gas tight syringe?

[jerry_76]

I need to analyze quantitatively vapors of different organic solvents. I thought (perhaps naiively) the easiest way is to have the solvent in a bottle with GL cap, wait until it reaches equilibrium with its own vapor and then get a sample through the septum in the cap and inject it into GC. Is it that simple, or there's more I should know about before I begin? (I bet there is...)

[chromatographer1]

You are assuming there is no air or nitrogen in the bottle remaining after if was filled with the solvent. You are also assuming that you can precisely hold in the syringe the same number of gas molecules for injection each and every time. And lastly you are assuming that the ideal gas law applies ideally for each and every solvent you will test. Now the reason they call it ideal is because it is not of the real world.

Three assumptions........... would you bet your life or your research on them all being 100% true?

Would you?

By the way, I have a bridge for sale in NYC spanning the East River for sale at a very reasonable price. Would you care to consider purchasing it? I also have some (swamp) land in Florida available for purchase at a very reasonable price as well.

best wishes,

Rod

[Johnny Rod]

Whata re you trying to achieve with the analysis?

PS is it a nice bridge?

[jerry_76]

I'm not really casino-like person, so no betting this time, sorry. What I actually do is that I pass the air (or N2) stream over the surface of solvent contained in the flask (gas-tight, or as gas-tight as it can be in a non-ideal world) and I have to figure out the concentration of vapor in this air stream. So being a simpleton me I though to get a sample of vapor with a syringe while I'm passing the air stream over the liquid and inject it into GC.
Having said all that, I still can't get the concentration of the vapor in the air stream quantitatively

Do you also own Ambassador Bridge in Michigan? Matty, is this you?

[Peter Apps]

You CAN do it, but with a manual, room temperature type of set up the accuracy and precision might not be very good. Echoing Johnny Rod - what are you trying to find out ?

I wouldn't take the bridge offer too seriously - I am already in negotiations with a Nigerian politician's widow to have the purchase price transferred to my bank account.

Peter

[Don_Hilton]

Depending on the configuration of your flask and plumbing, mixing (or lack thereof) could be a problem. The gas contacting the surface of the liquid will have higher conentration of gaseous components. Diffusion is gas is fairly fast, but only on a relative scale. So, if you are trying to sample from a moving gas stream in this flask, you could have a problem of reproducability - just because of the location of the needle relative to the conentration gradients in the gas stream.

If you are trying to get the average gas concentration after passing nitrogen over the liquid surface in this flask, I would suggest passing the gas into a gas sampling bag and sampling from the bag. This will allow you to resample the same sample multiple times and, if needed, ensure the sample you are working with has been properly mixed.

If you are trying to use the gas tight flask at the end of the stream to catch the gas, displacment of the gas already in the flask along with mixing would be a problem unles you have evacuated the flask to start with. And with that you have to be sure of how thoroughly you evacuate the flask each time and the effect of decreased pressure in your experiment as the vacuum draws gase into the flask. Thus the suggestion of a gas sampling bag. The gas sampling bag starts with zero internal volume (or pretty close) and the gas is always at atmospheric pressure (ore pretty close), no matter the volume.
---

Added in edit: And sampling from the gas sampling bag is with a gas tight syringe - and the technique works well. They even make gas sampllign bags with septum ports on them for this purpose.

[jerry_76]

Don, your answer is very helpful - I actually didn't even know those bags exist. I agree, this option is best for throughout mixing.

My goal is really simple - to get the concentration of the evaporated liquid in the gas stream, preferably using GC. To achieve this I need some reference, hence the idea of analyzing of what comes off "naturally" from the vapor in the air-tight flask...

So bridge is taken? First this analysis problem and then this. What a day.

[Peter Apps]

Depending on how volatile the liquid is the simplest way might be to weigh the solvent before and after, and multiply gas flow rate by time to get volume. You then have mass of solvent per unit volume of gas, and if you want mole fraction the rest is just arithmetic.

