) but I can't come up with a way to miss 2.5ppm if it were actually in the water. It's also much too large a discrepancy to be explained by bioactivity of the samples (though @Rick Mathew work above found that a tenth or so Si does get lost in bioactive samples sent to ICP).
Very interesting. I tend to agree with you that it's unlikely Triton missed such a large discrepancy. However, at the time I tested I did have a bit of a diatom bloom. I wonder if my silicates were somewhere in-between the two tests or if I somehow got some of the diatoms in the ATI tube. Maybe as they died in shipping they released a larger percentage of silicate into the tube than would be found in the DT?
 

taricha

5000 Club Member
View Badges
Joined
May 22, 2016
Messages
6,484
Reaction score
9,996
Rating - 0%
0   0   0
dissolution of diatom Silica is slow, diatom frustules stick around in the sand for a pretty long time under seawater conditions. I don't think quick re-release of 2+ ppm into the water is likely.

The 0.23 and 0.14 ppm Si reported by triton is perfectly consistent with a noticeable diatom population in the tank. No need to assume it must've been higher.
 

Randy Holmes-Farley

Reef Chemist
View Badges
Joined
Sep 5, 2014
Messages
66,580
Reaction score
62,879
Location
Arlington, Massachusetts, United States
Rating - 0%
0   0   0
dissolution of diatom Silica is slow, diatom frustules stick around in the sand for a pretty long time under seawater conditions. I don't think quick re-release of 2+ ppm into the water is likely.

The 0.23 and 0.14 ppm Si reported by triton is perfectly consistent with a noticeable diatom population in the tank. No need to assume it must've been higher.

If diatoms, or diatom bits, are actually suspended in the water, different companies with different (or any) procedures for removing particulates may yield different results.
 
Back
Top