static headspace method

[seet88]

Hi all,

I would like to ask the valid method for headspace. Headspace method is the solid/liquid when heated the gas and solid/liquid in equilibrium state. For water is very direct mean i prepare std in 1ml, sample also in 1ml. So calculation is straight forward.

However, let said for ethanol in viscous source or solid phase, it require to weight by gram. Example i prepare 100ppm ethanol from 10000ppm ethanol ( mean the absolute weight is 100ug - 10ul of the 10000ppm). Absolute weight no need consider the volume top up. So the quantitative is the absolute weight straight devide by the weight of sample. My question is do i still need top up water same as the standard? Another question is i compare one calibration vs sample with std addition (adding std on top of sample and do as a full cal curve). The result will vary. I believe std addition full cal curve the last point wil be bias low, so the result will cause it bias high. Agree?

[Peter Apps]

In equilibrium headspace the matrix of the samples and the standards must be the same. So if your samples are in water, then your standards must be in water.

As long as the matrices are the same there is no reason why a standard addition curve will be biased relative to an external calibration.

Peter

[seet88]

Hi Peter,

If let said i would like to test ethanol content in biscuit. So is it mean i ned weight 1g sample plus 1ml water in headspace vial; at the same time std will b in 1 ml water? But will it be in equilibrium state since both gaps of headspace will be vary because one of it contains additional 1g sample (bcos ethanol cannot be in gram, if in gram the concentration will be in very high). For the standard additional case, due to u keep on adding std on top of sample it will diluted ur sample as well. Hence the result will be bias high for the case. Mean i add 1g of sample to different vial and added 10000ppm ethanol base on the absolute weight, eg
1. 5ul 10000ppm ethanol added into sample
2. 10ul 10000ppm ethanol added into sample
3. 25ul 10000ppm ethanol added into sample
4. 50ul 10000ppm ethanol added into sample

Thanks
Seet

[MSCHemist]

I use full evaporative (FET) HS/GC/MS SIM
I use acetonitrile as an ITSD
I dilute the sample in water and spike with acetonitrile ITSD.

I then put 20ul in a 20ml vial (actually 7.5ul in a 9.2ml vial Tekmar 7000)
heat to 120 deg C
and inject 2ml 5:1 split
1.3ml/min He in a 30m 0.32mm 0.5um df column.

I calibrate at 10ppm 50ppm 250ppm and 1000 ppm ethanol.

With an FID you probably will have a LOQ at the 50ppm level.

[Peter Apps]

Hi Peter,

If let said i would like to test ethanol content in biscuit. So is it mean i ned weight 1g sample plus 1ml water in headspace vial; at the same time std will b in 1 ml water? But will it be in equilibrium state since both gaps of headspace will be vary because one of it contains additional 1g sample (bcos ethanol cannot be in gram, if in gram the concentration will be in very high). For the standard additional case, due to u keep on adding std on top of sample it will diluted ur sample as well. Hence the result will be bias high for the case. Mean i add 1g of sample to different vial and added 10000ppm ethanol base on the absolute weight, eg
1. 5ul 10000ppm ethanol added into sample
2. 10ul 10000ppm ethanol added into sample
3. 25ul 10000ppm ethanol added into sample
4. 50ul 10000ppm ethanol added into sample

Thanks
Seet

You will not get accurate results if the samples have biscuit in them and the standards do not. Samples and standards must have the same matrix. This is a more serious problem than the difference in headspace volume.

When you are doing standard additions you must either add very small volumes of concentrated standard, or you must take the volume into account in your calculations, or you must add clean matrix to each level to make the total volume the same. So in the example you give; sample plus 50 ul extra water, sample plus 5 ul 1% etoh plus 45 ul water, sample plus 10 ul 1% etoh plus 40 ul water, sample plus 25 ul 1% etoh plus 25 ul water, sample plus 50 ul 1% etoh.

Peter

[rb6banjo]

I'm with Peter on this one. You also have to make sure that when you make your additions, that you "get back what you put in". First, you need to verify that that the EtOH you add, partitions into the matrix in a known and repeatable way and that you know how long that takes. Adding the same amount of standard to the matrix and analyzing it with your method as a function of time will tell you how long you must equilibrate the sample. Once you know that, then you can add at multiple levels to check your linearity (do I get back what I put in?).

I do this sort of thing with packaging materials. A biscuit sample will probably behave very well (I don't know for certain but for ethanol, it should be good) as a partition system but some polymers are not good at all (e.g., polystyrene). If this is the first time you've ever heard of "the method of standard addition", please read about it in any undergraduate analytical chemistry text. As MSChemist mentioned, the LOD/LOQ for EtOH might not be too great if you add too much water.

Another approach is to completely dissolve the polymer (in this case a biscuit) and change the matrix essentially to the solvent. Your detection limit generally goes way up (poorer) if you go this way - especially for something like EtOH in this matrix. If you were looking for residual styrene in the biscuit, you might be ok with total dissolution of the matrix. Analysis of solids can be tricky!

[seet88]

Hi Peter,

Ok i think i get your point. Mean for the headspace method we can not centralized the method for ethanol by using external cal curve vs sample due to the sample matrix effect ( solid, liquid, viscous). So have any idea to have a general method to done on it? Example sample varities are biscuit, light sauce, dark sauce - viscous type, marinated sauce.

Thanks
Seet

[rb6banjo]

All of those matrices will be likely be different. Perhaps the sauces might be the same or similar. A chunky marinade won't likely be the same as a thin soy sauce (something like that). Certainly not the same as a biscuit.

The only way that I know of where you can get a "generalized" method for things like this involve elimination of the matrix. Generally, that means some type of solvent extraction. You'll have a tough time with EtOH an any solvent-extraction procedure that I can imagine.

[seet88]

Hi rb6banjo,

Then u receive different matrix test, how u do it? Can u draft the idea out? Appreciate ur info

Thanks
Seet

[Peter Apps]

Hi Peter,

Ok i think i get your point. Mean for the headspace method we can not centralized the method for ethanol by using external cal curve vs sample due to the sample matrix effect ( solid, liquid, viscous). So have any idea to have a general method to done on it? Example sample varities are biscuit, light sauce, dark sauce - viscous type, marinated sauce.

Thanks
Seet

That is exactly what I mean.

You might be able to get close to a general method if you work with saturated salt solution as your matrix for standards and to dissolve/dilute your samples. The salt will even out differences due to having dissolved solids. A lot will depend on what detection limits, precision and accuracy you need, and you should validate the method in house.

A problem that you need to address is whether you have any other compounds in the samples whose peaks co-elute with ethanol.

Peter

[rb6banjo]

I'm sorry. That would be too involved for me to type it all out. It sounds as if you're pretty new to all of this. I suggest that you get a good book on headspace gas chromatography and start reading. I can recommend:

"Static Headspace Gas Chromatography" by Kolb and Ettre, Wiley-VCH, 1997.

I think there's a more recent version than 1997. That's just the one I have on my desk. There are many good examples of how to develop headspace gas chromatographic methods in that book.

Here's the more recent version. Only $100 US for the Kindle version. It is a great resource:

https://www.amazon.com/Static-Headspace ... +headspace

I'm not sure if you can access ebay where you are but there's a copy of the 1997 version on there right now for $27.50 US. Having the most recent version does not matter that much.

[James_Ball]

This sounds like an analysis better suited to purge and trap than static headspace. With a heated purge you can obtain results that would be more consistent across the various matrix and only need water as the matrix of the standard, or possibly water/sand or salt water. With 10 minutes of purge time at 40-60C you would extract a major portion of the ethanol from the samples leaving less variation in results.