Reefs dot com article on triton testing

[hart24601]

I didn't see any links to this, sorry if I missed it. Curious what people think about the article. Randy?


Reefs Magazine - Skeptical Reefkeeping XII:

 

[Greenstreet.1]

Great read thank you for posting.

 

[Randy Holmes-Farley]

It is an interesting paper and such tests are worth doing, but I have a few comments.

1. For most of major and minor ions, it seems plenty accurate to me, and that is good to see since I had a lot of trouble getting a highly accurate number for calcium using the same basic method. Iodine seems off significantly, which is somewhat of a concern since many reefers target iodine. The acidification of the standard by the manufacturer might have an impact on iodine since iodate (the predominant form in seawater) has a pKa around 1. Acidification to near that level or lower (which is done in the author's standard but not in normal tank samples sent to Triton) might cause some loss of HIO3 to the air before the sample gets to the ICP plasma. If so, that issue won't reproduce in tank samples, but I have no idea if it is the reason or not (or even if the standard sample contained iodate). The article's analysis and comment on acidification does not address the potential for this element to go into the gas phase and leave the sample when the pH was dropped to near or below the pKa.

2. For the trace elements, being off by up to a factor of 2 (what they'd call 100%) seems not optimal, but is likely accurate enough for any reef husbandry purpose. Take copper as an example. The standard claims 10 ppb. The test gave 13 ppb. That's "33%" error, but the way we'd use that number is exactly the same, no matter what the value was with that same % error. Either too high, about right, or too low. 33% error won't change that for a trace element (IMO). A few are off by more than that (Cd, Mn, Zn, Ni). That deviation of more than a factor of two is a concern if those elements give a result that deviates from NSW in a way that is important to the user. Several of these are elements are those which aquarists have been concerned with, either because there's too much or too little, and having a better handle on these values would be highly desirable, IMO.

3. The "guaranteed standard" was evaluated by the manufacturer using the same basic method that Triton used (although the details might be very different). Without further testing, one might properly assume the guaranteed analysis is more likely to be correct, but that isn't a certainty in the case of the trace elements, IMO.

4. There has been some discussion about the sample supplied being acidified with nitric acid by the manufacturer. Such a "matrix" effect might impact the results for some elements (the effect is well known, but usuallyn smaller than the large discrepancy for some of the trace elements; the paper shows a small subset of data from the literature, but shows no data for the elements most off in the Triton study). Hence a dispute between Ehsan of Triton and the authors of the study. Interestingly, the acidification is supposed to prevent binding of ions to the sides of the containers and equipment used in the study. Since Triton doesn't acidify, there may possibly be some inherent loses of some ions in normal tank samples, but not in acidified ones. So if Triton's method inherently has some loses in normal samples (and their own standards) that are not present in the acidified samples sent to them by the authors, then that might possibly be why they got higher values for some elements: because Triton's method "assumed" there would be losses in their equipment that didn't happen in the standard. Or it may just be a faulty method.

IMO, the only way to answer the acidification issue with certainty is to measure a normal seawater sample, acidify it, and measure it again.

 

[hart24601]

Awesome! Thanks for your comments Randy, it's very interesting. Could one make a standard to test using very pure lab grade chemicals to make NSW levels from scratch so to speak? Wouldn't that remove the issue of having a standard that was examined using the same process. Would that be accurate if using an extremely sensitive balance and/or making a large amount so the trace elements could be added accurately - or would there be issues getting everything in solution correctly?

 

[Randy Holmes-Farley]

Awesome! Thanks for your comments Randy, it's very interesting. Could one make a standard to test using very pure lab grade chemicals to make NSW levels from scratch so to speak? Wouldn't that remove the issue of having a standard that was examined using the same process. Would that be accurate if using an extremely sensitive balance and/or making a large amount so the trace elements could be added accurately - or would there be issues getting everything in solution correctly?

The problem with such a standard in general is that even very low impurities in the NCl, CaCl2, MgCl2, and MgSO4 used to make a seawater standard can be as high or higher than the low levels of some of the trace elements in seawater.

But what you suggest is likely what the company did that made the standard used in the study.

 

[UK_Pete]

Does anyone know if we can see tritons response to the article that Randy mentioned?

 

[joefishUC]

Hello. Thank you for the response Randy. It does help clear things up a bit. Curious your thought on these few items. Some of the trace elements you mentioned like zinc, cadmium, nickel were not listed in the article as being "certified". They were in the certified sample as reference only (they are listed with an asterisk in the certificate). This would imply to me that those values are questionable from the start and therefore should not be scrutinized as closely as the certified elements. Is this a fair way to look at this?

