Reefs dot com article on triton testing

[joefishUC]

Thank you Electrobes for your great points.

I don't think I was clear enough. It wasn't so much that Ehsan's first few posts contained points that were lost in translation- actually you did a great job Ehsan! I was simply introducing myself as someone who will be helping to guide the discussion and highlight some of the key points that may not have received enough emphasis. I also want to point out that I represent Triton-US.

 

[Electrobes]

Thank you Electrobes for your great points.

I don't think I was clear enough. It wasn't so much that Ehsan's first few posts contained points that were lost in translation- actually you did a great job Ehsan! I was simply introducing myself as someone who will be helping to guide the discussion and highlight some of the key points that may not have received enough emphasis. I also want to point out that I represent Triton-US.

It didn't even occur to me that people may not recognize your username (I assume UC is for Unique Corals?). :xd:

 

[dankreef]

Randy,

Thanks for putting your results out and going through them step by step was great to get a better understanding of all them at once.

Secondly i've seen a few posts from Triton that said the tests were not 100% perfect, and are to help us in the Reefing hobby more. It seems like you are a little down on this companies ability to offer this leg up in testing. If not Triton then what would you recommend using that is $50 and 10 minutes of work? :nerd:

 

[UK_Pete]

What Triton are doing is pretty amazing and they are the first company to make it affordable, maybe none will follow for a long time, who knows.

But Triton guys, you have to be honest about the limitations of this technique. Having a physics background I understand how complex this technique is (although not the chemistry aspect of any preprocessing), separating the spectral lines of such a large array of elements with often very close interfering and strong emission lines like the sodium doublet at concentrations of 35 ppt, and getting data of elements at sub ppb levels, more than 7 orders of magnitude apart, must be very difficult and sensitive to all sorts of interferences. If you find that you can frequently get say an accuracy of 100% for a particular element but under some circumstances your accuracy drops to 500% for that element then you need to explain this, and ideally say in what circumstances accuracy will drop (ie which elements with interfering spectral lines to the ones you are using to quantify the element in question).

We (users) need to see regular independent testing by third parties we trust, like the testing reported here, for us to have any confidence at all in such a tricky technique, and if you are not up front about the pitfalls with the technique you are using then you will not get the confidence that I am think you deserve for this using technique from users.

Additionally the way that you report 'below detection limit' as zero might be interpreted differently to how you intend too. For example with iron displaying ppb to 2 decimal places when your LOD is 0.3 ppb for instance, this is an error to display it like this. You display < 0.3 as 0.00, meaning that a true iron level of 0.24 ppb is displayed as being below 0.01 ppb. Thats a significant error in displaying the results. The same goes with all the elements in your reports I think. When I look at the results people publish, I see 0.00, 0.00 etc, when in fact I should be seeing <0.3, <0.5 etc. Its important because everyone interprets these results as meaning absolute zero, and so when they hear otherwise they are surprised and dismayed. I do feel you should make sure its clear in the results of the limitations of the technique for each element (IE next to each element result).

If you don't explain to the users that the limitations of testing using this technique are not yours but limitations of the physics of the test itself (that is, the test that you can do for that price, not the best accuracy that can be attained if you spend 1000 times more on a test and take a month to do it), then it looks like any inaccuracy is yours. This is simply what can be done for the price. If users want to pay far more I am sure you can offer a service for that and achieve greater accuracy.

I think triton testing is great and I look forward to using the service myself but I really, really want to know what the technical limitations are. So if you have high iodine then your accuracy of tin and zinc will drop, that kind of thing. Ballpark figures for how inaccurate results will be under certain circumstances etc. And until I see a fair few independent results confirming your reported accuracies then I will be quite cautious of believing the reports completely.

 

[dankreef]

I am no chemist but I do know that if certain ions are measured at .000 descimal in NSW then it would be smart to have 3 decimals on the ones that need it. There is no point to having double decimal standards that come in triple decimal results. Obviously your going to see 0 everytime, which is a false reading.

