Manual Injection woes

[IdiotFool]

I was recently hired as a Quality Control Analyst for a wood stain manufacturing facility. They've not had a chemist in the lab for at least 8 years and only Chem Engineers have been around to do basic tests. They have a HP 6890 GC with manual injection that I've been able to get up and running. However, I only have experience with auto-sampler GCs.

Here's the deal, as far as I can tell: they're using a 13 ft. packed glass column with a simple glass liner/insert inlet and ECD. I believe this indicates an on-column injection, but I don't know for sure. My experience with GC was a split/splitless system with capillary column and FID. This thing is a new beast entirely.

About 30% of the time, my injection shows no peaks. I'm using a 0.5 uL gas-syringe and injecting 0.25 uL of sample. My technique does not deviate, so I have no idea what the problem could be. Any thoughts? What IS acceptable injection technique, given these conditions?

[Consumer Products Guy]

What IS acceptable injection technique, given these conditions?

Acceptable injection technique for quantitative results is to get your new boss to buy an autosampler, so you can (1) obtain better results (2) work on something more important than do injections.

Believe me, I did manual injections for years before decent autosamplers were available; injection technique is very important for any type of quantitative work. You'll have a tough time getting good %RSD with manual injections.

That said, I've never heard of a 0.5ul gas syringe. Wouldn't any syringe that small need to have a wire-in-needle set up? Is your 6890 GC even computerized? You really need a computerized GC, with EPC inlet control (do you have that). If it was me, I'd lobby for a computerized, autosampler-equipped GC, and likely convert my test methods to use capillary columns. Tell the pointy-haired boss how you'd be moving into the 1990s, and be more productive. You can probably get the GC field-upgraded for all that by Agilent instead of buying a new GC. Tell boss the vendors will deal, want sales.

[Don_Hilton]

If you have a packed glass column, I expect that the end of your column comes up into the inlet - and your injection is made directly into the end of the column. Correct insertion depth of the column into the inlet is important.

For maual liquid injection technique, I would suggest a sandwich injection and, if you can, a larger syringe - like 10 microliters. (My memories of injections on a packed column was one or two microliters.) Draw about 1 microliter air into the syringe, followed by your sample (one to five microliters) and then an additional 1 microliter of air.

On injection you have two possibilities - hot or cold needle. The analytes and solvent may work better with one than the other. Cold needle is push the syringe through the septum and quickly inject the sample (like the Agilent autosampler). Hot needle is insert the needle to depth, hold if for a few seconds so that it heats and then inject at an a moderate speed - it should take a maybe half a second or so. The air plug drawn into the syringe should ensure that your sample is expelled, so you should not have to heat the needle after you push the plunger.

[AICMM]

Idiotfool,

A couple more comments for what they are worth, then a question. Don is right about wanting to know where your column packing begins. It would be worth your while to cool your inlet and pull the column out and see where it was situated in your inlet. Second comment, I would suggest you look at the type of syringe you are using. With 13 foot of column, you probably have a lot of head pressure so if you syringe is leaky you might be just blowing the sample out. More information about your instrument conditions is always useful.

The question, a wood stain manufacturer using an ECD? I would have expected an FID, sure, but what are you tracking by ECD (if you can say?)

Best regards,

Matthew

[Suresh Seethapathy]

I can't think of a reason why you see peaks sometimes and sometimes not. If you are using a 0.5 uL syringe, i assume that it is a plunger-in-needle syringe which in my experience has a larger needle OD which in turn might damage your septum quickly (resulting in leaks). If it is a plunger-in-needle syringe that you are using, you obviously wont be able to do a "hot needle" injection as Don is suggesting.

Besides, i would never do a manual injection (for quantitation purpose) without an internal standard. I have done manual injection with internal standards and it works like a charm.

Suresh

[IdiotFool]

Why would the hot needle method not work if there's a wire insert? I've had much more success leaving the needle embedded in the syringe for a few seconds before injecting over the "sandwich method".

I've turned the pressure of the column and inlet down, and haven't seen any other problems with missing injections. The peaks are much larger than before, and everything is going much better. It's still a pain that nobody has any info or methods for what they used to do, but such is life.

As for getting an autosampler, I'd love it, but the GC hasn't been run in 5 years. That I'm here to run it all is much more than they used to have. I'll be hard-pressed to convince them to spring for an autosampler for a long while.

Thanks, all, for the help.

[Suresh Seethapathy]

In the hot needle injection technique you have to move all the sample into the barrel of the syringe and keep the needle empty by retracting the plunger. You will then have to heat the needle first for a few seconds (hence the name hot needle) before you let the sample solution enter the needle part. In the plunger-in-needle syringes, and with the ones i have handled, even if you retract the plunger to its maximum the sample still will be left in the needle part. So when you are in the heating phase of the technique, you still have liquid in the needle and the solution keeps evaporating the whole time the needle is in the injection port. I hope that explains why you can't use hot needle technique with your syringe.

[IdiotFool]

That makes sense, yes. Thanks.