Triton hasn't released an official statement regarding the article yet. Triton only claims accuracy of tests when performed on the matrix of seawater, not on an acid. The preparation, calibration, and method all rely on starting with actual seawater. This is absolutely a must according to Ehsan. Like Randy mentioned above, Iodine for example will test differently when in an acid vs seawater according to Ehsan. He accounts for this loss through calculations and customized methodology.

If any one else reading this or the article thinks that purchasing and sending a certified acidified sample would be a fun thing to do please consider that the ph in the sample is <2 and extremely dangerous to eyes, skin, etc. These samples SHOULD have been labeled for the safety of Triton-US and Trition GMBH who regularly handle the samples. They were sent disguised as seawater samples to Triton. On the msds sheet for CRM-SW it is listed as a hazardous material so the fun fact in the article is false.

 

[hart24601]

I hope you update this with Triton's response. Randy, did the article change how you view Triton's results at all either good or bad?

 

[Randy Holmes-Farley]

Not that much, no. I was happy to see how well they matched things like calcium, magnesium, and potassium.

I'm still not sure how to respond to trace element variations, even with 100% accuracy in measurement. But I suppose that if you see elevation in one of the ones mentioned above that read high in the standard, I might not do anything to try to reduce it until there is further evidence on the accuracy.

FWIW, of those that the standard in the article read more than 2x high, (Cd, Mn, Zn, Ni), my personal tank results showed none detected for all of those except zinc, which was normal. So even assuming Triton reads too high, it doesn't suggest I have too much of any of these.

My Triton Testing Results: By Randy Holmes-Farley - REEFEDITION

 

[craigbingman]

It is an interesting paper and such tests are worth doing, but I have a few comments.

Hi Randy.

1. For most of major and minor ions, it seems plenty accurate to me, and that is good to see since I had a lot of trouble getting a highly accurate number for calcium using the same basic method.

I would agree with that for "hobby purposes" it seems good enough (so are most titration tests too) but it isn't good enough to recover alkalinity by charge balance. One issue with this kind of testing is the lag time between the collection of the sample and getting the result. For calcium and magnesium, the numbers may be substantially different by the time you get the results back. I worry about chasing lagging numbers for highly dynamic values.

Iodine seems off significantly, which is somewhat of a concern since many reefers target iodine. The acidification of the standard by the manufacturer might have an impact on iodine since iodate (the predominant form in seawater) has a pKa around 1. Acidification to near that level or lower (which is done in the author's standard but not in normal tank samples sent to Triton) might cause some loss of HIO3 to the air before the sample gets to the ICP plasma. If so, that issue won't reproduce in tank samples, but I have no idea if it is the reason or not (or even if the standard sample contained iodate). The article's analysis and comment on acidification does not address the potential for this element to go into the gas phase and leave the sample when the pH was dropped to near or below the pKa.

Iodine was off somewhat. Iodine redox is complicated. I'm not as worried about iodic acid losses to the vapor phase as I am about losing elemental iodine into the walls of the tube. Colorimetric tests for iodine/ide/ate are complicated and having a reference method that I could really trust would be very nice.

2. For the trace elements, being off by up to a factor of 2 (what they'd call 100%) seems not optimal, but is likely accurate enough for any reef husbandry purpose. Take copper as an example. The standard claims 10 ppb. The test gave 13 ppb. That's "33%" error, but the way we'd use that number is exactly the same, no matter what the value was with that same % error. Either too high, about right, or too low. 33% error won't change that for a trace element (IMO). A few are off by more than that (Cd, Mn, Zn, Ni). That deviation of more than a factor of two is a concern if those elements give a result that deviates from NSW in a way that is important to the user. Several of these are elements are those which aquarists have been concerned with, either because there's too much or too little, and having a better handle on these values would be highly desirable, IMO.

I think some of the deviations seem large enough that one might begin to question "fit for purpose" in some use cases.

3. The "guaranteed standard" was evaluated by the manufacturer using the same basic method that Triton used (although the details might be very different). Without further testing, one might properly assume the guaranteed analysis is more likely to be correct, but that isn't a certainty in the case of the trace elements, IMO.

It was certified reference material, and I have a high degree of confidence that the reference material is as described for the certified components.

