GC-MS Analysis of Metaldehyde

[Ento_Joe]

Hi all,

I am currently involved in a project looking to investigate the contamination of water with metaldehyde, a common pesticide used for slug and snail control in many parts of the world, but seem to be having some issues with the instrumental analysis.

The extraction and analysis method I have been using is described in pages 8 to 13 of the UK Environment Agency's methods paper, found here: http://webarchive.nationalarchives.gov.uk/20140328104725/http://www.environment-agency.gov.uk/static/documents/Research/Metaldehyde-226b.pdf. However, I'm having some issues replicating their instrumentation settings and ultimately the reported limits of detection.

Firstly, within the document linked above the authors state they used helium as the carrier gas at a rate of 30 ml/sec, however I'm unsure how that translates into ml/min as it seems the flow would be insanely fast?

Secondly, the authors state they injected 1 ul of sample using a pulsed splitless method, however they don't give any further detail. The GC asks for the pressure and duration of the pulse, which I can't replicate. Any suggestions?

I'm guessing that I'm struggling to replicate their LODs because I can't fully replicate their instrumentation settings accurately. I am unable to detect any metaldehyde below 10 ug/L at present, but as the EU limit for metaldehyde in drinking water is 0.1 ug/L this isn't good enough for my study.

Does anyone have any suggestions as how to progress?

Thanks,

Joe

Potentially useful papers:
http://pubs.rsc.org/-/content/articlepdf/2017/ew/c7ew00039a
http://www.mdpi.com/2073-4441/7/6/3057/htm

GC-MS instrumentation used by me:
Agilent 7890 B Gas Chromatograph coupled to an Agilent 5977 A Mass Spectrometer

[Rndirk]

Firstly, within the document linked above the authors state they used helium as the carrier gas at a rate of 30 ml/sec, however I'm unsure how that translates into ml/min as it seems the flow would be insanely fast?

Indeed, it's impossible that the 30ml/s is a carrier flow over their 0.25mm column.

I wouldn't worry too much about their flow and use 1-2 ml/min of He gas on a 0.25mm column.

Secondly, the authors state they injected 1 ul of sample using a pulsed splitless method, however they don't give any further detail. The GC asks for the pressure and duration of the pulse, which I can't replicate. Any suggestions?

Pulsed splitless sometimes helps to push the analytes on your column. It's splitless, but after the sample is introduced the pressure is increased for a short time. Typically a couple of seconds at a pressure 2-3x the normal pressure.

I suggest to start with normal splitless.

I'm guessing that I'm struggling to replicate their LODs because I can't fully replicate their instrumentation settings accurately. I am unable to detect any metaldehyde below 10 ug/L at present, but as the EU limit for metaldehyde in drinking water is 0.1 ug/L this isn't good enough for my study.

Do you take into account the method starts from 250-500ml sample and concentrates it by SPE to a couple of ml? This means the 100ng/L can be 50µg/L on the instrument.

They express the concentration in the calibration curve directly as concentration on sample basis, this can be confusing and i personally don't like it.

[Ento_Joe]

Hi Rndirk,

Thanks for your response, much appreciated.

I had assumed that 30 ml/sec was unrealistic, thank you for confirming my suspicions! I had been trying this method using a 1 ml/min flow rate anyway, so that doesn't seem to be the issue.

Running samples using normal splitless doesn't seem to detect metaldehyde levels belo 10 ug/L. I may try the pulsed splitless method on Monday when I'm back in the lab. A paper I found after posting my initial question adapts the methods of the Environment Agency note I posted earlier to this:

GCMS was performed using an Agilent 789A GC interfaced to an Agilent 5975C MSD mass spectrometer operating with electron ionisation at 40 eV and scanning from m/z 50 to 600 at 2.7 scans/s based on a protocol adapted from the Environment Agency for determination of metaldehyde in water. The GC was equipped with an Agilent 7683B auto sampler and a programmable temperature variation (PTV) inlet. The samples were injected in pulsed spit-less injection mode (1 µL; inlet pressure of 25 psi for 0.25 min) and separated on an HP-5 capillary column (J&W scientific column 5% diphenyldimethyl polysiloxane; length 30 m, I.D. 250 µm, film thickness 0.25 µm). The samples were run at a constant flow (1 mL/min) with He as a carrier gas. The heated interface temperature was set to 280 °C, with the mass source temperature at 230 °C and the MS quadrupole at a temperature of 150 °C. The samples were injected at 35 °C and the oven was programmed to 260 °C at 20 °C/min at which it was held isothermally for 0.5 min. Peaks were acquired both in scan and SIM mode at ion peaks of m/z = 45 and 89.

