Tried something new with my GCMS, strange results

[James_Ball]

Ok, so I have been working on switching to H2 for carrier with not so much luck. So today I tried Nitrogen carrier just for fun. Tunes will pass for BFB, sensitivity is lower as expected but overall it works. Problem I am having is I am getting a high background of m/z 42. Ar=40 CO2=44, so what is 42? The 42 is about 10x higher than 40 so I figured it wasn't related to Ar, unless it is something really strange like ArH2 or ArD (no way it is those lol).

This is for volatiles analysis so the only problem has been needing to pump up flow to 0.5ml/min at desorb so I can get enough flow to desorb the trap, then drop the flow back to 0.2ml/min to keep the analyzer pressure down about 5*10^-5. Lowering from 0.3 to 0.2 ml/min actually increased the response for m/z 69 five fold. Using a 40m x .18mm column. I have strange half split peaks for the xylenes, ethylbenzene, but otherwise the chromatography looks good. It would probably work better on one of my turbo units instead of the diffusion pump 5973, but it is a first try, not expecting miracles .. yet

[JMB]

If m/z 42 is the molecular ion, then possibilities are ketene (CH2CO), isomers of propane (C3H6) and isomers of CH2N2.

[Don_Hilton]

With plenty of N2 and background air, CNO as a possibility?

[gapeev]

Answering the question asked, it is likely N3+.

Answering the question that has not been asked, using N2 is one of the last things you want to do.

[Peter Apps]

Answering the question that has not been asked, using N2 is one of the last things you want to do.

Why ??

Peter

[Don_Hilton]

I tried N2 as a CI gas (trying to avoid the source fouling one gets with methane for ECNI) and there was no damage to the instrument that we were aware of. A CI gas is used at a simiar rate to column flow or even a bit higher. Nitrogen turns out to be a poor CI gas - and I was finally able to find that this was known and published back in the late 60's - not as easy to find as newer literature. (We decided, despite the nature of the beast, to use NH3 - and it works quite well.)

Some tripple quads instruments use N2 in Q2.

As a carrier, N2 requires lower linear velocity in the column. In the mass spec, you need to start your mass range for acquisition high enough to avoid N2.

So, I am also curious as to why N2 should be the last thing one would want to do.

[James_Ball]

Answering the question asked, it is likely N3+.

Answering the question that has not been asked, using N2 is one of the last things you want to do.

N3, that is one I never thought about, but it works out nicely, especially with so much nitrogen in the source. I doubt it was any of the others since it was coming through a cool column with only carrier flowing, and had just had H2 running through the system which should have knocked out most of any contamination present.

As for using N2, we are looking for alternatives to Helium since we currently use at least two tanks of Helium each week. As the price continues to increase alternatives are looking more and more attractive. Hydrogen being the preferred alternate carrier gas but working out the problems passing a BFB tune have not been easy. Getting the carrier flow down to 0.3ml/min gets it very close but will not pass tune consistently. I did see that someone (I believe it was from the Tel Aviv university) suggested trying a mixture of H2/N2 in which there is a ratio that would mimic the flow characteristics of He and be less reactive. I don't have the equipment to do a blend so haven't tried that yet.

In my little 3 hours or so experiment last week I found with N2 the chromatography works well except for the first freon peak which I expected, and the Ethyl Benzene, Xylenes peaks near the middle of the chromatogram which had shoulders on the tailing side almost to the point of being split peaks. I also had to raise the threshold since I was getting an across the scan range background of about 100-150 counts for almost all the masses which I did not expect along with about a 1-2% background level of m/z 42 when compared to m/z 69. This last problem only interfered with one compound so it was something I could work around if needed. Just above optimum linear velocity at 15cm/sec I was getting an analyzer pressure of about 6*10^5 torr, at 0.2ml/min using a 5973 with oil diffusion pump. Using a high performance turbo model would probably help with the sensitivity loss due to the higher pressure.

One plus was being able to begin the analysis run at a higher oven temperature and still have good retention of the early peaks, which allows the oven to reach the beginning temp faster on cool down.

I did notice that with N2 you get a larger sensitivity change with carrier flow change than with either H2 or He, most likely because of the size of the N2 molecule and changes in mean free path are greater for a smaller change in flow. