GC/MS Large volume injection

[sav]

Right now I am using the Restek 5.2mmID Silcosteel single gooseneck liner, which is very thin walled to give more volume for evaporation.

Instead of the Siltek deactivated borosilicate wool I have been using, I am going to take some plain glass wool and silanize it in the liner and see if that works better.

IMHO, Silcosteel isn't best choice as well as manually silinized glass wool in this case (its not help). You need just good deactivated single gooseneck glass liner with carbofrit (Restek Sky liner etc. with carbofrit). The carbofrit works much better than any super best deactivated glass wool when you dealing with active analytes but, honestly, i don't know how it will work in CSR LVI. Also doble gooseneck is better than single gooseneck in CSR LVI.

[Peter Apps]

whoops double post, sorry

[Peter Apps]

My experience with carbofrit is that it is great for everything that does not have an aromatic ring. An inlet with a glass frit might be worth a shot because it ensures that only sample vapour gets to the column tip, rather than a messy mix of vapour and droplets.

Peter

[James_Ball]

I have some double goose neck cyclo liners from Restek I may try, though I will be sacrificing volume for expansion with those. I have used them in the past with good results.

I was wondering about the carbofrits, I had never used them before. If the problem is with aromatics, then that would not be good since I have at least 50 compounds in the mix that are aromatic. Most would be PAH or chlorinated aromatics.

[Rndirk]

Since James will be using an Agilent MMI and you're discussing liners and LVI i'd like to trow in a question: I'm doing 25µl injections with solvent vent using the Agilent MMI with 5190-4006 liners.

Happy times, but they are quite expensive. The thing is i can't find alternatives without glass wool. If i would use a liner with glass wool for analytes that doesn't require the inertness; would I have to change my solvent vent parameters?

[sav]

I had some "miracles" with active compounds which not had aromatic rings (although they had conjugated pi bonds) and carbofrit, but i did not try carbofrit with mid and high bp compounds with aromatic rings (but https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3824841/ ). Theoretically a carbofrit can bring trouble for mid and high bp planar aromatic molecules (some PAH etc.). May be deactivated glass frit is more "safe" choice in this situation, but definitely its not best choice for CSR LVI.

I had a lot of headache with single gooseneck cyclo liner (Restek) in split mode - poor RSD of peak area (low bp compounds; i did a lot of optimization procedures and did not solved this issue). Also i think its liner (even doble gooseneck) is not appropriate for CRS LVI (relatively long distance between solvent deposition place and entrance of retention gap; also relatively low effective vaporization volume). The CRS LVI also work with empty liner (single or doble gooseneck), but recovery and reproducibility is better with glass wool (also much lower backflash). So may be the easiest (from chromatography point of view) way is hydrogen (single taper liner with glass wool).

[James_Ball]

Since James will be using an Agilent MMI and you're discussing liners and LVI i'd like to trow in a question: I'm doing 25µl injections with solvent vent using the Agilent MMI with 5190-4006 liners.

Happy times, but they are quite expensive. The thing is i can't find alternatives without glass wool. If i would use a liner with glass wool for analytes that doesn't require the inertness; would I have to change my solvent vent parameters?

For the full EPA8270 list I am thinking the MMI will be the way to go versus CSR since I can inject into a cool inlet. I don't have any experience yet with this technique but I hope others here have and can answer the question. The CSR method I do believe will work for the less labile tests such as PCBs, DRO, and others.

[James_Ball]

I am still working on this, and I have had some success with the CSR-LVI, but balancing the peak shape of the late eluting PAHs with breakdown of labile compounds is difficult using the standard injection port.

Doing a ton of literature review and I came across a comment I believe was from Jack Cochran over at Restek where he described the fast injection speed as having a "thermospray" effect. That got me to wondering, which is usually dangerous, but I came up with an idea I wanted to throw out to everyone. What if a change could be made in the liner that would place then end of the needle in a narrower liner bore, with the lower half having a larger diameter for more volume? This would give the carrier gas a much higher linear velocity at the syringe tip, which could lead to better dispersion and evaporation of the solvent, preventing backflash and getting the sample volatilized before it contacts the lower gold seal/column tip which should reduce breakdown and also improve response of the high boilers. I am thinking something like this.



Crude drawing I know, and I was thinking maybe the exit of the narrow part could be flared or restricted whichever would lead to better evaporation. I'm not much on fluid dynamics but I know how an ESI and APCI probe works this design is also similar to the nebulizer used on the ICPMS. Could even give the exit of the narrow part a guide to cause the gas to swirl in the liner for better mixing.

[Peter Apps]

As far as I know you only get thermospraying if the needle is in the inlet long enough for the needle to heat up - which the fast injections on modern autosamplers were specifically designed to avoid. This is partly why you can get much better performance with higher boiling solvents if you increase the pre-injection dwell. What you don't want though is for the sample to fractionally evaporate out of the needle, leaving some of the heavier analytes inside it. The heat exchange between gas and liquid as really poor - most evaporation happens when the sample drops hit a solid surface, which is where glass wool, frits and baffles help; the Grobs showed that it is quite possible for a stream of liquid sample to pass all the way down an inlet liner and hit the bottom seal. Even when the drops hit a surface all is not necessarily well, they can skate around and jump back into the carrier gas due to the Liedenfrost effect (spit on a stove).

