GC-TCD for moisture?

[MSCHemist]

Hi all I could use some advice.

I work for a flavor company that produces savory flavors. We make both encapsulated and simple plated flavor powders.

Loss on drying works very well for encapsulated flavors. I just take a moisture balance to 85 degrees C which is high enough to drive off the water but low enough not to cause sugars to react or decompose.

The second category the simple plated flavors are more troublesome. Upon heating they lose flavor volatiles along with water giving inaccurate inflated readings for moisture. I am looking into Karl Fischer but they contain sugar polymers and a lot of salt which will not solubilize and many contain aldehydes, diketones, mercaptans, siloxanes all which are known KF interferrants.

I am wondering if a HS-GC-TCD system might be the simplest solution. The setup isn't much more than a KF and there are likely no solvents. The trouble I am wondering is if I heat up the sample enough to release all the water and avoid headspace matrix effects will it decompose and produce water and how to deal with the issues.

[Consumer Products Guy]

In the 1970s we occasionally used packed-columns and TCD for water by GC.

However, I would lean towards Karl Fischer for your samples. You don't care if the salt dissolves or not as long as the solvent can extract the water. Major KF instrument suppliers have systems where the sample is weighed into a vial, sealed, and the contents heated and the headspace transferred to the KF cell to determine water content.

[MSCHemist]

I'm not sold on the KF oven. A lot of the interfering flavor compounds are also volatile and heating the sample above 85C or so will cause reactions that can generate water (sugars and amino acids).

[itspip]

CPG is right. As long as the water in the sample dissolves into the KF solvent you should be fine. You may end up cleaning out the vessel more often though, but it should work.

I wonder if you could add a derivitising agent that would react with the water which you could then analyze by GC.

[Peter Apps]

Maybe a loooong shot, but long ago with the help of a few beers I suggested adding calcium carbide and measuring the acetylene to someone. To my surprise he told me later that it worked. I don't have any details I'm afraid.

Peter

[MSCHemist]

I came across a lit reference for it. The matrix was ammonia and they were looking at trace ppm levels. My samples are solid so I need a mid or high boiling anhydrous solvent probably n-methyl pyrollidine or pyridine or something and in the range of 0.5-3%. Looks like they put the carbide and sample in the headspace vial and incubated it at 40 deg C for 2 hours. Naturally you don't want to generate enough acetylene to blow explode the headspace vial.

[MSCHemist]

Actually there is a device that just does the carbide reaction and measures the pressure generated by acetylene and gets the moisture based on the pressure generated. It is designed for soil. Not sure what the precision/accuracy is.

[Peter Apps]

It's got to be worth a shot if you can easily get your hands on some carbide.

Who were the authors on the paper you found ?

I think that soil moisture is about ten times higher than the range you want to measure, but you could get around that by using a bigger sample.

Peter

[MSCHemist]

The latest issue of LCGC was in my mail box this morning and they had an article on water by GC-TCD. They use a 60m Watercol trifluoromethylsulfonate ionic liquid column and newer headspace GC-TCD. They indicate they had trouble with older GC-TCD with sensitivity which caused them to have to lower the split from 100:1 to 5:1 which caused significant peak shape deterioration. Also they were analyzing syrups in the 30-85% range vs my flavor powders in the 0.5-3% range.

Looks like the carbide reaction method is still preferable as I am not looking to buy a new state of the art GC system for moisture. I am thinking more along the lines of an old Tekmar 7000 and 5890/6890.

[MSCHemist]

Here is the main reference I found for the carbide method
-H.S. Knight and F.T. Weiss, Anal. Chem. 34, 749–751 (1962)
http://pubs.acs.org/doi/pdf/10.1021/ac60187a010

Trace Moisture in Ammonia: Gas Chromatography Using Calcium Carbide
-Journal of the IEST 41(1):21-25 · January 1998

[Peter Apps]

There is truly nothing new under the sun - when that 1962 paper was published I had only just started school ! I love the bit about the GCs being built in house - those were the days !

Peter

[osp001]

Perhaps you could go the gravimetric route with a sealed container using phosphorus pentoxide and heat. Convert the P2O5 to phosphoric acid by reaction with water, giving you weight difference. Might react with other compounds, but it could be worth a shot.

[MSCHemist]

Unless they repealed the law of conservation of mass it would need to be unsealed and since there is no gas evolution there shouldn't be any loss of mass.

P2O5(s) + 3 H2O(l) ---> 2 H3PO4

The trouble with any gravometric method based on water reactivity is the atmosphere especially in summer has quite a bit of moisture. The reaction will never end.

I'm leaning back towards KF titration. If I use the ketosolver which replaces most of the methanol with 2-methoxy ethanol to reduce the acetal formation from butter diketones and aldehydes it should be sufficient for my needs. Mercaptans are very strong smelling and most of our flavors shouldn't contain ppt levels enough to significantly affect results.

[osp001]

Unless they repealed the law of conservation of mass it would need to be unsealed and since there is no gas evolution there shouldn't be any loss of mass.

The system would have to be sealed to avoid loss of moisture, not the sample or the desiccant.

Page 3 section (A)(1) of this document describes the FDA technique for residual moisture determination in biologics using phosphorus pentoxide.

[MSCHemist]

I see the surroundings are sealed but the sample is not. The phosphorous pentaoxide is used to desiccate the sample which is on a balance. It also uses a moderate vacuum 1mmHg which would probably also drive off the flavor volatiles.