 

[DAPG8GT]

Along for the knowledge and like others have stated I won't have anything to contribute other than an open mind to the truth..

Thanks to all the scientific minds putting their knowledge into the thread and especially Eshran for responding openly to real legitimate questions from people that know what questions to ask =)..

 

[Randy Holmes-Farley]

Randy,

Thanks for putting your results out and going through them step by step was great to get a better understanding of all them at once.

Secondly i've seen a few posts from Triton that said the tests were not 100% perfect, and are to help us in the Reefing hobby more. It seems like you are a little down on this companies ability to offer this leg up in testing. If not Triton then what would you recommend using that is $50 and 10 minutes of work? :nerd:

I don't think I've said anything about the accuracy one way or the other.

 

[Randy Holmes-Farley]

Hi Ehsan,

My current job is a consultant to other companies. So I do not have access to any ICP machines.

 

[Sangheili]

Additionally the way that you report 'below detection limit' as zero might be interpreted differently to how you intend too. For example with iron displaying ppb to 2 decimal places when your LOD is 0.3 ppb for instance, this is an error to display it like this. You display < 0.3 as 0.00, meaning that a true iron level of 0.24 ppb is displayed as being below 0.01 ppb. Thats a significant error in displaying the results. The same goes with all the elements in your reports I think. When I look at the results people publish, I see 0.00, 0.00 etc, when in fact I should be seeing <0.3, <0.5 etc. Its important because everyone interprets these results as meaning absolute zero, and so when they hear otherwise they are surprised and dismayed. I do feel you should make sure its clear in the results of the limitations of the technique for each element (IE next to each element result).

I think this is a very valid point that should be quickly implemented. I assume the reason it is currently displayed at 0.00 is due to software/programming complications.



Ehsan, I believe the discussion for me is just 2 questions:

1) Why does Triton not instruct users to acidify their samples? What are the advantages/disadvantages of this?

2) Is the "certified sample", assumed to be tested by an outside laboratory, more or less accurate then the Triton testing? I assume it was "certified" by testing it using a similar process that Triton does.



I would like to also point out that the article may be somewhat misleading. The way I interpret it is that only some of the elements in the pure sample are certified and others are provided without certification. A quick comparison with the Triton results grid shows that most (not all) of the very divergent test results are from these non-certified elements.

 

[GainesvilleReef]

I don't want to have to acidify my samples. I understand it may introduce error. But I don't want to have to acquire very pure and strong acid. Nor do I want to keep such acid around for something I may only do couple times of year. This would make a simple to use inexpensive test much more complicated and costly. I would not go through the trouble because I am a hobbyist.

 

[Ehsan@triton]

Hallo Pete

What Triton are doing is pretty amazing and they are the first company to make it affordable, maybe none will follow for a long time, who knows.

But Triton guys, you have to be honest about the limitations of this technique. Having a physics background I understand how complex this technique is (although not the chemistry aspect of any preprocessing), separating the spectral lines of such a large array of elements with often very close interfering and strong emission lines like the sodium doublet at concentrations of 35 ppt, and getting data of elements at sub ppb levels, more than 7 orders of magnitude apart, must be very difficult and sensitive to all sorts of interferences. If you find that you can frequently get say an accuracy of 100% for a particular element but under some circumstances your accuracy drops to 500% for that element then you need to explain this, and ideally say in what circumstances accuracy will drop (ie which elements with interfering spectral lines to the ones you are using to quantify the element in question).

Well nobody says that what we do is easy, but that don´t mean it will not work. If everybody could do it and it would be easy then a lot of people would have done it already i guess.....
TBH we prooved the abilitys of the tests and the products over 7 years now and in my and a lot of users opinion ,this discussion is lets say " to late " .
Why do you think the Interference is as high ? Out of knowledge because you work with the same mashine, or do you just think that it need to be high interference ?
With an ICP-OES of the kind we use the Interfrence aspect is not as problematic as it looks like in this article ( wich is very poor part of it ).
Seperating lines is depending on the resolution you scan. Our resolution is quite high so. we have around 22000 line we could choose to test on, you just need to choose one that has no big interference, wich you do out of experimenting. what exactly is the Problem ?
This is just The part of the article that should fear people away without any scientific argument.

