Pesticides using API2000 LC/MS/MS info

[-zwitterion+]

We have an API2000 LC/MS/MS that we are getting ready to set up. We are basically novices in LC/MS/MS. We intend to test for a list of approx 60+ pesticides down the road. Our intentions are to get trained on this particular model that we have and then maybe purchase an API4000 later on.

Questions I have are:

We need to reach mostly 10-20 ppb levels (with some higher exceptions).
Can the API2000 detect those levels?

Can the API4000 detect those levels?

I have read somewhere else that the most current LC/MS/MS technology is barely achieving the levels mentioned above. And when I say current technology I am referring to Shimadzu's 8060 and Agilent 7000 series latest LC/MS/MS units.

I am somewhat confused since technical papers on the older instruments talk about 1 pg detection levels.

Any help/suggestions/guidance will be greatly appreciated. BTW, API4000 units are listed on eBay from 90K all the way to close to 140K. Some with warranty and complete some without and incomplete set up. I have a quote of about 140K from a reliable vendor with 1 yr serv contract, complete set up. So to throw in another question, for this kind of money, is this the best unit?

In advance, I thank you for your help.......

[Rndirk]

I'm not experienced in the differences between the models you mention.

You could ask SCIEX but your information is not sufficient.

1) What kind of samples do you want to analyze? Are all the pesticides compatible with the LC-MS/MS conditions you're planning? Do you want to measure them in a single injection? Is it a screening or a quantification?

2) 10-20ppb are pesticide levels on sample basis (I assume): for example 10 µg/kg on a certain vegetable. This is not very relevant to compare different detectors.

It depends on your sample preparation and chromatographic conditions. If you want to compare different detectors, it's better to speak in terms of concentration of final extract or better, absolute amount of pesticides on-column.

For example (assuming 100% recovery in extraction):

- There is 10 ppb of pesticide X in vegetable Y (10 µg/kg).
- Your sample prep consist of 1g sample ending up in 10mL of solvent => this leads to 1 ppb in the vial you're going to inject (1 µg/L = 1 pg/µL)).
- You inject 10µL so there is 10pg on-column

The number of 1pg you have found in the paper you mention is most likely the amount on-column.

In the above example you see that you there is room for adjustments to develop the method for the limit your detector can handle. Adjusting the extraction or the injection volume can easily lead to 100 or 1 pg on column.

Be aware that a higher concentration/volume injected means more matrix (dirt) injected and a higher chance of interferences & more maintenance is needed, possible overloading of the column/detector,... This is the advantage of having of a more sensitive instrument.

I have read somewhere else that the most current LC/MS/MS technology is barely achieving the levels mentioned above. And when I say current technology I am referring to Shimadzu's 8060 and Agilent 7000 series latest LC/MS/MS units.

This is wrong. Do you have a source?

Google application notes for pesticide analysis using LC-MS/MS: you will finds dozens.

[-zwitterion+]

Rndirk, thx for the reply, to answer one of your questions, this all has to do with pesticides testing in plants .... Based on your on-column calculation example and a Restek's Quechers sample preparation for this type of analysis, it will go like this:

Assuming we are looking for a final Reporting Limit of 10 ppb ...

Preparation is as follows:

1 gram of sample
20 mL of solvent (10mL water/10 mL ACN)
10 grams of some Quechers salts
[ Up to this step DF=20 ]

From above take 6 mL, vortex and centrifuge.
Then take 2 mL (Out of the 6 mL) and dry under nitrogen
then reconstitute to 1 mL
[This step has a Concentration factor of 2]

Total DF = 10

Therefore, I need to see 1 ppb on-column, multiplied by 10x, the extraction factor, that 1 ppb will become 10 ppb reporting limit. And as you indicated, there is room to play with dilution factors by increasing sample size, injection volume, etc......

So in conclusion, I need to find out if this instrument can detect at a minimum 1 ppb on column for the pesticides of interest....

Rndirk, or anybody else, please confirm that what I stated/concluded/calculated is correct.

Thank you......

[Rndirk]

First step in your calculation has a dilution factor of 10, not 20.

In Quechers, water and acetonitrile phases separate after adding salt. So the pesticides in 1g of sample go to 10ml of acetonitrile.

