Chromatography Forum

Hello,
I am separating a C8-C20 series of n-alkanes dissolved in chloroform on a narrow-bore 25m x 0.15mm i.d. x 0.25µm film BP-1 column with detection by TOF MS. 1 uL is injected at 1/10 split (1 mL/min He column flow, 40cm/s average velocity, 41 psi head pressure at start column temperature of 40°C) using a CIS-4 (Gerstel) injection port in constant temperature mode at 320°C. The liners used in this injection port have a nominal volume of ~200 µL. Under the injection conditions used the solvent expansion volume (for chloroform) will be ~160µL.

If I use a deactivated glass wool liner, I get 2-3X the peak height for all peaks (so no evidence for BP discrimination) compared to using a baffled liner. Why should this be, since both liners have the same volume? The outcome suggests I am losing sample out the top of the baffled liner - but this should not be so as liner volume > expansion folume and there is effectively a 10mL/min He flow through the liner. Is glass wool known to slow down the vapour cloud expansion process?

I've also tried a cold split injection with both liners where the injector temperature is ramped from 10°C to 320°C over 25s. This should reduce any liner overfilling effect that might happen during a hot split injection. However, with either liner, the response for all peaks is lower with cold split. The peak shapes for the more volatile alkanes are broader with the cold split technique, which is understandable given that they won't be efficiently trapped on the liner at 10°C. However, for C20 alkane, which should be fully retained on the liner at 10°C, the response is lower for either liner in cold split compared to hot split, even though the peak shapes are identical. I can't explain this either!

Any insight would be greatly valued.

thanks
Tony

What goes on when a volatile organic solvent is injected into an inlet at 200 C above its boiling point is not simple, and vapour overflow is an issue only with splitless injections. Best guess with the baffled liner is that droplets are skidding off the hot surface like spit on a hot stove, and getting past the split point without vaporizing completely. With the glass wool evaporation is more efficient because the liquid wets the surface and heat transfers more effectively.

As an additional complication you will find that the time the needle spends in the hot inlet before the sample is injected will also affect the peak areas, and their repeatability from run to run.

There is a good chance with the cold split that the split ratio is changing as the inlet heats up.

Are peak areas affected in the same way as peak heights ? - it could be that the peaks are getting higher and narrower.

Peter

Hi Peter,
Thanks for your useful comments.

I think the incomplete vaporization theory is a good one for the baffled vs glass-wool liners.

However, I'm not so sure of the reason why cold split should give a poorer response than hot split for a relatively high BP analyte, like C20 alkane. I don't quite follow your reasoning for the split ratio changing during heating of the cold injector. For lower response to be observed, the split flow would have to be higher than its set point during sample transfer onto the column. I guess this might happen momentarily during liner heating if vapour volume and pressure increases and the electronic flow control (Agilent 6890) lets this excess out by opening the the split valve up. But is this likely?

The peak shapes I have with cold vs hot split methods are identical when normalised - i.e. the height/width ratios are the same.

cheers
Tony
T

Hi Tony

The difference in areas between hot and cold could be due to a fluctuation in split ratio when injecting to a hot inlet - in other words that the split is "correct" with the cold injection. And unfortunatley EFCs are prone to erratic behviour in the face of pressue and flow transients. How does the repeatability compare between hot and cold ?

Peter

Tony,

I agree with Peter on the wool versus baffle. Wool will hold things up while a baffle lets them bounce around and possibly out. Carlo-Erba had some really great video on this stuff.

The more interesting one is the C20 discrimination. Two things to chew on. First, the EPC is setting for the column, not the inlet so the rapid change in injector temperature could be affecting the gas viscosity in the inlet with the EPC happy as a clam. My problem with that is that the volume in the inlet is pretty small but relative to the total volume of gas in the system at any one time, maybe not. Second, have you tried a really fast inlet rate, like 5 seconds to 320? Maybe the issue is that the C20 is not peeling off correctly, not that it is not sticking correctly.

Just some thoughts to add in.

Best regards.

Repeateability is fien with cold split injections. I haven't tested thoroughly with hot split. For now, I am just accepting the response is lower for cold split (for whatever reason), as I don't have sensitivity issues and I think it is at least in theory kinder to the sample and better for discrimination to stick with the cold split technique.

Also, I cannot ramp my PTV any faster than I am doing now - it is already heating at its maximum rate.

thanks
Tony

HI Tony

I would agree that cold split is the way to go. Do not worry about heating rate - as long as the inlet is hotter than the column ( and with a 40 C column start temp it will be) and the inlet is clean (NB) stationary phase focussing will sharpen the peaks for all except the very early eluting compounds.

Happy New Year

Peter