Peter

[Don_Hilton]

You can purchase a large gas tight syringe - about 1L, fill it with air, put that in a bag and inject a known volume of solvent. Mix the bag well and take out a mL or so to inject in the GC. Add more solvent - mix and sample again. And with a few additions, you have built a calibration curve. For most of us the loss of 5 mL volume from 1L in such a calibration would be acceptable.

[Johnny Rod]

You could just take a tee off the exit gas stream and send this through a gas sampling valve to inject into the GC. If your pressures are low you might do this with a small bellows on the outlet of the sample valve, and a flowmeter (rotameter) Say squeeze the bellows once, wait for the ball(s) to drop back to zero, press start at that instant, should be fairly repeatable. In terms of calibration, you can buy a cylinder of your solvent in nitrogen (or whatever gas you're using), but there is a trade-of with its volatility. Less volatiles solvents like n-butanol mean you either have a cylinder at lower concentration or you get a higher conc but at a lower pressure so less gas in it. If you inject the same way for both e.g. at zero flow, you'll get something sensible out of it.

A good starting point is to look up some vapour pressures and just take a stab at the ballpark concentration and go from there, ifit's a mixture then just proporotion each component, once you have the right order of magnitude then you can start to think about making accurate measurements.

[jerry_76]

Johny, I actually tried that and it worked (with the flowmeter) - the calibration is the main problem. I'm not sure if I can get all the mixtures I need from the gas supplier. That is why I thought about taking out a defined volume for analysis. Then again I need to know the concentration somehow.
Would it be possible to calibrate the detector with solutions and use it as a reference for vapor analysis? Supose I make the calibration for different solutions with varying concentration of my vapor sources - would it be possible to refer it to the gas analysis of pure vapors? (I use FID detector for that).

[Don_Hilton]

I have added small portions of solvents to gas in a gas sampling bag for the purpose of calibrating for measurement of a gas mixture. YOu do have to makd sure the lisuid vaporizes and there is always the question of adsorbtion onto the bag - so you need to take a bit of care with higher boilers to be sure you are gettign what you think you are getting.

[Johnny Rod]

Yep you can make standards in a bag or we prefer a glass flask which has a septum for syringe withdrawal. Which solvents are you looking at? I'm not sure about the US (I think are are over the pond from me?) but in the UK I work for BOC (Linde group) and our Special Gases division can make a mixture of pretty much anything.

Failingthat then yes you can inject solutions but you need to calculate the amount on column to be usre you're in the right ballpark. Don't get your w/v and v/v units confused, the old calssic!

[chromatographer1]

Your first post was:

"I need to analyze quantitatively vapors of different organic solvents."

The problem is and was getting an accurate determination of the mole volume of the sample and then accurately taking a precise portion for injection, and then injecting ALL of the measured portion into the GC.

One approximation I used to perform was to take a precise volume of a pure solvent liquid and blow it into an empty (0 PSIA) gas cylinder by using helium and pressurizing the cylinder to a precise pressure. Knowing the capacity of the cylinder it was possible to determine the mole % (v/v) of the gas contained in the cylinder and use a sample loop valve for consistent injection into a GC. Sometimes a little warming of the cylinder was performed to ensure the liquid was vaporized.

This procedure has its limits but was quite useful in making 1-100 ppm solvent/helium samples for repeated analysis.

best wishes,

Rod

[jerry_76]

Thank you all for help. I'm actually trying now to get a reference mixture from gas suppliers (Linde included). This would be definitely best solution to this problem.
Rod, I agree this is possible to make such gas mixtures in the lab, except that I don't have experience with it, and I don't think the H&S responsible would let me play with pressurized gases... So I think evaporating a known volume in a gas-tight bottle or bag is more probable option.

Failingthat then yes you can inject solutions but you need to calculate the amount on column to be usre you're in the right ballpark.

Could you possibly elaborate on that?