May direct injection gives different results than HeadSpace?

[LucasOrtiz2]

Hi, My name is Lucas and I work for a laboratory in a University on Chile.
I am new working with GC-FID, and I have an abnormal issue.
My work consist on measure different samples of MeOH+EtOh+NPropanol in brine (salinity 150000ppm). The problem is that both EtOH and Npropanol gives higher values of concentration using HeadSpace (with capillary column - Agilent) than Direct injection (packed column - Shimatzu). The opposite case occurs with MeOH, concentration are higher (or same value) on Direct Injection than HeadSpace.
I cant explain why this is happening. I need a little bit of help please
Every time we start a new batch of samples, we make a new calibration curve (new matrix) for the HeadSpace.
GC-FID with HeadSpace is "mine" so i can tell you whatever you need about it. But the Shimatzu belongs to another lab far from here.
Thats all. Thanks!

[Rndirk]

If i'm correct you and another lab are getting different values for the same samples right?

Getting the same results on the same samples with a different instrument, a different injection, and a different person performing the analysis can be troublesome. What errors are we talking about?

[LucasOrtiz2]

Thanks for reply. All of you are saying is true but the error is two folds, i mean if with HS injection i get 400ppm then for Direct Injection the value is 800ppm. So that is why i think the problem should be something else and not a common mistake.

[LucasOrtiz2]

If i'm correct you and another lab are getting different values for the same samples right?

Getting the same results on the same samples with a different instrument, a different injection, and a different person performing the analysis can be troublesome. What errors are we talking about?

Thanks for reply. All of you are saying is true but the error is two folds, i mean if with HS injection i get 400ppm then for Direct Injection the value is 800ppm. So that is why i think the problem should be something else and not a common mistake.

[Peter Apps]

If i'm correct you and another lab are getting different values for the same samples right?

Getting the same results on the same samples with a different instrument, a different injection, and a different person performing the analysis can be troublesome. What errors are we talking about?

Thanks for reply. All of you are saying is true but the error is two folds, i mean if with HS injection i get 400ppm then for Direct Injection the value is 800ppm. So that is why i think the problem should be something else and not a common mistake.

Hello Lucas

I we are to help you we need an accurate description of what the problem is; in your first post you say "The problem is that both EtOH and Npropanol gives higher values of concentration using HeadSpace" which is exactly opposite to your most recent post.

We need to know what intruments you are using, your analytical conditions and how you are doing your calibrations, and the same for the other lab.

Have either of the labs ever analysed a refererence material with a known concentration ? What were the results ?

Peter

[LucasOrtiz2]

If i'm correct you and another lab are getting different values for the same samples right?

Getting the same results on the same samples with a different instrument, a different injection, and a different person performing the analysis can be troublesome. What errors are we talking about?

Thanks for reply. All of you are saying is true but the error is two folds, i mean if with HS injection i get 400ppm then for Direct Injection the value is 800ppm. So that is why i think the problem should be something else and not a common mistake.

Hello Lucas

I we are to help you we need an accurate description of what the problem is; in your first post you say "The problem is that both EtOH and Npropanol gives higher values of concentration using HeadSpace" which is exactly opposite to your most recent post.

We need to know what intruments you are using, your analytical conditions and how you are doing your calibrations, and the same for the other lab.

Have either of the labs ever analysed a refererence material with a known concentration ? What were the results ?

Peter

Hi Peter. Thanks for your reply. I agree with you, i missed up. Let me be more specific. My lab recive a group of samples witch contain MeOH, EtOH, and Npropanol in brine as solvent. Since we start our values of concentration (for EtOH and Npropanol) are twofold than the other lab (They are a private lab so they dont want to share anything). The opposite happens to the MeOH, sometimes values are same but othertimes our values are lesser.
Both Labs must report results to the same company.
We use a Agilent GC-FID with HeadSpace Injector. Every year we do the offical service and maintenance for all the equipment.
For calibrations the company provides the brine, so we just prepare differents concentrations (100, 200, 400, 600, 800, 1000ppm for example) for triplicate each. Then we make the numerical regretion and start to measure. We do all at the same day. Also we prepare checkpoint samples (put it random on the sequence) to be sure that all is running well.
On the other hand the samples are not coming direct to us. First the company send the samples to the private lab and then they send it to us. The samples comes in a plastic 20ml bottle. Then we take an aliquot of 2ml and put it in a HS-glass vial sealed with a metalic cap that contain a rubber septum.
We asked they to measure a reference material but they answer that is not in their contract so they wont do it. All I know about their equipment is what I posted first.
Think that is all. Thank you for your support.

[GOM]

Hi Lucas

In addition to the above replies

You asked

My work consist on measure different samples of MeOH+EtOh+NPropanol in brine (salinity 150000ppm). The problem is that both EtOH and Npropanol gives higher values of concentration using HeadSpace (with capillary column - Agilent) than Direct injection (packed column - Shimatzu).

If you stop to think about it and consider why then yes, the two approaches (direct injection vs headspace) will give a different result. Why do you think that may be so?

With headspace analysis the usual approach is single or multiple standard addition of your analyte to your sample

Kind regards

Ralph

[Peter Apps]

Hello Lucas

Unless you can determine which lab's results are correct (assuming that one of them is getting correct results) there is not much prospect of any progress with this.

We could try to troubleshoot your method, if you provide complete details of every step, but that seems pointless unless you know that your results are wrong.

Do you know that your results are wrong ?. Does the company that produces the brine say that your results are wrong ? Why are they sending samples to two labs ?

Peter

[James_Ball]

Methanol will be more soluble in the solution than Ethanol or Propanol, so the headspace values will most likely recover better for the Ethanol and Propanol than for Methanol, which I can see giving you lower results for Methanol by headspace. Of course if you calibrate in the brine solution, then that should compensate for that difference on solubility.

What bothers me is the private lab does not want to analyze a known reference standard, seems maybe they are not sure their values are correct. If it was my lab I would analyze it even if not in the contract just to prove I am doing things correctly. Maybe you can take one of your standards, send it to the company wanting the samples analyzed and have them submit it to the private lab as a blind sample. I would also ask the company to send you a split of the original samples so you are not needing to analyze a sample that has already been opened in the other lab. Since the target analytes are volatile, they could be lost when being handled by the other lab, and if they have methanol in the lab there, you could be getting contamination from their lab air into the sample before it is sent to you.

It is difficult to show two different methods are comparable if you are not using identical calibration standards or sample handling protocols. The best you can do it carefully document your method and process and be able to show there are no errors in your measurements, then it will fall on the other lab to do the same to back up their results. Measuring splits of a reference sample is the only way to compare the two processes on equal ground.

[ronery]

Headspace and direct injection are not interchangeable, in other word, you need to re-validate your analytical method if you want to switch from direct injection to headspace.

You can try standard addition method on headspace then compare results. Headspace can analyze more diluted samples than headspace, but it is also prone to more significant matrix effect and partition coefficient.