Page 1 of 2

Glyphosate is not reproducible

Posted: Wed Mar 13, 2019 4:07 pm
by PaulaL
Hi,

I am having trouble with analyzing and developing a method for glyphosate.

I had the method working and could see glyphosate down to 0.01mg/kg and then all of a sudden the peak deteriorated and I haven't been able to get it back.

I am using a Shimadzu LC stack with the mass spec; Sciex API 4000. An Acclaim Trinity Q1 column 3um 3x100mm (new column)

Mobile phase A: 50mM ammonium formate, formic acid 2.9pH, H2O:ACN 6:4, Mobile phase B:ACN

Gradient Binary Flow: 0 mins 100% A, 10.1 mins 10% A, 13.1 mins 100% A, 18 mins 100% A

Flow rate of 0.5ml/min

(this is from the EU reference glyphosate method: http://www.crl-pesticides.eu/userfiles/ ... urlSRM.pdf)

I prep the samples using the QuPPe method therefore, it is not derivatized.

Since the problem arose, I have made fresh mobile phases and cleaned the ion source. I don't think it's a detection problem as the detector works on other methods.

Does anyone have any idea why it was working and now it is not and how to get it working again? Any help would be appreciated.

Re: Glyphosate is not reproducible

Posted: Wed Mar 13, 2019 5:14 pm
by Vlad Orlovsky
Do you see it at higher concentration or you don't see glyphosate at all? Are your injections from the same vial or you made new samples? Did you try to inject glyphosate without the column to see if it related to column or HPLC system?

Re: Glyphosate is not reproducible

Posted: Wed Mar 13, 2019 5:47 pm
by PaulaL
Thanks so much for replying, this glyphosate is driving me slightly mad!

When the samples are being extracted there is barely a peak and very noisy at 2mg/kg. When I run a solvent standard at 2mg/kg I am getting a better peak but the sensitivity is still terrible and isn't anywhere near as close to where I had it last week with the sensitivity for an extracted 2mg/kg sample.

I have tried lots of new extractions, tried some from old vials in case it was a new extraction but they all seem to give the same terrible noisy peaks.

I've injected the solvent glyphosate std without a column and it is present, when I tried taking the column out the ferrule was really stuck in the column and was a hard job to get out. I'm just about to change that and see if the problem is just as simple as having to change a ferrule.

Edit: Unfortunately it was not as simple as changing the ferrule. To give you an idea of what I had last week compared to now I'll try and attach the images below. Image

Re: Glyphosate is not reproducible

Posted: Wed Mar 13, 2019 7:33 pm
by Vlad Orlovsky
is it the same response ad with the column? If it is more without column, then you column traps glyphosate or noise level from the column reduces you sensitivity.

Re: Glyphosate is not reproducible

Posted: Wed Mar 13, 2019 8:27 pm
by PaulaL
Those are both with the column, the same column. The one on the left was from last week when the method was working and the other is from today. The one on the left is actually only 0.6mg/kg and the one on the right is 2mg/kg.

Re: Glyphosate is not reproducible

Posted: Thu Mar 14, 2019 7:31 am
by Rndirk
I agree with Vlad, it seems a good test to replace the column with a zero-volume connection and do the same injection of a solvent standard. Don't forget to adjust the MS method for this test: no divert valve, and your analyte will obviously elute earlier.

On a sidenote, I was interested in your column, googled it and I have to say it's quite a special one (for example, it can not be used with MeOH). Your mobile phase choice seems to be compatible, altough the pH might be a bit close to the lower limit. Let's see what the no-column test gives?

Re: Glyphosate is not reproducible

Posted: Thu Mar 14, 2019 10:10 am
by Gaetan Glauser
The link below might be interesting for your issue:

http://www.waters.com/webassets/cms/eve ... ebcast.pdf

In slide 31 they mention strong peak tailing possible due to metal ion contamination which they could mitigate with an acid wash. No more details though,.

Re: Glyphosate is not reproducible

Posted: Thu Mar 14, 2019 3:09 pm
by PaulaL
So without the column the intensity of the solvent peak is lower than with the column.

Thanks for that link, it looks as though that wash could be of help, very annoying they don't say what sort of acid wash they use, that would've been very handy. I had a google and the Acclaim trinity manual has its own suggestions of washing the column so I will give that a go. https://assets.thermofisher.com/TFS-Assets/CMD/manuals/Man-065475-Acclaim-Trinity-Q1-Man065475-EN.pdf

It is a special column. From doing research it seemed that this column gave the best results for glyphosate so we purchased it hoping it would be a case of using the mobile phases that it suggests along with the same gradient and doing the QuPPe extraction. It did work very well for about a week and then the peak just started to get worse and become inconsistent and not usable. I think that's the most frustrating part, I had it working.

It could be a case that the column needs to be maintained regularly in order to give good chromatography. I shall see what happens after doing the washes they suggest.