The 500% out value is what is one of the worse things in that "supposed scientific" Article.
Cd ( Cadmium ) is tested on 0.6 ppb and should be in the CRM at 0.1ppb. Not only that it is not certified Value , the autors highlite it with a diviation of 500%.
While the real deviation is just 0.5 ppb ( 0.00005 mg/l ), you could also acount it as the most accurate Result in the test.
Who tells you that the deviation at 100ppb would not also be 0.5 ppb wich then would be just 0.5% relativ deviation ?
Manipulativ things like that are horrible to me, and do not belong into a scientific assament of Data.
Just fearing users of this is the only purpose of such a act.

We (users) need to see regular independent testing by third parties we trust, like the testing reported here, for us to have any confidence at all in such a tricky technique, and if you are not up front about the pitfalls with the technique you are using then you will not get the confidence that I am think you deserve for this using technique from users.

Well we have a lot confident users already, and the nicest thing to it is, in our case, the users don´t need sombody else they need to trust even not us.
Like I said in my MACNA speach.
I want a kind of reefkeeping that is based on KNOWLEDGE not on TRUST.
So the best way to get a Valid answer on "will the Test work for it´s purpose for me" , is using it for its purpose.
It is a bit like you would ask sombody else if the burger you want to eat will taste good or not... and the guy you ask take the burger to a lab do some tests , and even not eat it on his own... and then tells you " well probably it will not taste good".
Better just try a burger.
So why regular Tests from third parties you TRUST, why don´t you TRUST regular tests by yourselfe and the constitution of the Animals you keep ?
So many users worldwide have done such kind of tests on there own, as it works anonymous ( we have done that on purpose), or just asked us about accuracy of this and that, I can not at all understand why you want to just TRUST sombody else.

Additionally the way that you report 'below detection limit' as zero might be interpreted differently to how you intend too. For example with iron displaying ppb to 2 decimal places when your LOD is 0.3 ppb for instance, this is an error to display it like this. You display < 0.3 as 0.00, meaning that a true iron level of 0.24 ppb is displayed as being below 0.01 ppb. Thats a significant error in displaying the results. The same goes with all the elements in your reports I think. When I look at the results people publish, I see 0.00, 0.00 etc, when in fact I should be seeing <0.3, <0.5 etc. Its important because everyone interprets these results as meaning absolute zero, and so when they hear otherwise they are surprised and dismayed. I do feel you should make sure its clear in the results of the limitations of the technique for each element (IE next to each element result).

This is a very valid point I will change that asap.

If you don't explain to the users
that the limitations of testing using this technique are not yours but limitations of the physics of the test itself (that is, the test that you can do for that price, not the best accuracy that can be attained if you spend 1000 times more on a test and take a month to do it), then it looks like any inaccuracy is yours. This is simply what can be done for the price. If users want to pay far more I am sure you can offer a service for that and achieve greater accuracy.

We will do that explanation in the next Accuracy and LOD PDF the way you would like it i think Interference , best detection areas and worse case recoverys will be included wich will make the TEST even a better tool.
But such things need Time and need to be Tested under real conditions, so sorry if we are to slow in this point, but we try to be very exact.

I think triton testing is great and I look forward to using the service myself but I really, really want to know what the technical limitations are. So if you have high iodine then your accuracy of tin and zinc will drop, that kind of thing. Ballpark figures for how inaccurate results will be under certain circumstances etc. And until I see a fair few independent results confirming your reported accuracies then I will be quite cautious of believing the reports completely.

We have not reported any accuracies yet pete.

All the Best Ehsan

 

[Ehsan@triton]

Randy,

Thanks for putting your results out and going through them step by step was great to get a better understanding of all them at once.