4. There has been some discussion about the sample supplied being acidified with nitric acid by the manufacturer. Such a "matrix" effect might impact the results for some elements (the effect is well known, but usuallyn smaller than the large discrepancy for some of the trace elements; the paper shows a small subset of data from the literature, but shows no data for the elements most off in the Triton study). Hence a dispute between Ehsan of Triton and the authors of the study. Interestingly, the acidification is supposed to prevent binding of ions to the sides of the containers and equipment used in the study. Since Triton doesn't acidify, there may possibly be some inherent loses of some ions in normal tank samples, but not in acidified ones. So if Triton's method inherently has some loses in normal samples (and their own standards) that are not present in the acidified samples sent to them by the authors, then that might possibly be why they got higher values for some elements: because Triton's method "assumed" there would be losses in their equipment that didn't happen in the standard. Or it may just be a faulty method.

About the acidification issue: the commentary on that issue (elsewhere) has veered pretty deeply into comedic territory, from my perspective. A low point for me was concern that their machine had somehow been compromised by the presentation of an acidified sample, when in fact, ICP-OES machines are built to withstand such samples BECAUSE IT IS EXPECTED THAT ALL TRACE METAL SAMPLES WILL BE RUN IN AN ACIDIFIED MATRIX. Trace metals in a seawater-like matrix will NOT be stable for days without acidification. This is simple chemistry, and is completely independent of whatever a certain testing service is saying. You might be able to redissolve a substantial amount of the lost metals by acidification of the samples upon receipt. It seems likely to me that Triton is doing this. That nearly invalidates the matrix argument, since we are looking at matrix differences between a 1x and 2x acidified sample, which will be a lot less than matrix differences between a 1x and 0x acidified sample.

IMO, the only way to answer the acidification issue with certainty is to measure a normal seawater sample, acidify it, and measure it again.

I suspect that the actual concentrations of trace metals in most sea surface water samples will be below the limit of quantification for this method, so that test might not be as revealing as you think, unless both samples were spiked to bring up the levels to the quantifiable range. Since we aren't supposed to send acidified samples to Triton, the test is impossible to perform. Unless they revise their opinion on that issue, I will never be able to have a high degree of confidence in this method. Period.

 

[craigbingman]

Triton hasn't released an official statement regarding the article yet. Triton only claims accuracy of tests when performed on the matrix of seawater, not on an acid. The preparation, calibration, and method all rely on starting with actual seawater. This is absolutely a must according to Ehsan. Like Randy mentioned above, Iodine for example will test differently when in an acid vs seawater according to Ehsan. He accounts for this loss through calculations and customized methodology.

Unless we can send stable matrices of known composition to Eshan, then there is no way to have confidence in the testing. You may not be aware of this, but it is completely normal procedure for people to send samples of known composition to a testing service. This is what everyone does. It would be much more impressive to me if we were working toward a system where it is possible to send such samples, rather than disallowing them.

I don't understand the basis for any calculation of loss of anything in these samples when it is not known how long has elapsed between the time that the sample was collected and delivered for testing, or given the large differences in the poise of water from different systems to precipitate carbonates, which will be a major loss mechanism for a number of transition elements. Losses will be a function of time, and saturation. Any correction is completely _ad hoc_ without knowing more about those two variables.

If any one else reading this or the article thinks that purchasing and sending a certified acidified sample would be a fun thing to do please consider that the ph in the sample is <2 and extremely dangerous to eyes, skin, etc. These samples SHOULD have been labeled for the safety of Triton-US and Trition GMBH who regularly handle the samples. They were sent disguised as seawater samples to Triton. On the msds sheet for CRM-SW it is listed as a hazardous material so the fun fact in the article is false.

Joe, this is a subject that we should be approaching with a little more caution than I have seen so far. I think that there are a large number of supplements that are shipped around without supplying MSDS for all of them. Perhaps we should, but one needs to be a bit careful here not to call attention to the mote in thy neighbor's eye while walking around with a 2x4 in one's own. My understanding is that normal seawater would require a MSDS, if we are really going there. The level of personal protective equipment required for handling seawater and seawater in a 2% nitric acid matrix is exactly the same. (Eye protection, lab coat, gloves.) Tank water that has been sitting around in a sealed tube for a week or so is likely to be bio-funky in ways that scare me more than a dilute acid sample. Bio-funky tank water is a hazardous material, too.

