Headspace Problem

[ZlausD]

Hey Guys
I´sorry for spamming for help in my first thread but i have a problem on my headspace analysis.

I use an agilent G1888 Headspace with Agilent 6890 GC
I use a stock solution of:

Methanole: 213µg/ml
Ethanole:1000µg/ml
Isopropanole: 1000µg/ml

I dilute 400µl Stock solution with 4600µl Water for chromatography to get my calibration standard

The calibration standard is a mixture of
18µg/ml Methanole
80µg/ml Ethanole
80µg/ml Isopropanole

The Headspace Programm is:
Oven Temp: 90°C
Loop Temp: 120°C
Transferline Temp: 100°C
Cycle Time 23min
Vial Equilib Time: 10min
Temp Equilib Time: 1min
Pressurisation Time:0,2min
PressEq Time: 0,3min
Loop fill time Time: 0,15min
Loop Equilib Time: 0,5min
Inject Time: 1min
Valve Purge Time: 10min
Vial Size: 20ml
Shaking: High

GC Parameters:
Inlet:
Split Ratio: 7
Inlet Temp: 180°C

Column:
I use a Wcot fused silica column from Varian:
lenght: 30m ; Filmthickness: 1,8 µm ; Diameter: 320µm

FID Detector:
Temp: 260°C
Makeup flow (N2): 45ml/min
H2 Flow: 40ml/min
Air Flow: 450ml/min

Oven:
40°C for 8 minutes then 60°C/min to 200°C hold for 2 minutes

I get areas in my standard solution chromatogramm:
Methanole:0,9pa/min
ethanole: 7pa/min
isopropanole: 12 pa/min

after that i inject a vial filled with 5ml Water for Chromatography:

There i find:
methanole0,009 pa/min
ethanole: 0,06pa/min
isopropanole: 0,05pa/min

this is 1% carryover...
i read in your forum that your Systems have a carryover of 0,003% how can i get to that?

should i reduce the split or something?

[Peter Apps]

Welcome.

The carryover is most likely due to your analytes sticking to the internal surfaces of the plumbing in the headspacer, or in the GC inlet.

Do you have deactivated plumbing ? (probably called Siltek) or is it bare metal ?

How have you connected the transfer line of the headspacer to the GC ?, is it coupled into the carrier gas feed line close to the inlet, or do you have a needle going through the septum ?

Has the headspacer been used for samples that might have high boiling components that will accumulate in the tubing ?

Do you have any kind of packing in the inlet liner ?, and what kind of liner is it ?

What is the stationary phase in the column ?

Peter

[Bigbear]

What is your loop size. You only need 5-10 loop volumes durring injection. You may try reducing the injection time to give more time flushing the loop between runs.
You may try increasing the loop fill time to insure that it's full of the intended sample.

[ZlausD]

Wow guys thanks for your quick reply

I use a Siltek deactivated transfer line
But the 6 port valve istnt Siltek its bare metal I think (agilent part nr.: 9040700026 )
The Sample Probe is also deactivated (agilent part.no.: 2322700011)
Also the transferline needle is deactivated (agilent part.no: 2322590004)
The transfer line is stick through a septum into a Liner
It’s a glass liner with 0,8mm diameter there is no packing material in the liner

As a column I use a CP select from Varian with
6% cyanopropylphenyl, 94% dimethylsiloxane as stationary phase
I use the Headspace to analyse samples for two methods one method is for samples up to 100 µg/ml
The other method is for higher concentrated samples up to 5 mg/ml.
The samples have a water based matrix. Mostly I test immuneglobuline samples.


I try to reduce reducing injection time and increasing the loop fill time today.... but i dont know witch loop size i have.... i think its 2 ml

thanks

[Peter Apps]

There is nothing obvious wrong with the set up or the method. Try increasing the valve purge time to the maximum that you can use with the cycle time that you have, and increase the transfer line temperature to 120C. There might be a bit of a cold spot where the needle joins the tip of the transfer line, try wrapping aluminium foil around the connection and the injector nut to transfer the heat, and put some insulation around the whole thing.

Peter

[ZlausD]

Thanks Peter

Ill try to increase the valve purge time to 20 minutes and heat up the transferline to 120°C
its a good idea to isolate the injector connection ill try this too.

[ZlausD]

So I increased the maximum purge time to 20 minutes and isolated the point where transfer line fits into the GC inlet.
I heated up the transfer line to 120°C.

Also i did the changes Bigbear requested to do.

