poor reproducibility of RRF and peak areas

[Plunkett79]

Dear members,
I am new to the forum.
I have a problem of poor reproducibility in GC peak areas form injection to injection, and I’d like to understand the possible causes.
I’m currently doing manual injections on a Clarus 580 GC with FID detector.
My GC method is: Isothermal 110°C for 10 min, Injector temp = 220C, Carrier gas: 2ml/min (N2), Split ratio: 10, FID temp: 220C, Air flow 450 ml/min, H2 45 ml/min, Column DB WAX 30m x 0.25 mm x 0.15um
The sample if Diethylhydroxylamine (dissolved in water). Injection volume: 4 ul
The peak shape is very good.
Even when using an internal standard (DMAc for instance), I am observing a poor reproducibility of peak areas and Relative retention factors, with 7 % RSD for 12 injections.
What could be the problem?

[Consumer Products Guy]

I have a problem of poor reproducibility in GC peak areas form injection to injection, and I’d like to understand the possible causes.

I’m currently doing manual injections on a Clarus 580 GC with FID detector.

Wow, I didn't realize that some folks were still using manual injections !!!!

The sample if Diethylhydroxylamine (dissolved in water). Injection volume: 4 ul

This information is important. Water expands a ton in the inlet and can cause strange things to happen ! And when we (rarely) used water as solvent, we kept injection volumes to 0.5 microliters.

[Plunkett79]

Well, I am stuck with this system for now. An autosampler is installed on the other channel, but it's for headspace.

So if I want to inject a 0.5 ul for instance, I should get a 1 or 2 ul syringe, I guess?

[Jake]

As CPG points out, you're doing a split injection with water as the solvent. For an injection of 4 uL of water, the vapor volume will be over 4 mL. If your liner volume isn't about twice that (I've never seen an inlet/liner with an 8 mL volume), you're going to have injection issues.

Although Agilent-specific, here's a good overview of the challenges of injecting water into a GC:

https://www.agilent.com/cs/library/slid ... VbbvZxIPqB

[DR]

While I agree that a smaller syringe will tend to increase your accuracy, your reproducibility will be very technique driven.

When I had to do injections of "wet" things, I'd program my injector to make the injection very slowly (like 10 seconds/µL or even slower).

Reproducing this with manual injections would be very difficult, but it helps prevent the water from flashing into everything.
Also, use a big, wide open liner.
As you're using an internal standard, the only RSD you should care about is that of your response ratio. Is that OK?

[Peter Apps]

Welcome to the forum.

I also do manual injections !

Why are you injecting such large volumes, is that necessary to get the required signal:noise on the analyte peak?

Have you checked whether decreasing the injection volume actually decreases the peak size - at 4ul you will be creating a huge pressure pulse in the inlet that will play havoc with your split ratio.

Your carrier gas flow rate is way above optimum, is there a reason for 2ml/min and for using nitrogen?

Is this a method that you are not allowed to change, or can you alyer settings as necessary?

Peter

[Plunkett79]

Many thanks all for your kind replies.

Jake:
I’m using this 4mm liner (part number N6121020) which has deactivated glass wool, based on a suggestion from PE.

Peter:
I have all the leeway that I want to alter the method.
Peak response is very good, so I guess I can decrease the injection volume. I will order some smaller syringes.
For the carrier gas, I’m also stuck with nitrogen for now. What will be the effect of altering the flow? It’s about Van Deemter, isn’t it?

DR
I understand that I just need to care about the RSD of the response ratio, but at the moment it’s way too variable.