Let me tell you a story. A long long long time ago, when I was much much younger, I was doing some water testing for a major aquarium products vendor. They sent me samples from a number of systems, like ten or twelve total samples, and when I was testing them, I noticed that three of them were really suspiciously similar and looked a lot like seawater, while several of the samples were highly divergent. It turns out that they had sent me a triplicate sample of natural seawater. I was initially ******, since these were fairly laborious tests for me to conduct, and on some level that effort seemed to have been "wasted." However, I very quickly realized that this was something they needed to do in order to have confidence in my methods, and I really had nothing to hide. I hope that soon, you can feel the same way.

 

[Randy Holmes-Farley]

Thanks for the comments, Craig.

I suspect that the actual concentrations of trace metals in most sea surface water samples will be below the limit of quantification for this method, so that test might not be as revealing as you think, unless both samples were spiked to bring up the levels to the quantifiable range. Since we aren't supposed to send acidified samples to Triton, the test is impossible to perform. Unless they revise their opinion on that issue, I will never be able to have a high degree of confidence in this method. Period.

What I meant by a normal sample, was a tank sample. Sorry for not making that clear.

A tank sample that did show some trace elements, before and after acidification (or a split sample, one half of which is acidified on collection). That should help clarify the issue.

 

[craigbingman]

Thanks for the comments, Craig.

What I meant by a normal sample, was a tank sample. Sorry for not making that clear.

A tank sample that did show some trace elements, before and after acidification (or a split sample, one half of which is acidified on collection). That should help clarify the issue.

Hi Randy. If we can get Triton to agree to accept samples that have been put into a standard matrix, even if that requires us to make some separate declaration, that would be great progress. If we can convince them and the people in the supply chain to follow standard lab practice and use at least a bare minimum of personal protective equipment (eye protection, lab coat, gloves) that would also be good for everyone involved. Right now I would be worried to send a sample for fear that someone at Triton would wind up with a Vibrio infection and I would feel responsible, since I am forever plagued by an excessive sense of responsibility about everything that I touch. Sending the samples you mention would allow us to understand more about how the lack of acidification negatively effects the accuracy and reliability of the Triton testing, but it unfortunately won't tell us everything we need to know about how well "the mittens are on the kittens" (apropos the cute illustration in the article.) For that we need to be able to send reference materials. No other testing service in the world would freak out if I did that. They would of course ask that I pay full price for testing my standards, but they would do it with a smile, and presumably a wink when they nailed it. The upshot of Eshan's rules are that they make his service unverifiable, and given how strident they have been about making sure no one sends reference samples, that seems intentional rather than just ignorant. Make no mistake about it, Eshan's rules about samples place him beyond scrutiny. That would be a big leap of faith for me if we were dealing with someone who had extensive credentials in analytical chemistry. And such a person would never insist that I take that leap of faith! So, given all of this trust and faith that we are supposed to be placing in his unpublished methodology, and given that we can't verify his work, it seems appropriate at this point to ask Eshan to directly comment, here, on his educational background and credentials. What do you think, Randy? Does any of that seem unreasonable to you?

 

[Spracklcat]

Regarding the safety of shipping the standard, I've read the SDS. I review SDS's almost daily for various reasons, and this one strikes me as no more dangerous than most household products you all pick up at the grocery store (and use with no protective gear). If you are not used to reading these documents I could see how it would look scary dangerous, but if you compare it to a smattering of normal laboratory materials, it raises no flags. Good Laboratory Practices would have all sample handlers wearing protective gear anyway (coat/gloves/glasses) especially with unknowns.

 

[craigbingman]

Regarding the safety of shipping the standard, I've read the SDS. I review SDS's almost daily for various reasons, and this one strikes me as no more dangerous than most household products you all pick up at the grocery store (and use with no protective gear). If you are not used to reading these documents I could see how it would look scary dangerous, but if you compare it to a smattering of normal laboratory materials, it raises no flags.

Right. I found this in a Safety Data Sheet just today:

SECTION V: HEALTH HAZARD DATAThreshold Limit Value: Not established Primary Routes of Entry: Eyes; OralEffects of OverexposureConditionsto Avoid:May cause eye irritation with symptoms of pain, redness and swelling of the conjunctiva.May be harmful if swallowed with symptoms of abdominal pain, nausea and diarrhea.Avoid contact with eyes . Do not swallow.Discontinue use if signs of skin irritation or rash appear .Conditions Aggravated by Use: Use of this product may aggravate preexisting skin; eye and respiratory disorders including asthma anddermatitis.Emergency & First Aid Procedures :Eyes: Flush eyes with plenty of water for at least 15 minutes. Remove contact lenses. Get medical attention.Skin: Rinse skin with water. Get medical attention if irritation persists.Ingestion: Do not induce vomiting. Get medical attention. Do not give anything by mouth to an unconscious person.