Secondly i've seen a few posts from Triton that said the tests were not 100% perfect, and are to help us in the Reefing hobby more. It seems like you are a little down on this companies ability to offer this leg up in testing. If not Triton then what would you recommend using that is $50 and 10 minutes of work? :nerd:

I don't think I've said anything about the accuracy one way or the other.

Hi Dan,

I don´t feel that Randy is trying to get our abilitys down , i need to say that I am very thankfull to his work with helping the users of the Tests to understand it better.
And his acting IMO is very profesional and open minded.
A official thank you Randy :yo:


But regarding to the Article your point us absolutly Valid, and also belong to the Fit for purpose discussion. Like on the Cadmium amount in the Test, just because we have been 0.0005 mg/l (500% in that case) out is it better not to use that parameter at all ? what would be the Alternativ , not testing ???

All the best Ehsan

 

[Ehsan@triton]

Hi Ehsan,

My current job is a consultant to other companies. So I do not have access to any ICP machines.

OK no worrys

 

[al*]

Ehsan, if you talk about knowledge and not trust as the important part in our hobby there needs to be some way of independently confirming the values so an interested person can see what the limits for the tests are.

We have been doing this with salts or additives for years. If you can't test for e.g. calcium or alkalinity content in a salt you'll need to trust the manufacturer, you'll never know what's in it for certain. "The proof is in the pudding" doesn't really apply to the reef hobby, or we'd all be putting ginger in our tanks in order to get rid of ick or similar.

 

[Ehsan@triton]

I think this is a very valid point that should be quickly implemented. I assume the reason it is currently displayed at 0.00 is due to software/programming complications.



Ehsan, I believe the discussion for me is just 2 questions:

1) Why does Triton not instruct users to acidify their samples? What are the advantages/disadvantages of this?

2) Is the "certified sample", assumed to be tested by an outside laboratory, more or less accurate then the Triton testing? I assume it was "certified" by testing it using a similar process that Triton does.



I would like to also point out that the article may be somewhat misleading. The way I interpret it is that only some of the elements in the pure sample are certified and others are provided without certification. A quick comparison with the Triton results grid shows that most (not all) of the very divergent test results are from these non-certified elements.

Hi Sangheili,

1. to acidify a sample you would use 65% HNO3 (Nitric acid) suprapur garde, that is normaly used in ICP testing, You don´t want to do such things at home.
and we as a professional working company will never recommend such a thing to you as working ,storing... would be dangerous to you and your family.
PLEASE only work with such things under Professional Lab conditions.
For shipping ( in a professional way) dipending on the amount of acid used, it would be needed to propper label it.
At the end everything would get dangerous, complicated and expensive.
The question is why do you want to do it ? Is it realy needed ? how much more accuracy you will get out of that ?

TBH we could also acidify the sample when we start testing... if you wich us to do, but the real question is why?


2. Yes that is also what Randy already mentioned a bit.

here some info wich I would acount to the geek questions:

A. the Lab-devices used for certification are ICP-OES and IC / HPLC wich are pretty the same we have here. Just we don´t know about how old there devices are.

B. the Method they use is a internal developed laboratory method so "unknown"

C. the CRM is just 170$ wich is quite cheap when certified with the same devices like we have, and having the same problems.
how could such a cheap price be achived as, CRMs are not mass products usualy. And the production of such an standard is not at all easy ????

I could go on here but that will get boring

the Authors ( want to ) TRUST the other Laboratory more then they TRUST us.

I also need to say more and more i read the Article i find it writen in a Biased and Missleading way enbeded in a scientific jargon.

ALl the best Ehsan

 

[Thales]

Hi, it's Rich Ross, one of the authors of the article. I have stayed out of discussion of the article because Trition and Triton US said I was going to be contracted by their lawyers after they reviewed a pre publication version of the article. I would like to be involved in this discussion, but would like Ehsan's assurance that the lawyer threats are over. Ehsan, Joe, are the threats of lawyers still active or can we discuss without that sword hanging over our heads?

Thanks in advance.

