Determination of Conc. of the Gas Inj. through Liq Inj.

[tmalew2]

We have a the following specs:
Agilent 6890 GC
2 FID's both have DB17ms column
One FID is connected to Liquid Injection Port and
the other FID is connected to gas injection.

Is there a way of finding out the ppm levels of a substance in a gas by doing calibrations through liquid injection. Lets say I inject n-decane in the liquid injection port, 10ug/ml. I end up getting a peak of certain size. Now I inject a argon gas that has traces of n-decane. When I inject the gas, I know the pressure of injection, volume of sample loop, and maybe the temperature.

Is there a way of calculating the concentration in ppm of n-decane in the gas injection by using the liquid injection area and concentration (micro grams/ml)?

[thohry]

We have a the following specs:
Agilent 6890 GC
2 FID's both have DB17ms column
One FID is connected to Liquid Injection Port and
the other FID is connected to gas injection.

Is there a way of finding out the ppm levels of a substance in a gas by doing calibrations through liquid injection. Lets say I inject n-decane in the liquid injection port, 10ug/ml. I end up getting a peak of certain size. Now I inject a argon gas that has traces of n-decane. When I inject the gas, I know the pressure of injection, volume of sample loop, and maybe the temperature.

Is there a way of calculating the concentration in ppm of n-decane in the gas injection by using the liquid injection area and concentration (micro grams/ml)?

Yes if
* you know exactly how much sample enters the columns after injection
* The two FID must be in he same condition, oven temp and flow rates are also the same
* Probably the tolerance is higher.

[tmalew2]

thohry,

Both columns are in the same oven so their temperature is the same throughout, the flow rates are the same as well.

As for the volume of injection, the gas injection we know since its a 100 microliter sample loop, the liq. injection we use a splitless method, and we always inject 1 microliter through a manual syringe.

Would this be enough or more is needed?

[thohry]

thohry,

Both columns are in the same oven so their temperature is the same throughout, the flow rates are the same as well.

As for the volume of injection, the gas injection we know since its a 100 microliter sample loop, the liq. injection we use a splitless method, and we always inject 1 microliter through a manual syringe.

Would this be enough or more is needed?

So I think the only thing you have to carry out is to correlate the concentration of a peak in 1ul (liquid) and 100ul vapor. Then try to do some spikes.

[GasMan]

I would do this as follows:

1. First check how good your manual injections are. You will need to fill your syringe to 1 ul, withdraw all of the sample into the syringe barrel and note the volume. Inject the sample and then see how much is left in the syringe. Calculate how much you have injected and then calculate the weight of decane that has been injected. Repeat this several times and record the area of the peak.

2. You should NOT assume that both FID's will give exactley the same results. I would move the end of the column that you have just used, and install it into the the other FID, and see if your results are the same.

3. IF both FID's give similar results, then you can use the response factor that you get on the FID for the manual injection on the valve injection FID.

4. This will give you then the weight of the compound in your 100ul of gas injected, and from this you can calculate the concentration.

Gasman

[JI2002]

The right way doing it is to run a new calibration with n-decane gas standard and analyze the samples the same way as the standards.

[tmalew2]

I just want to clarify this, when I inject 100ul of gas, the gas is mostly argon, with n-decane and other species present if ppm levels. The liquid injection we inject is 1 ug/ml, so its 1 ug of n-decane in 1 ml of dichloromethane, but we inject 1ul of this solution into the GC. We don't inject any pure liquids or gas into any column. Gas is always mostly pure Argon, and liquid is mostly dichloromethane.

Does this change how I would calculate the ppm in the gas.

Let's just say for now that the response is the same on both FIDs and columns. I can find the mass of the substance I inject into the liquid injection and get an area. When I inject the gas, I need to divide the response (area) by the pressure of injection. Therefore, we always get Area/(Pressure Inj) versus the concentration. So how would I be able to find the mass of the substance I inject based on the information I have. I have the volume of the gas injection, but I don't know how much in that volume was n-decane.

Any thoughts?

[larkl]

I think you're getting into dangerous territory here. You may be able to get a number, then you report it and it becomes the 'gospel'. Then you'll be defending it at a later date and youu're not sure it's right without a gas standard. You can make a gas standard in air in a glass bulb of known volume. Just inject 0.5 uL of decane in the bulb, warm it slightly to vaporize the hexane and do the math. You could even fill the bulb with some other gas (argon?) first, as long as you ensure it's a atmospheric pressure.

[tmalew2]

Thanks for everybody's help, but the reason I ask its n-decane isn't really the problem doing calibrations in gas phase. It's the really heavy species that we can't do gas calibrations with cause of absorption issues to the surface of the tubing. n-Decane was just an example of a species.

[AICMM]

tmalew2,

I was going to suggest a static dilution bulb except your last post suggests that the compounds you are really interested in will go splat on the walls making this option useless. Correct?

Best regards.

[tmalew2]

Correct, right now in our lab we are measuring heavy hydrocarbons, biphenyl, anthracene, naphthalene, and heavier. To prevent absorption, we are using tubing where the surface was deactivated or coated with a solution to prevent absorption, but nevertheless, when running calibrations glass bulbs don't work, we noticed absorption problems with heavy species. This is why I am trying to figure out calibrations technique through the use of a liquid injection port with the same column type, both DB-17ms in the same oven.

[AICMM]

tmalew2,

I would propose taking you in a completely different direction based on your last post. I would propose collecting your sample on a cartridge and eluting the cartridge with a solvent. Then you have a solvent to solvent calibration. The only one that would be even the slightest concern would be Nap since it might break through depending on the sorbent. I am confident you could find an appropriate sorbent that would work for these components.

Something to consider.

Best regards.

[tmalew2]

In our lab. we tried that already, we collected a sample from the experiment in a vessel, flushed it with dimethylchloride and injected the liquid. We had good and bad results.

I just really need to know if its possible to do calibrations of the gas through liquid. What do you need to know to be able to or is it not possible or accurate enough to use.