SECTION VII: SPILL OR LEAK PROCEDURESSteps to be Taken in CaseMaterial is Released or Spilled: Small spills of one gallon or less may be flushed with plenty of water to sanitary sewer system (Ifpermitted by local sewer regulations). Dike and contain large spills with inert material and transfer liquidto containers for disposal. Keep spill out of storm sewers and waterways.

It sounds a little scary, right?

It is from a safety data sheet for waterless hand cleaner.

https://www.spartanchemical.com/products/product/296004

I'm not picking on them in any way. It is just that the safety data sheet for something that you are supposed to intentionally rub on your hands sounds a little scary.

Good Laboratory Practices would have all sample handlers wearing protective gear anyway (coat/gloves/glasses) especially with unknowns.

Absolutely.

In the case of aquarium water, the biological considerations probably are probably dominant (unless it is acidified or otherwise preserved) and standard Universal Precautions for handling biohazardous material dictate that it should be handled like infectious material.

e.g.
Diversity of Vibrio spp. populations in several exhibition aquaria ... - PubMed - NCBI "Diversity of Vibrio spp. populations in several exhibition aquaria with a shared water supply."
[Vasodilator activity of CO2 in newly formed subcutaneous gas pocke... - PubMed - NCBI23 "Vibrio damsela associated with diseased fish in Denmark."
http://www.ncbi.nlm.nih.gov/pubmed/19836641 "Tropical dermatology: marine and aquatic dermatology.
and more are all consistent with treatment of shipped aquarium water as material with non-trivial biohazard potential. Doubly so since people are likely to ship samples from aquariums that are in some degree of distress from unknown causes. The underlying issue might be chemical or it might be biological in nature. I think reasonable people would conclude that standard lab coat/gloves/eyewear would be used at all steps where there was the possibility of exposure.

 

[Randy Holmes-Farley]

FWIW, my biggest concern with trace elements continues to be the issues relating to bacteria and how they are counted or not counted. Regardless of the answer, I'm not sure how one should interpret the results for those biological elements (such as iron) which may be mostly in organisms, especially since the amount in the bacteria may change during the collection of the sample and the testing.

My understanding from what Ehsan said is that they were not part of the sample tested.

Acidification might kill bacteria and release trace elements. Not acidifying might allow continued growth and sequestration of trace elements. Either way, if the bacteria are not tested, it is a confounding issue. If they are part of the tested sample, the results may not be reflective of the "in water" concentrations.

I agree that evaluating standards is appropriate, as long as they fit the requirements of the expected analytical process. I just don't know what those are at the moment.

 

[Spracklcat]

Hey Craig, I did a big bacteriological survey of a bunch of aquaria a few years back. Know what I found, aside from Vibrio? Salmonella. Staph. Other pathogens. So yeah, acidified water would be a safer thing (assuming no PPE)

 

[craigbingman]

I agree that evaluating standards is appropriate, as long as they fit the requirements of the expected analytical process. I just don't know what those are at the moment.

I'm glad that we are in agreement about standards being important.

This situation is actually much more simple than you might have been led to believe. All analytical processes meant to examine trace metals in seawater by ICP start with an acidified matrix. (Hopefully filtered prior to acidification.) I'm open to read any citations indicating that other procedures are appropriate, but I don't really think there are any. If anyone finds such a reference, please send me a URL or DOI, as I'm always willing to learn something new. Thanks.

Eshan should feel welcomed to step in at any point and explain his procedures. I've seen "Eshan@triton" pop up while I have been looking at this thread, so I am pretty sure he is aware that it exists. <waves hello!>

 

[craigbingman]

Hey Craig, I did a big bacteriological survey of a bunch of aquaria a few years back. Know what I found, aside from Vibrio? Salmonella. Staph. Other pathogens. So yeah, acidified water would be a safer thing (assuming no PPE)

Yick.

 

[Electrobes]

Holy cow, two monster chemists that got me where I am today with reefing. :hail:

I've used Triton's testing services twice, and while I found some of the information interesting I don't feel compelled enough by it to worry about the other, outside of Calcium, Magnesium, Alkalinity, etc parameters (Actively trying to raise or lower it).

I will admit that article above and the discussion here has made me more suspicious about their unknown methods. I hadn't realized he was essentially blocking what I consider important knowledge that would further ensure confidence in his methods. But I guess it shows, having used the service twice, that curiosity got the better of me instead of actual research. :lol: