How to quantify methanol in low ppb-range?

[JelmerD]

I was wondering if someone knows a technique which I can use for quantifying methanol in water in the range of, let's say, 50 ng/L to 50 µg/L.

I tried HS-GC-MS but that doesn't work, because methanol is too small a molecule for the MS. So I experimented with a derivatisation using sodium nitrite and sulfuric acid (methylnitrite is formed), but that didn't solve the problem. I also thought about some sort of LLE followed by derivatisation with BSTFA, but I figured this probably wouldn't work due to the polar character of methanol (it will never migrate to the apolar phase where the derivatisation should take place).

The equipment I've got is:
Agilent 6890N GC + 5975 MSD (capable of HS, TDU and liquid injection), colums: HP-5 (split/splitless inlet) and HP5-MS (TDU/CIS4 inlet)

Agilent 1200 LC + 6410 QQQ (don't know too much about this system)

Shimadzu 2010 GC-TCD with HP-molesieve column installed

Can anyone help me on this?

[chromatographer1]

JelmerD,

In the June 1997 issue of Analytical Chemistry I published a GS-HS chromatogram of methanol at 20 ppb from water using FID. (see the 0.5ng spike of methanol in the 25µL water sample.

The peak is actually a bit larger than expected due to the residual methanol in the lab air. There was an LC running in another lab on the other side of the multistory building where I worked. Thus I had a blank of the equivalent of 10-20 ppb of methanol in the air when I processed my water samples. Still the true LOD is about 5 ppb when using a small 25µL sample of water. (I also saw hydrocarbons and acetonitrile in the air. The building was a block away from Interstate 70 highway.)

In order to improve this detection limit I would use a 10 meter 0.53mm ID 1µm Carbowax 20M capillary followed by a 10 meter 0.25mm ID fused silica tube before the FID using an oven temperature of 40°C.

I would inject directly a 2mL to 5mL sample using a 20mL/min He flow rate if possible (at least 10mL/min). I would also try using 50 or 100µL of water and add potassium sulfate or sodium chloride salt to improve recovery of the methanol.

If you have a Varian 3500 that would also help as that FID is 4-10 times more sensitive than other FID designs.

It is possible to meet your LOD for methanol. Other options are to use multidimensional Gc using a backflush to detector configuration using a packed 1% CW20M on Carbopack B column and a multimicroliter liquid sample injection.

best wishes,

Rodney George
Senior Research and Development Scientist
Gas Separations Research
Supelco
595 North Harrison Road
Bellefonte, PA 16823

814-359-5737 voice
814-359-5459 fax
rodney.george@sial.com

[JI2002]

The only detector I can think of is MS in SIM for methanol in 50 ppt to 50 ppb range. Don't understand what you mean by saying methanol is too small a molecule for MS.

[Peter Apps]

To do methanol itself you have to have a polar column, if you are stuck with the 5% phenyl columns you could try some wet chemistry to make methyl esters of a smallish acid - say butyric or hexanoic. A lot will depnd on what interferences are in the samples.

I am really intrigued by why you need to measure methanol at ppt levels - as Rod says you will never get a clean blank - breath on a piece of glassware and you will get more methanol than this in any water that you put into it.

Peter

[JelmerD]

Thank you all for your response! As I am not the head of laboratory, I will discuss the options you gave me with my labmanager.

@ JI2002:
MS SIM doesn't work because the ions 15 and 31 will not give enough response. Besides that, the other available ion, m/z=29, is an isotope of nitrogen, which is interfering in the start of the chromatogram when you're using HS injection. But isn't it commenly known that MS is not sensitive enough for compounds with molecular masses of <50? I tried to google up some articles focusing on the determination of methanol, and none of them came up with MS as detector, it's all FID.

@ Peter Apps:
We need to measure methanol at ppt-levels because we have a project running about adsorption and oxidation of MTBE (which is present in concentrations up to 50 ppb). We know that MTBE can break down to TBA and methanol during oxidation. Our researcher suspects that TBA sticks to the adsorbent during oxidation, so she wants know if that is really the case by monitoring the formation of methanol.

[chromatographer1]

JelmerD

It certainly looks like you will be using a cold trap to concentrate the methanol to higher concentrations in order to perform your analysis well.

You should look into finding another absorbent to trap the TBA as its response to a FID is much superior to methanol.

Good luck in your search.

Finding a reagent to react with the alcohols and perform the analysis by GC-MS or LC would be a wonderful idea.

best wishes,

Rod

[AICMM]

JelmerD,

Kolb and Ettre cite (Static Headspace-Gas Chromatography) an analysis of 100 ppb methanol from pharmaceutical buffer by using a cryotrap. 2mL of sample, 2 gram of K2C03, 80C for 60 minutes. Cryotrap is 55cm of 0.32 id 5% phenyl methyl silicone 1 micron thick. Does not say how cold they kept the trap though. Colum they used was a 0.32 wax but since you are using an MS you should use a 0.25 wax. You will pay a price for the column diameter switch but it should work okay. Kolb is partial to pressure balanced injector so the application does not cite an injection volume.

Best regards.