Chlorohydrin

[MikeSimone]

I need to develop an assay for chlorohydrin (3-chloropropane-1,2-diol) in biological media. I have GCMS and HPLC-UV at my disposal.

The molecule does not have a good UV absorption, and when I try to analyze by GCMS I see the peak areas for replicate injections from the same vial decreasing dramatically. I thought it might an adsorption effect with the glass autosampler vials, so I repeated the experiment using polypropylene vials. Same thing.

Any thoughts? If I do not solve this I will have an angry toxicologist coming after me.

Thanks!

Mike

[rb6banjo]

We could use some more details. Diols are very difficult because they are very water soluble. What detection limit do you need? A flame detector for GC might be a better route than MS:

http://blog.restek.com/?p=6662

The mass spectrum for your target compound is not very distinct. It could look like a host of other materials in your matrix. The chlorine atom could help you - electron capture detector could be used to exploit this. Certainly, if GC is an option, a polar stationary phase would be recommended.

[MikeSimone]

I am running a DB35 column, peak shape is decent. I am injecting a 30ppm standard in MeOH, split at 20:1. I am running the MS in SIM mode monitoring m/z43, 44, 66, 79 and 81 which are the main ions in the mass spectrum. GC conditions are pretty standard, 50C to 150 @10C/min. There are no other analytes of interest in the matrix. Everything looks fine except that peak area decreases dramatically from around 8E6 area counts to around 1E6 over a period of 4hours. Overlaying chromatograms the loss of response appears almost linear with time, so I don't think it is something simple like a bad syringe or leaking septum. I have reproduced this so far on three separate days, with two different vial types (glass and PP). Each time I start out with strong peak response, around 8E6 counts and it drops to almost nothing after a few hours.

I am using a Restek Sky deactivated 4.0mm liner and a new Siltek injector seal.

I really can't figure out what is going on here....

[rb6banjo]

Interesting. Anything else growing in? I can't think of a reaction where the glycol would react with the methanol. To eliminate the chromatograph and the methanol solvent as sources of the problem, can you make your 30 ppm standard in something else (methylene chloride for instance) and try the same timed experiment?

[Peter Apps]

How many injections do you do in the four hours ? Are they all from the same vial, or from different vials ? If from the same vial, what happens if after the four hour sequence from the one vial you inject from a different vial ?

Do the glass and polypropylene vials have the same septa ?

Peter

[X]

Derivatize this compound with MTBSTFA1 may give you stable response.

[MikeSimone]

Thanks all for the suggestions. To answer some questions: Both the glass and PP vials have PTFE lined septa. I am setting up the sequence to inject 1ul every 75 minutes. 15minutes for the actual run, 60 minutes with a timed event to turn the detector off and cycle back to starting temperature. So overnight I can make 10 runs or so with an hour between each injection. Overlaying the chromatograms you can see a steady decrease in response from a nearly overloaded peak to non-detect the next morning.

I may try an alternate solvent such as MeCl2 which was suggested. I have not yet tried derivitizing.

[Peter Apps]

60 minutes with a timed event to turn the detector off

Turning the MS off between runs is unorthodox to say the least (and difficult to achieve I would have thought). What happens if you leave it running ?

Peter

[MikeSimone]

Someone at Agilent suggested to me that this would be a good way to build a time delay between runs into the sequence, rather than just collecting MS data for the full 75 minute run.

[Peter Apps]

Someone at Agilent suggested to me that this would be a good way to build a time delay between runs into the sequence, rather than just collecting MS data for the full 75 minute run.

To be frank you have had very poor advice. It's a long time since I used Chemstation and I have never used Agilent's latest software but I find it had to believe that you cannot simply programme a post-run delay - at the very worst you could set a very long stabilization time on the GC.

What does the MS actually do when you "turn the detector off" ? - surely it is not venting and pumping down again, which is what turning it off and on involves. If it is it would certainly explain a loss of peak area.

Why are you putting in a delay between runs ? - it would be normal practice to run a sample as soon as the previous one was finished. Do you see the same decrease in peak size if you do not have a delay between runs ?

Peter

[MikeSimone]

Chemstation E02.02 does not have a function for a time delay between injections. In the method editor, there is a table for timed events on the MSD. Selecting "detector off" at 15minutes turns the filaments off and stops sending voltages to the ion lenses, stops the quadrupole scanning etc. It does the same thing as the MSD would do in between runs or when you program a solvent delay into the method. It does not vent the system in between runs.

[Peter Apps]

Chemstation E02.02 does not have a function for a time delay between injections. In the method editor, there is a table for timed events on the MSD. Selecting "detector off" at 15minutes turns the filaments off and stops sending voltages to the ion lenses, stops the quadrupole scanning etc. It does the same thing as the MSD would do in between runs or when you program a solvent delay into the method. It does not vent the system in between runs.

OK, that sounds like usual practice.

Peter

[MikeSimone]

Ok - a couple more data points...

It appears that my stock solutions are not stable over time either. Making fresh and diluting to WL restores peak area. The loss of response occurs rapidly even with no time delay in between injections. It was suggested to me that the analyte may hydrolyze in contact with water and that my methanol probably contains some water, makes sense.

Diluting in MeCl2 as was suggested here still shows degradation, but the effect is not as dramatic. Again, presence of water seems to make sense.

If anyone is interested here is a link to a good paper describing the analysis of this compound in food stuffs:

http://www.efsa.europa.eu/sites/default ... s/779e.pdf

Looks like I will have to derivatize which I wanted to avoid since I will probably see half a dozen or so samples for this analysis this year.

Any better ideas?

[Peter Apps]

If the chlorohydrin is hydrolysing in moist methanol it will surely hydrolyse in the biological matrices of your samples.

How about UV light ?, are you using clear or amber vials ?

Peter