Your final extracts, at the reporting limit, will have 2ppb (µg/L = pg/µL) of pesticides following that protocol.

If you choose to inject 10µL of 2 pg/µL, then you put 10*2 = 20 pg on-column.

Be aware that there are orders of magnitude in differences for detector sensitivity among pesticides. If you can freely choose your pesticide selection (which is rarely the case) you might get by with an older and cheaper MS!

[-zwitterion+]

Rndirk, OK, thx for the correction on the DF, I have never used Quenchers and thanks to you I just learned one important fact about their use.

I found this interesting MDS SCIEX presentation that compares several API LC/MS/MS instruments.... Here is for those interested in checking it out...

http://www.traceorganic.com/2008/presen ... obarah.pdf

API2000 sensitivity wise does not look good compared to the API4000 or API5000....Bottom line I think that I just need to do a DL study as well as a spike and recoveries study based on the method to determine the capabilities of the instrument for that particular test.

API2000 is from around 1997, the API5000 was built around 2005

[Rndirk]

Another heads-up: instrument manufacturers will naturally push you towards buying their newer models, it's their job. I doubt they will offer you anything below the TQ or QTrap 4500.

Their newer models *are* better (in terms of sensitivity and selectivity), but i think it's a good strategy of you to start off with the system you have and get experience with playing around with sample prep and instrument parameters.

Coming back to your original question: In my personal experience, i think 140K for an API4000 is a lot. It was probably a good deal 10 years ago.

[James_Ball]

I am not sure of the difference in the API2000 versus API4000 but on the books I know the API4000 has about 5x the sensitivity as the API3200 which we have, and the API5000 is about 10x of the API3200. The difference in sensitivity is the length of the quad rails and the vacuum system. More ultimate vacuum gives better sensitivity and longer quads gives better resolution.

I was working on a pesticide list of over 150 analytes in a tobacco matrix a few years back on the API3200 all in one run.

For an example I did 3g sample to 15ml solvent and I could take Malathion down to 0.0013ug/g(mg/kg) which would be 1.3ppb which is a detection limit based on having a peak on the instrument that gave a minimum of 2000 area counts. You can see lower than that if you have low noise but it was a good practical and reproducible detection and reporting limit. Carbamates all gave detection limits around 10ppb but something like Captan was closer to 1ppm due to solvent solubility(it works better as a GCMSMS analyte).

You won't be able to get all of your analytes down to the same limits and the limits will vary quite a bit among a list as large as yours so be prepared for that.

Restek has standard kits for both GCMSMS and LCMSMS, look at those and you will see what works better on each instrument and which ones work on both. Their applications also give pretty good ideas of what you can expect on the instrument, most were developed on the API3200 or API4000 I believe.

[mckrause]

The 2000 is a less-than-ideal system for QuEChERS work. It's switching speed is not great and the TurboV sources didn't come out until the 3200. You can definitely get 1-2 ppb with conventional QuEChERS on a 4000; actually, a 3200 is more than sufficient. The 2000 also has outdated software that you will absolutely not be happy with.

Your example of a QuEChERS extraction is flawed. A conventional extraction for fruits and vegetables (the matrices QuEChERS is validated for) is 15 grams of sample extracted with 15 mL of acetonitrile. You then can perform a solvent exchange. Depending upon the pesticides you are interested in you might be able to concentrate the extract if needed, say 5 mL of the ACN evaporated and reconstituted in 1 mL of 80:20 water:ACN. This is completely dependent upon the pesticide and the matrix.

There is no reason that I can think of that you would only extract 1 gram of sample unless you are working with a matrix that is non fruit/vegetable (such as marijuana) that has extraordinarily high co-extracted material. In any case you would still extract the 15 grams (since that is the validated method) and then dilute the extract.

We currently dilute most of our QuEChERS extracts a factor of 5-10 before running them in order to avoid matrix issues. We are running older MicroMass triple quads and have plenty of sensitivity to see 1-2 ppb of >95% of the pesticides we look for on LC-MS/MS. Be advised that most of the issues you will run into have to do with ionization source limitations; we still heavily rely on GC-MS since many of the pesticides of interest do not ionize under ESI conditions.