Re: Glyphosate is not reproducible

Posted: Thu Mar 14, 2019 5:33 pm
by Andy Alpert
Possibly your column is interacting with the phosphonate group. Passivate the metal surfaces of your column (as well as the rest of the HPLC system) by eluting it overnight at a low flow rate with 40 mM EDTA.2Na. Next morning, flush out the EDTA, condition your column, and evaluate the recovery of injected glycosate.

Re: Glyphosate is not reproducible

Posted: Thu Mar 14, 2019 6:33 pm
by James_Ball
In the past we used a Hamilton column for Glyphosate along with one from Pickering and both needed to be cleaned at times due to iron contamination. Usually a few injections of 100ul of about a 5% Acetic acid solution would do the trick. Have to see if that would be compatible with this column you are using.

Re: Glyphosate is not reproducible

Posted: Tue Mar 19, 2019 5:30 pm
by PaulaL
Thanks for the replies. Cleaning the column over night did not seem to help. I'm having to resort to derivatization and hopefully I can now get consistent results.

Currently prepping the samples now so I will have to wait and see.

Thanks for all the help :)

Re: Glyphosate is not reproducible

Posted: Wed Mar 20, 2019 7:01 am
by Rndirk
PaulaL wrote:
So without the column the intensity of the solvent peak is lower than with the column.

It could be a case that the column needs to be maintained regularly in order to give good chromatography. I shall see what happens after doing the washes they suggest.


This means that the column is not the problem, or do i misinterpret the line in bold above?

Re: Glyphosate is not reproducible

Posted: Fri Jun 21, 2019 2:05 am
by nchamkasem
Andy Alpert is correct. Glyphosate is a chelating agent and will bind with trace iron or metal in the HPLC flow path. I am the author that method 1.5 in the QuPPe method version 10.1 is using. I have a paper to demonstrate that glyphosate will have poor peak shape in the HPLC system with a trace amount of iron. To fix this is, one, passivate the flow path with EDTA or two, add EDTA in the mobile phase, or three, add EDTA in the extracting solvent. QuPPe method use 1:1 water:methanol which will extract lots of non-polar interference and will cause matrix suppression. Adding EDTA in the mobile phase will cause significant matrix suppression as well. Contact me and I will send the ref paper to you. I developed the method on Q1 to cover 20+ polar analytes not retain on the C18 column from glyphosate (neg) to paraquat (pos). Please check my ReserchGat publications.

Re: Glyphosate is not reproducible

Posted: Fri Jun 21, 2019 7:17 am
by Rndirk
nchamkasem wrote:
Andy Alpert is correct. Glyphosate is a chelating agent and will bind with trace iron or metal in the HPLC flow path. I am the author that method 1.5 in the QuPPe method version 10.1 is using. I have a paper to demonstrate that glyphosate will have poor peak shape in the HPLC system with a trace amount of iron. To fix this is, one, passivate the flow path with EDTA or two, add EDTA in the mobile phase, or three, add EDTA in the extracting solvent. QuPPe method use 1:1 water:methanol which will extract lots of non-polar interference and will cause matrix suppression. Adding EDTA in the mobile phase will cause significant matrix suppression as well. Contact me and I will send the ref paper to you. I developed the method on Q1 to cover 20+ polar analytes not retain on the C18 column from glyphosate (neg) to paraquat (pos). Please check my ReserchGat publications.


Interesting post! I was not aware that EDTA is an LCMS-compatible additive. Can you comment on the concentration of EDTA in the mobile phase, and possible negative effects? I haven't been working with glyphosate myself, but I was getting the same problems with similar analytes - which fade (briefly) upon washing the system with EDTA (as in the post of Andy).

I found this Agilent poster that shows some results with/without EDTA, and they apparantly found medronic acid to be superioir additive with the same function.

It may sound absurd, but perhaps glyphosate itself could be a candidate as a mobile phase additive - if the goal is not to measure glyphosate of course, but compounds with similar problems.

Re: Glyphosate

Posted: Fri Jun 21, 2019 1:06 pm
by nchamkasem
Please check out this paper (Anal Chem 2018 page 9457) This paper adds 5 uM of EDTA in the mobile phase B on a HILIC column. This method will have EDTA thru out the ran and eluted as a hill. It will suppress the signal of many analytes. It is better to add in the extracting solvent to minimize the glyphosate-metal binding in sample containing metal ions such as calcium in milk. At the same time, when you inject the sample (20 uL), it will prevent glyphosate from binding with metal (Fe) present in the flow path. This paper also shows that you will not have bad glyphosate peak in the non-metal HPLC system. The concentration in my solvent which is 100% water is 10 mM EDTA + 50 mM acetic acid). The acid is to precipitate protein (in milk). I used Acclaim Q1 in the anion exchange mode, not HILIC mode as described in the method 1.5