Carryover or ghost peaks of ethanol and propan-1-ol

[PhilLeigh]

I analyse ethanol content in pharmaceuticals using propan-1-ol as internal standard. The matrix is 100% aqueous and the column is PoraBond Q 30m x 0.32mm x 5um PLOT column.

A System Suitability requirement is that after an injection of Standard solution, an injection of deionsed water shall show no carryover. In practise, very small carryover peaks are always seen. Repeated injections of water are 'cleaner' but can still show some trace solvent peaks.

I inject 1ul through a 1:50 split. Despite the large injection volume of a hugely expansive solvent I get good symmetry and reproducibility using a stainless-steel liner (5.2mm i.d.) with a "homemade" insert of paper (for example filter paper)

I use wash vials (either water or methanol) before and after inejction.

The wash vials have polymer needle guides but the waste vials are open.

I have used new injection septa - still the same problem. Likewise I have tested new syringes (5ul) - agan the same problem.

My equipment is Agilent 6890. Column 150C; Inlet 150C; Detector 240C.

Ideas would be appreciated!

PhilLeigh

[dblux_]

Do you use separate wash vials before and after injection ? How many washes ? Do you wash syringe with your blank (deionised water) ?

Check whether washing is sufficient enough.

Is filter paper in your liner resistant to 150 degC ?

[chromatographer1]

You have drug component flashed somewhere in your injector and/or on your column.

Every time you inject a sample or std containing the alcohols they leave behind some of the alcohols. These get rinsed out with the blank injection following, thus your carryover.

You must find and remove this material.

Good luck,

Rod

[Peter Apps]

Symmetrical peaks do not necessarily mean that you have no adsorption of analytes - it can also occur when the adsorption is so strong that the tail is so flat as to be invisible. In this case I think that you have very strong adsorption to the filter paper as it dries in the hot inlet - then when you inject water again some of the adsorbed material is displaced by the water to give the carryover peaks. Try with a different inlet packing.

As you note - there will always be some carryover, so you need to change the criterion to a not greater than X, where X will not affect subsequent analyses.

Peter

[KM-USA]

How small are these carryover peaks? Less than 0.5% of the peaks in real samples or standards?

Have you tried using the autosampler in the slow inject or slow plunger mode? That's what we use for our ethanol assays; but we use USP <611> and capillary column, with different intermal standard.

[PhilLeigh]

How small are these carryover peaks? Less than 0.5% of the peaks in real samples or standards?

Have you tried using the autosampler in the slow inject or slow plunger mode? That's what we use for our ethanol assays; but we use USP <611> and capillary column, with different intermal standard.

The carryover peaks are tiny, typically <0.05% of a standard response. I shall certainly try the 'slow injection' option. However, my recent findings have shown that the paper insert that I use (which improves reproducibility and symmetry with 'difficult' sample matrices) may itself be the cause of the problem. When injecting standard solution using a glass cup-splitter liner (with no additional insert) the carryover problem disappears.

Many thanks for your response

Phil

[houseofbird]

First, it is hard to split water well. Not impossible, though.
I suggest a re-vamp.
I do oxygenates of gasoline in water. MEoH, EToH, N-butyl, 2-Heptanol, sec-Butyl and many other alcohols.

I use 6890 FID
I have great success with this setup:

Column: Restek Rxi-1ms, 30M, 0.32mmID, 4.0um film thickness. CAT#13396
Guard Column: Restek Polar-Deactivated Guard column 10M. CAT#10069
Inlet liner: 4mm Cyclosplitter from Restek too. CAT# 20706

Inlet: 220C Split 1:1
Oven:80C for 0.5min..then 15C/min to 120C...then 30C/min to 200C
FID TEMP: 340C
Set to "Constant Flow" with flow rate @ 2.5ml/min
Inject only 0.5ul
Detector flows: H2 25ml/min....AIR 250ml/min....Makeup (N2) 25ml/min no constant column makeup
Carrier gas: He (really important)
Use a Merlin for your septa if possible
Also, I use 1,4-dioxane as ISTD. I find 1-propanol a little unstable and prone to tailing

Key point, I don't have carryover problems and I am looking at ug/L levels of detection. And often times I will get samples that are at % levels.
If all this doesn't help you, try removing your septa purge lines and give them a good cleaning. Maybe replace the chemical trap.