This gives me something to work with regarding the pulsed splitless settings.

How would you express the concentration? I hadn't taken that into account, but I wasn't working with 250 ml samples either. I actually only pushed 25 ml through as the vacuum manifold for my SPE tubes packed in and it took too long, plus this was only a preliminary investigation to see if I could get the GC-MS method to work satisfactorily.

[James_Ball]

I don't understand them using 40eV ionization, most normal GC/MS work is done at 70eV.

The 30ml/s is probably supposed to be 30cm/s, which would be in the range of 0.8ml/min He, which is much more believable.

Also the SIM mode will help you get lower detection limits on the instrument.

Most analysts will use the concentration of the standard that is being injected onto the instrument such as 10ppm would be 10ug/ml or 10ng/ul. If you use 10ug/ml and you do an extraction that is 1000ml sample through solid phase and you elute with solvent, evaporate and bring to final volume of 1ml, then the 10ug/ml is equal to 10ug/L. If your instrument can detect 0.1ug/ml(ppm) then the sample concentration at 1000ml and final extract volume of 1ml would be 0.1ug/L(ppb)

If working with 1L or 500ml sample sizes it can take 1-3 hours to pull the sample through the SPE cartridge, then another hour to concentrate down the elution solvent(don't want to evaporate too quickly, you could lose your analytes).

I have been working on a drinking water project using EPA 525.3 and if I can get started by 8:30 I can get a set of 10 samples ready for the instrument by 16:00 if I am lucky.

[Rndirk]

Hi Rndirk,

Thanks for your response, much appreciated.

I had assumed that 30 ml/sec was unrealistic, thank you for confirming my suspicions! I had been trying this method using a 1 ml/min flow rate anyway, so that doesn't seem to be the issue.

Running samples using normal splitless doesn't seem to detect metaldehyde levels belo 10 ug/L. I may try the pulsed splitless method on Monday when I'm back in the lab.

I actually only pushed 25 ml through as the vacuum manifold for my SPE tubes packed in and it took too long, plus this was only a preliminary investigation to see if I could get the GC-MS method to work satisfactorily.

I suggest to start developing the method without doing the sample preparation. So first check if you can produce a satisfactory peak of the compound in a standard solution (use the same solvent as you'd inject by doing the sample preparation).

Determine how low you can get in concentration in terms of the solution you're injecting. If you know that, for example, your instrument can measure 10µg/L reliably, then you know that extracting 100mL water sample to 1mL organic solvent can give you 0.1µg/L. You will have to validate that of course!

I agree with James that you will want to use SIM. You can either use the ions from the paper or determine them yourself by running in full scan and then selecting the ones you get. They should be the same as EI-MS is quite reproducible but it doesn't hurt to check.

[rb6banjo]

I know that Wikipedia is not always the best source of information but for things like this, I've found it to be pretty valuable. I'd be alarmed that it supposedly decomposes at 80 °C. A 300 °C inlet temperature isn't good.

[Ento_Joe]

I don't understand them using 40eV ionization, most normal GC/MS work is done at 70eV.

The 30ml/s is probably supposed to be 30cm/s, which would be in the range of 0.8ml/min He, which is much more believable.

Also the SIM mode will help you get lower detection limits on the instrument.

Most analysts will use the concentration of the standard that is being injected onto the instrument such as 10ppm would be 10ug/ml or 10ng/ul. If you use 10ug/ml and you do an extraction that is 1000ml sample through solid phase and you elute with solvent, evaporate and bring to final volume of 1ml, then the 10ug/ml is equal to 10ug/L. If your instrument can detect 0.1ug/ml(ppm) then the sample concentration at 1000ml and final extract volume of 1ml would be 0.1ug/L(ppb)

If working with 1L or 500ml sample sizes it can take 1-3 hours to pull the sample through the SPE cartridge, then another hour to concentrate down the elution solvent(don't want to evaporate too quickly, you could lose your analytes).