Rich

 

[Ehsan@triton]

Ehsan, if you talk about knowledge and not trust as the important part in our hobby there needs to be some way of independently confirming the values so an interested person can see what the limits for the tests are.

We have been doing this with salts or additives for years. If you can't test for e.g. calcium or alkalinity content in a salt you'll need to trust the manufacturer, you'll never know what's in it for certain. "The proof is in the pudding" doesn't really apply to the reef hobby, or we'd all be putting ginger in our tanks in order to get rid of ick or similar.

HI Al,

I would like to be able to write like you but if you keep it a bit simple I can understand it better.

In reality " The proof is in the pudding " have only been the thing that reefers all around the world have done and TBH still do in this hobby. The put Aminos , Color up, Acro magic.... in there Tanks , they do Waterchanges with salts they even can not test for Bromid or Strontium in them or any Haevy metal... they claim all ths things good without Knowing anything about them....

And Now in the case of testing like you see abouve you can ether TRUST in us or TRUST in another Lab, wich one is correct you will never know ?
But if you just take a bit of time and effort, just add a bit of knowledged amount of Vanadium to a test and send the same sample without adding Vanadium or ( any other element of your concern ) you will know about your very own Accuracy in our test. That is why we made it Anonymous.
This would be a way to get real knowledge.

Our Testing will open up informations that will get you out of "the proof is in the pudding" a bit, But as a reefer I still beive in "The proof is in the pudding " much more then in a PHD ( sorry no ofense ).
I have a lot of PHDs in my Family and they all can not do what I reach in my tank. And also i have seen PHD using the Trust of people very often by writing lies in articles even in this Hobby.
I remember a PHD Chem. Writing there will no Aluminum lich out of Al PO4 remover , do you remember Randy ;-) : Chemistry And The Aquarium: Aluminum In The Reef Aquarium ? Advanced Aquarist | Aquarist Magazine and Blog
So I decided to do something against it and created the TRITON TESTING SERVICE to have a secoundery proof to the existing "proof i in the pudding".
Wich is working very well as in the iver 10000 test we caried out until now , lets say no tank with a 50 ppb of Pb ( lead ) realy looked good.
while a lot of tanks with similar results like the NSW we tested or our Tank look realy good. (Just a Chemical point of viwe)

I hope that is understanable

All the ebst Ehsan

 

[hart24601]

It is interesting where this will head. I like the changes in how results are delivered as UK Pete mentioned. I do think it's important to keep in mind while we can debate the accuracy, for detecting chemistry that is really out of balance it seems to work very well. Heck, I tested ESV and saw super high lithium. While not particularity toxic I contacted ESV and they confirmed with triton and another lab that yes some lots were contaminated. They also made some base changes to their formula to being it close to NSW levels (told to my by the president of ESV) but he didn't say what exactly. None of that would have happened without triton testing.

 

[UK_Pete]

Ehsan thanks for your replies.

I hope you don't feel that my points were too critical because thats not how they were meant. I think you offer an honest and unique service which is extremely useful. But I think that since you are currently the only company to offer this at a practical price, you are vulnerable to be blamed for the limitations of the technique, since theres no one else to compare you to. Its not possible to say 'all the companies have this particular limitation in their testing'. And when I say limitation I fully appreciate your comment that the real variation of say Cd is only a very small amount, IE 0.5 ppm, but I think the problem is that users hope to know exactly what the possible error margins are, since they are often interested in these very small differences - frequently thats what draws them to do a triton test, to determine things below the detection limit of their other tests.

When you say that the article in question is manipulative, I partly know what you mean - it might be a bit sensationalist, so to speak. But I think thats to be expected since you stand out as the highest authority as far as testing goes for ordinary hobbyists, and the authors did turn up a worrying result. You are always going to be under scrutiny since everyone holds triton testing up as always correct. IE if someones home test kit reads a different value to their trition test, they will immediately say the triton test is correct, and often they will come on forums and let everyone know about that. So if the triton test then turns out to have flaws which have not been disclosed, everyone is disappointed. Having said that though I think the evidence currently is that the article confirmed that the triton test IS indeed always better than the home tests, with the small caveat that some elements are not quite as accurate as everyone thought. Or perhaps its just that at close to the limit of detection, some greater deviances creep in.