The idea of slow injection is a good one, I'm gonna try that myself, today.

Hope this helps.

[PhilLeigh]

Dear chromatographer,

Thank you for interesting and comprehensive reply. There are many similarities in our applications. The major difference seems to be that I have involatile excipients which has influenced my choice of inlet liner.

Some of your suggestions I can carry out now without any changes to our issued methods for example choice of septum, cleaning out purge lines, change chemical trap.

Thanks again,

Phil

First, it is hard to split water well. Not impossible, though.
I suggest a re-vamp.
I do oxygenates of gasoline in water. MEoH, EToH, N-butyl, 2-Heptanol, sec-Butyl and many other alcohols.

I use 6890 FID
I have great success with this setup:

Column: Restek Rxi-1ms, 30M, 0.32mmID, 4.0um film thickness. CAT#13396
Guard Column: Restek Polar-Deactivated Guard column 10M. CAT#10069
Inlet liner: 4mm Cyclosplitter from Restek too. CAT# 20706

Inlet: 220C Split 1:1
Oven:80C for 0.5min..then 15C/min to 120C...then 30C/min to 200C
FID TEMP: 340C
Set to "Constant Flow" with flow rate @ 2.5ml/min
Inject only 0.5ul
Detector flows: H2 25ml/min....AIR 250ml/min....Makeup (N2) 25ml/min no constant column makeup
Carrier gas: He (really important)
Use a Merlin for your septa if possible
Also, I use 1,4-dioxane as ISTD. I find 1-propanol a little unstable and prone to tailing

Key point, I don't have carryover problems and I am looking at ug/L levels of detection. And often times I will get samples that are at % levels.
If all this doesn't help you, try removing your septa purge lines and give them a good cleaning. Maybe replace the chemical trap.

The idea of slow injection is a good one, I'm gonna try that myself, today.

Hope this helps.

[Peter Apps]

First, it is hard to split water well. Not impossible, though.
I suggest a re-vamp.
I do oxygenates of gasoline in water. MEoH, EToH, N-butyl, 2-Heptanol, sec-Butyl and many other alcohols.

I use 6890 FID
I have great success with this setup:

Column: Restek Rxi-1ms, 30M, 0.32mmID, 4.0um film thickness. CAT#13396
Guard Column: Restek Polar-Deactivated Guard column 10M. CAT#10069
Inlet liner: 4mm Cyclosplitter from Restek too. CAT# 20706

Inlet: 220C Split 1:1
Oven:80C for 0.5min..then 15C/min to 120C...then 30C/min to 200C
FID TEMP: 340C
Set to "Constant Flow" with flow rate @ 2.5ml/min
Inject only 0.5ul
Detector flows: H2 25ml/min....AIR 250ml/min....Makeup (N2) 25ml/min no constant column makeup
Carrier gas: He (really important)
Use a Merlin for your septa if possible
Also, I use 1,4-dioxane as ISTD. I find 1-propanol a little unstable and prone to tailing

Key point, I don't have carryover problems and I am looking at ug/L levels of detection. And often times I will get samples that are at % levels.
If all this doesn't help you, try removing your septa purge lines and give them a good cleaning. Maybe replace the chemical trap.

The idea of slow injection is a good one, I'm gonna try that myself, today.

Hope this helps.

That method has some interesting details, it's always good to see alternative ways of tackling things.

For trace alcohols in water I would not expect a methyl silicone column to separate the smaller ones very well - is that why you use a thick film ? And I would expect the water to eat the phase pretty quickly - what kind of column lifetimes do you get ?

Most GCs struggle to maintain very low split ratios, and 1:1 ends up giving the worse of both split and splitless injections. Have you tried higher split ratios, or real splitless ? How do they look ?

The programme rates are very high for the length of column - presumably you have resolution to spare towrds the end of the chromatogram ?

Why is it imprtant to use helium as carrier gas - and compared to which other gasses ?

Thanks

Peter