I have been working on a drinking water project using EPA 525.3 and if I can get started by 8:30 I can get a set of 10 samples ready for the instrument by 16:00 if I am lucky.

Hi James,

Thanks for the response and apologies for my delayed reply.

I'm unsure as to why the authors are using 40 eV ionisation, but I have kept mine at 70 eV during analysis.

Having only worked with 25 ml samples thus far and it taking the best part of an hour, I can only imagine how tedious working with larger samples can be!

All the best,

Joe

I suggest to start developing the method without doing the sample preparation. So first check if you can produce a satisfactory peak of the compound in a standard solution (use the same solvent as you'd inject by doing the sample preparation).

Determine how low you can get in concentration in terms of the solution you're injecting. If you know that, for example, your instrument can measure 10µg/L reliably, then you know that extracting 100mL water sample to 1mL organic solvent can give you 0.1µg/L. You will have to validate that of course!

I agree with James that you will want to use SIM. You can either use the ions from the paper or determine them yourself by running in full scan and then selecting the ones you get. They should be the same as EI-MS is quite reproducible but it doesn't hurt to check.

I know that Wikipedia is not always the best source of information but for things like this, I've found it to be pretty valuable. I'd be alarmed that it supposedly decomposes at 80 °C. A 300 °C inlet temperature isn't good.

Hi Rndirk,

Thanks for your response and apologies for my delayed reply.

I think you are right, the best place to start developing the method is without sample preparation. I made up a series of metaldehyde dilutions today (in dichloromethane): 1000 ug/L, 100 ug/L, 10 ug/L, 1 ug/L, 0.1 ug/L, and 0.01 ug/L. All were detectable using SIM mode for ion peaks at m/z 45 and 89, although I suspect I should increase the number of ion peaks I am detecting.

I had noticed that metaldehyde decomposes at a relatively low temperature, but assumed the authors of the Environment Agency report had factored this into their method. Should I be using a lower injection temperature do you think? If so, what should I try?

All the best,

Joe

[rb6banjo]

The US EPA methods are usually worked out pretty well before they are published BUT there are often times some of those "why didn't they think of that?" moments. I know nothing about the scrutiny the methods experience before they are published.

I recall from previous exchanges (the mite problem) that you have a thermal desorption system. This seems like a perfect opportunity to use TD and the Gerstel "Twister" technology for extracting the metaldehyde from the water. You have almost complete control over the temperature used to desorb the analyte from the stir bar. No solvent needed to get the metaldehyde out of the water.

It looks like there's a precedent for my approach. Anatune in the UK:

http://www.anatune.co.uk/wp-content/upl ... 122013.pdf

[James_Ball]

Looks like it sublimes at 120C and melts at 246C so you could probably get by injecting at 180C. At least that would be a good starting point, then do injections of the same concentration going up or down 20C at a time and see if the response increases or decreases. Also experiment with the split valve timing until you see a decrease in response.

You may also want to try using a Uniliner which will put more analyte on column and keep it from contacting the hot metal surfaces.

[Rndirk]

I think you are right, the best place to start developing the method is without sample preparation. I made up a series of metaldehyde dilutions today (in dichloromethane): 1000 ug/L, 100 ug/L, 10 ug/L, 1 ug/L, 0.1 ug/L, and 0.01 ug/L. All were detectable using SIM mode for ion peaks at m/z 45 and 89, although I suspect I should increase the number of ion peaks I am detecting.

µg/mL, right?

If you get a good signal and calibration, i wouldn't worry about the possible decomposition unless you want to go lower.

[Ento_Joe]

The US EPA methods are usually worked out pretty well before they are published BUT there are often times some of those "why didn't they think of that?" moments. I know nothing about the scrutiny the methods experience before they are published.

I recall from previous exchanges (the mite problem) that you have a thermal desorption system. This seems like a perfect opportunity to use TD and the Gerstel "Twister" technology for extracting the metaldehyde from the water. You have almost complete control over the temperature used to desorb the analyte from the stir bar. No solvent needed to get the metaldehyde out of the water.

