Purge Flow for Splitless Injection

[Mike H.]

What is the flow range for this parameter? I'm using a 30m x 0.53 x 1.0 column. I tried to look this up. One site suggests a range from 100-200ml/min and another site suggests 10-25ml/min. Can anyone shed some light on this.

Also is it even necessary to use purge flow?

[tlahren]

You are using the 7890 right? For splitt-less i usually have the purge flow set to 50 mL/min. Probably the most crucial part is the split-less time you chose. This can and will greatly affect your analyte response and may cause irregular peak shape if it is too long. If you are analyzing low BP compounds I think 0.5 min or less is a good start. But I usually optimize it by starting with a really low split-time and move up until you stop seeing either increase in response or poor peak shape.

What is your injection volume and solvent? What's your column flow rate (mL/min)?

[Hornet]

Like tlahren said the most important thing is the purge time and not the flow, i would stay around 50 mL/min or higher anyway.

[Don_Hilton]

You need suffient flow to sweep the inlet clean quickly enough to stop transfer of any materials from the inlet to the GC column - like the tail of the solvent peak.

With a purge rate of 100, you will exchange the carrier gas in the inlet in a couple of seconds or less. While this is effective, it can use a considerable quantity of gas.

Try a few injections and see what it takes to cleanly end the transfer of materials from the inlet for your samples. If you are trying to analyze small peaks that come close to the end of the solvent peak - and you are using the purge time to trim the tail to avoid interference, you may need to be more aggressive with the purge of the inlet.

[Mike H.]

You are using the 7890 right? For splitt-less i usually have the purge flow set to 50 mL/min. Probably the most crucial part is the split-less time you chose. This can and will greatly affect your analyte response and may cause irregular peak shape if it is too long. If you are analyzing low BP compounds I think 0.5 min or less is a good start. But I usually optimize it by starting with a really low split-time and move up until you stop seeing either increase in response or poor peak shape.

What is your injection volume and solvent? What's your column flow rate (mL/min)?

I'm using an Agilent 6890. Solvent is Water. Analytes are MeOH, ACN, IPA and N-methylpyrollidone. Injection volume in 0.1µL. Column flow is 4.6 ml/min.

[tlahren]

As Don said, 100 mL/min purge rate will flush the injection port fast. If you want to calculate how long it takes to flush the injection port I use the following. Liner volume is about 0.5 mL for a 4mm ID Agilent liner. A good "flush" would be at least 3 liner volumes at 100 mL/min =(0.5mL × 3)/100mL/min = 0.015 min = 0.9 sec. You can use the same logic for calculating the time needed for an adequate transfer (split-less time) during the injection as well. Instead of using 100 mL/min as the flow rate you will only have the column flow rate of 4.6 mL/min. These calculations have been a good starting point for me in the past when setting up a method from scratch. Maybe someone here has a better suggestion. I think 100 mL/min will work and about 0.5 min split-less time. If you set your gas-saver setting to 20 mL/min at about 2 or 3 min or so this will save on wasted gas dramatically.

Hope this helps somewhat.

[aldehyde]

I think the benefit to additional purge flow is to sweep it through the unheated split line faster. As noted the volume of the liner is small enough that even at 30-50 mL its going to clean the liner area pretty quickly.

[Mike H.]

The best I've gotten with the single taper liner is about 14%RSD. The parameters were 25ml/min purge flow and 0.05 min purge time. It's a bit frustrating, but what are you going to do but keep trying?

Also on a side note when using a double tapered liner at a 200ml/min purge flow and 0.05 purge time, RSDs were <5%. However the actual method will contain nonvolitile sample which will quickly dirty up the liner. So this is why I'm trying to get a single taper liner to work.

[Peter Apps]

With a 0.05 min (= 3s) purge time you have all the disadvantages of both split and splitless injections, and none of the advantages of either. Increase your splitless (=purge on) time by a factor of ten and you will see a remarkable improvement.

Peter

[Mike H.]

Thanks I will try this.

[Mike H.]

How is it that the purge flow and purge time can affect analyte peak shapes when it occurs after the sample has already been injected on the column?

[tlahren]

"split" flow and split-less time control how much and how long the analytes are in the injection port. It is not necessarily true that once the sample is injected it all goes to the column immediately. Most of it will within a second but a small amount may hang up in the injection port. This small part enters the column later than the first part thus causing a later eluting "tail" on the peak. A higher split flow will effectively "wash" out this tail before it has a chance to enter the column. It's also good for flushing non-volatiles out of your injection port that may not have fully transferred to the column.

One of the most important aspects of GC is to control analyte transfer to the column. Injection port temp, flow, pressure, reactivity etc., all play a role in the transfer. And every compound reacts differently. This is why final parameters are usually determined through trial and error but with some sort of strategy applied.

[Don_Hilton]

The purge off time is quite critical. When the sample is injected, the vapors spread in the inlet liner. This depends on the liner, expansion volume of the solvent, injetion technique, and such. If you assume 1 mL volume for the inlet liner it would take 0.67 minutes to push the full volume of the liner onto the GC column if the flow rate were 1.5 mL/min (a slightly faster than optimum rate for a 0.32 column). In the first 4 seconds, only the first 1/10 of the inlet volume would be pushed onto the column - not enough to move much of the sample into the column. If your purge off time is too short, the contents of the inlet will be incompletely transfered to the column.

You should not have to transfer the full volume of the liner to the column, as very little of the sample will evaporate into the upper portion of the liner. But it is not bad idea to make a few injections to make sure that material is transferring properly to the column and that peaks are reproducable.

[Mike H.]

Ended up using these parameters. 3.5ml/min and 0.5min. Precision and linearity are very good now, but as luck would have it I ran into another issue that I didn't anticipate. I used a 5 point calibration curve of 100µg/ml-500µg/ml and assumed that I'd make 100 ppm the quantitation limit, but later learned that while the solvents in water are infinitely soluble, the compound I will be testing for solvents is not. Max solubility is close to 50mg/ml. So the actual amount of NMP injected on column in my 100ppm standard is ~ 10ng, but an actual sample consisting of 100ppm of NMP would only be injecting 0.5ng.

Does anyone have advice on how to proceed?

[Don_Hilton]

Avoid using water as a solvent. Use something that will disolve all your analytes and the compound you are testing. THF? MTBE?

(Without knowing your sample preparation, it is hard to give much useful information.)