The take home message is that expecting good things to come from flash injections is a triumph of hope over physics - PTVs make much more sense; you dispense a precisely metered volume of solvent into a cool container, heat it up to vapourise it and transfer it to the column nicely mixed with carrier gas, then focus the analytes onto the head of the column with stationary phase focussing. You also get the option to vent the solvent, so you can do large volume injections without flooding the column and crudding up the detector.

Peter

[James_Ball]

Thermospray was just what Cochran gave as a somewhat descriptive name for it. What I am thinking about is how the carrier gas is flowing in the area of the needle tip upon injection.

In a normal splitless injection liner which has an inner diameter of 4mm, with a flow of anywhere from 1-4ml/minute depending on if you are using a pulsed injection or not, the linear velocity around the needle tip is going to be low, which is what can allow for backflash to occur and does not promote evaporation of the solvent until it contacts solid surfaces, which also promotes breakdown.

If the linear velocity near the needle tip is much higher, then you will have the effect of the solvent being atomized giving much smaller droplets which will increase their volatility and should allow them to evaporate within the carrier gas stream. One problem I can see would be when the carrier gas enters the larger portion of the liner that the expansion would induce a cooling effect which could be counter productive. To make it work though I am thinking the inner diameter of the narrow portion of the liner would need to be very near the outer diameter of the needle to achieve the maximum carrier gas velocity.

This is mostly just fanciful thinking while I have time to sit between injections of each test shot so it may not be anything that would actually work, just wanting to run it by others to see if I am totally crazy or not.

[GOM]

Hi

Peter wrote "a triumph of hope over physics "

Love it - I am going to steal that one!

If you contact Ray Perkins at Anatune he may well be able to offer some useful and valuable insights.

I bought his (now old) ATAS Optic LVI system after a visit to Riva and was very happy with it (100uL injections)-

Try Googling anything to do with Koni Grob or Hans-Gerd Janssen and large volume injection

or http://www.glsciences.eu/html/lvi-injection.html

I am still looking for my notes and comments on this course in Riva

At a similar time Agilent released a LVI system but they considered 25uL to be large volume

Do we need to define what is large volume?

Ray was extremely helpful to me when I was experimenting with manual LVI. He enjoys theoretical discussions.

I have mentioned this LVI before in posts

And yes, it does come down to the solvent BP, speed of injection, PTV, split time and, quite importantly, the liner design

Regards

Ralph

[Peter Apps]

Getting a stream of liquid to spray out of a smooth needle is droplets that are small enough to evaporate quickly is more of a challenge than it sounds - that's why thermospray-MS interfaces are historical curiosities. If you pass a fast steam of gas axially along the needle all it does is to stretch the liquid stream out and make it break into big drops. You can get smaller droplets by rapidly vibrating the needle, either transversely or axially, or go the electrospray route by putting a large voltage on it (I recall hearing that an electrospray inlet was presented at Riva (?) a few years ago).

Even with a rapid flow of gas an injection from a needle that is a close fit to the inlet liner gives a sample film trapped between needle and inlet wall by capillarity - when you pull the needle of the inlet that liquid gets smeared onto the bottom of the septum. You can stop the capillary creep by having the liner above the needle tip wider than it is just at the tip - so the inlet profile is pinched in exactly where the needle tip sits as the injection is made - something like a Restek precision liner without the glass wool.

Air bubbles and solvent plugs above the sample in the syringe barrel are conventional ways of ameliorating some of the problems with injections to hot inlets - sometimes they work well enough to be worth the extra trouble.

Peter

[James_Ball]

Interesting tidbit to go with all this.

I just made some injections with a 2mm liner with a very small amount of wool at the bottom and 10ul injection volume. I have better peak shapes now than when I used the 5mmID liner. Using pulsed injection I am at about 2.8ml/minute carrier flow at injection and the liner volume is 140ul so I am sweeping the liner volume 4x with a 0.2minute splitless time. I just did it to see what would happen, expecting big blobs for peaks and was surprised at the result. I think this is proving what Jack Cochran was saying about the CSR technique in that the syringe is shooting the solvent in a tight slug directly to the bottom of the liner, which is why they were not seeing backflash, and it even works with smaller volume liners. By reducing the diameter of the liner from 5mm to 2mm the linear velocity at the same flow rates is more than 4x faster, which should help force the sample into the guard column and reduce the time it spends in the inlet, which may be good for the labile compounds. That is what I will look at next.

[Peter Apps]

With a 2mm liner the needle only has to be 1mm out of perfect alignment for the tip to touch the liner wall, so you might be transferring the sample direct to a hot surface (good) rather than squirting it into a gas flow.

What point style do you have on your needle ? - anything other than a cone squirts liquid sideways onto the wall rather than axially down the bore of the liner.

Peter

[GOM]

Something in Peter's previous posts stirred a recollection of using a side vent needle.

I have tried contacting Ray at Anatune today to discuss it on your behalf but he is in a meeting all day.

He will get back to me tomorrow and I will post any references / comments

You originally mentioned that 12.5uL is your large volume.

Is there any reason for this rather specific volume apart from the paper ( and being exactly half of 25ul )?

Is that the volume that you wish to inject?

Regards

Ralph