It looks like there's a precedent for my approach. Anatune in the UK:

http://www.anatune.co.uk/wp-content/upl ... 122013.pdf

I'd love to be able to purchase something like this to give it a try. The main problem I have is that I'm conducting this work as a proof of concept to write a full grant proposal, so have little money to spend!

Looks like it sublimes at 120C and melts at 246C so you could probably get by injecting at 180C. At least that would be a good starting point, then do injections of the same concentration going up or down 20C at a time and see if the response increases or decreases. Also experiment with the split valve timing until you see a decrease in response.

You may also want to try using a Uniliner which will put more analyte on column and keep it from contacting the hot metal surfaces.

I have been injecting samples at a reduced inlet temperature today, with relatively good peaks in SIM mode. I will try and post some chromatograms tomorrow. The split valve timing is set to 50 ml/min at 1 minute, any later than that and I start to see solvent appearing after my 5 minute solvent delay.

I have never heard of a Uniliner, what makes it more efficient at putting analyte onto a column compared to other liners?

µg/mL, right?

If you get a good signal and calibration, i wouldn't worry about the possible decomposition unless you want to go lower.

I have made the solutions in ug/L concentrations as the EU limit for metaldehyde in water is currently set at 0.1 ug/L. I can detect all the concentrations previously mentioned, but the calibration curve I make from these points isn't a particularly straight line!

[Rndirk]

I have made the solutions in ug/L concentrations as the EU limit for metaldehyde in water is currently set at 0.1 ug/L. I can detect all the concentrations previously mentioned, but the calibration curve I make from these points isn't a particularly straight line!

I find it hard to believe that you get a signal for a 1µL injection of a 0.01 µg/L = 10 ng/L standard with a GC-MS single quad (= 10 fg on column)

Remember that the EU limit you mention is on the sample, this does not mean your calibration curve has to go that low if your sample prep concentrates that sample.

[rb6banjo]

i doubt you need that "multiflex" sampling apparatus. I bet all you really need is a pdms twister stir bar and an empty thermal desorption sample tube.

If you can't buy the twisters, perhaps you could make your own. I bought some PDMS tubing from VWR (surgical tubing, very small i.d., 0.5 mm). Get some wire that has the same o.d.. Insert the wire into the tubing and make them however long you want. The wire gives it physical strength. If your vision is good, you can probably just use a small section of the tubing with no wire. Condition the tubing in the TD apparatus to remove the volatile materials. Throw your clean sampling device in the water samples and stir for whatever time you need/want. Analyze with your TD method. I was worried about putting too much liquid (inside the pdms tubing after sampling) in the sample tube prior to TD analysis. That's why I went with the wire support - plugs the interior of the pdms tubing.

My tubing is Sanitech Platinum-Cured Sanitary Tubing (Cat# 75829-390). It was about $43 US for 25 feet. A 1 cm piece is cleaned purged of all volatiles in one cleaning cycle on my Markes TD-20 cleaner. I just used some nickel wire I had in the lab.

[James_Ball]

I have made the solutions in ug/L concentrations as the EU limit for metaldehyde in water is currently set at 0.1 ug/L. I can detect all the concentrations previously mentioned, but the calibration curve I make from these points isn't a particularly straight line!

I find it hard to believe that you get a signal for a 1µL injection of a 0.01 µg/L = 10 ng/L standard with a GC-MS single quad (= 10 fg on column)

Remember that the EU limit you mention is on the sample, this does not mean your calibration curve has to go that low if your sample prep concentrates that sample.

For the UCMR4 project I am injecting 0.003 ng/ul using a 1ul injection for alpha Hexachlorocyclohexane using SIM on a 5973 and it gives a decent peak.

[Virolarufula]

Hi all,

I developed a method for metaldehyde and other pesticides using LC-MS/MS for surface water runoff. Don't know if will suit you in terms of LOD & LOQ.

http://onlinelibrary.wiley.com/doi/10.1 ... 4/abstract

Many colleagues are reporting some effective loss of metaldehyde during the concentration step. Notice the Henry constant for metaldehyde is relatively high making this pesticide relatively volatile.

I recommend you to try LC-MS/MS and you can easily achieve the DWD compliance value.