Now i have a Carryover of 0,5% so i got better with your help...Thanks guys

And the fact that my areas are really low is no problem?

can i still improve the carryover?

[Peter Apps]

Your samples are close to the lower limit for headspace analysis, so the peaks can be expected to be small. How repeatable are they ?

To check that you have genuine carryover (as opposed to long-term bleed from somewhere) do one injection from a concentrated standard, and then series of 5 or 6 from clean water - if it is real carryover you will see an exponential decline in peak area from the blanks.

It might be worth trying a different inlet liner with a wider bore - you can get strange flows as the pressure hops up and down as the loop switches into the flow path and with the narrow liner you have you might be getting sample into places in the inlet from where it re-appears in subsequent runs.

Peter

[amazhad]

You might be able to optimize your split ratio for larger area counts. Are you getting a true 7:1 split ratio? See below before answering.

Since the transferline is attached to the inlet via a needle through the septum, I'm assuming you have a manual pneumatic control (MPC) headspace requiring you to adjust the flow rate of gas through the transferline via the carrier flow control on the top of the G1888A headspace.

This type of setup provides two sources of gas flow into the inlet, one from the headspace (MPC) and one from the inlet electronic pneumatic control (EPC) module. However in this type of G1888A setup the 6890 does not know about the MPC gas source coming from the headspace transfer line, it only knows about the inlet EPC gas source. This changes your split ratio, what you see on the front panel of the GC is wrong, it is not 7:1 but something larger.

For example:
You remove the transferline from the inlet, so only the inlet EPC is providing gas flow. The column flow is set to 2 mL/min and the split ration to 7:1. The flow coming out of your split vent is 14 mL/min. A true split ratio of 7:1 just like what it says on the front panel of the GC. A gas flow meter on the actual split vent would confirm 14 mL/min.

You reconnect the transferline to the top of the inlet and adjust the carrier flow on the headspace so 20 mL/min of flow is coming through the transferline. Your inlet parameters are still the same, and the front GC panel still shows a 7:1 split ratio with 14 mL/min to the split vent. However if you put a gas flow meter on the actual split vent you should get a reading close to 34 mL/min (inlet EPC module split flow + gas flow through the MPC headspace transferline). The true split ratio is 17:1, not 7:1.

The only way to calculate your true split ratio is by using a gas flow meter when a headspace is using MPC. So are you getting a true 7:1 split ratio or something higher?

Now that I have said all of that, the G1888A can also be setup for EPC control where the inlet EPC is re-routed to the headspace and the EPC actually controls the gas flow through the transferline. There is only one gas source to the inlet. With this type of GC/headspace setup, the split ratio on the front panel of the GC is always the true split ratio. Unfortunately in my experience when GC headspace methods are traded between sites, the method often does not stipulate what type of headspace gas flow control was used (MPC vs EPC), leading either to too much sample on the column or too little.

[ZlausD]

Thanks Peter and thanks amazhad

I test 6 samples of the calibration standard in one sequence and there is a variance of 2-3%
Ill try a bigger liner on monday and also the point with the concentrated standard.

Also i found out that every time i test clean water some solvent peaks appear.

So i´ll try the steam clean programm too

The G1888 is controlled by EPC like you mentioned... i changed this when i started this project.
When started to work on this method the carrier flow was controlled by MPC and wrong..

So my Split ratio is the true split ratio.

[ZlausD]

Hi guys a little update...

I ran a the agilent steam clean procedure...

Still i have solvent (ethanole methanole ) peaks in my Clean Water samples...

So is the column the problem? Should i change it?
Also Ill try the broader liner today.

[Peter Apps]

The column is very unlikely to be the source of carrryover.

There might be multiple sources for the carryover, in which case you might not eliminate the carryover in one step. Did the steam clean make any difference at all ? Did you run a standard after the steam clean, or did you run clean water immediately ?

Peter

[ZlausD]

No the steam clean didnt make any difference...

I ran only two air and a clean water sample.
No standard

In every sample there was a 0,08pA Ethanol Peak.

Greetings

[Bigbear]

When was the last time you did injector maintenence? Your caryover may come from the split vent line, you can change it if you have GC grade 1/8" copper.
You can test this theory by injecting 1ul of MeOH directly into the injection port.

[ZlausD]

I had injector maintenance in march 2013

Should i change the line inside the headspace or the line wich comes out of the headspace and goes into nowhere?

What am i supposed to see wehn i inject MeOh in the injection port just one MeOh Peak?

But ill try this afterwards whenn the testrun with the bigger liner is finished.

thanks