Regardless though I do believe you need to explain this so no one is surprised. IE make sure that you say what the potential errors are and of what magnitude. Then, if an article like the one in question is published, you can point to your documentation and say 'we have already disclosed and explained this'.

Regarding the hamburger, I understand your feeling but I slightly disagree. Lets say we are interested in the pesticides in the hamburger. It might taste good but we want to know if its got anything poisonous in it. With triton testing, we might have a problem in the tank, and want to know why it might happen. We might jump on the high Cd, since its so toxic, and say 'there is our problem'. When in fact the Cd is 5 times lower than we expected and that factor of 5 might be significant, so we might blame the high apparent Cd and not fix the real problem in the tank.

Having said that, as far as the 'fit for purpose' question goes, there was never any question for me that triton testing is fit for purpose. Its clearly more accurate than home test kits, so its clearly a vast improvement. Thats its purpose. It might however not be fit for the purpose of scientific research of, for instance, iron depletion in an aquarium. Since I have been recently interested in iron, this is something I was interested in triton testing for. A quick look at triton test results shows iron below 0.01 ppm in many tanks, when the iron might actually be higher as we discussed (But you already said you would address that which is great). But is triton testing suitable to determine if iron is getting to limiting concentrations in tanks? No, as its not sensitive enough. But triton is fit to greatly improve peoples understanding of whats in their tanks, which I see as its purpose.

As far as the trust issue goes, I guess from the bold type you might feel offended by the suggestion that users need independent testing to trust triton. I would like to point out that independent testing is normal in most aspects of goods today, with regards to food, drink, consumer goods, etc, including tank equipment etc. Independent verification of quality is ubiquitous in todays market place, taking aquarium goods, people are using triton to test aquarium salts for instance. And recently a salt was shown to have a quality control issue which was caught by triton via a hobbyist. As a result the issue was fixed and potentially tank crashes were avoided. If people had not tested the water, perhaps they would have found out the problem in the end, but it might have been through tank crashes. Testing attempts to avoid that. Independent testing also increases confidence in services like yours and I really think you should welcome it. If you try to prevent independent testing, it makes triton look like they have something to hide, which I don't believe you do.

BTW for me, the articles reported issues with triton do not worry me much, as long as I know the results (which elements I should be wary of trusting tritons figures for). I was more worried and surprised when I read about the fact that triton do not welcome independent testing, and are so closed with respect to information about their techniques. I appreciate that some information has to remain yours for commercial reasons but I would like so see everything else disclosed. IE whatever is not commercially sensitive. The less that is disclosed, the more important independent testing is IMO, because if you don't say how you are testing, no one else can judge your methods reliability. A process that is not disclosed or independently tested is of far less use than one that has been proven to accurate and reliable.

Best regards, Pete

 

[UK_Pete]

Ehsan, I don't want to bombard you with too many posts but just wanted to comment on these 2:

the Authors ( want to ) TRUST the other Laboratory more then they TRUST us.

I think its obvious that here we have a typical difference between 2 sources (triton and 1 independent tester). We need more to know who is right, hence the need for several parties to conduct independent tests as I said I felt would be required. Thats also to protect you from erroneous or incompetent testing from independent testers.

The put Aminos , Color up, Acro magic.... in there Tanks , they do Waterchanges with salts they even can not test for Bromid or Strontium in them or any Haevy metal.

Thats what we are looking at ending now with triton IMO. Hence the need to hold triton up to a high standard. I hope that eventually triton testing and similar allows the reefing hobby to rise out of the medieval magic potion era we are in and into a new era of scientific keeping where we all understand what we are doing and have no further need for voodoo. But I do feel that triton has to be proven first for its results to be held as irrefutable, and checked (quality sometimes slips and regular verification and checks keeps us all in line and the quality high).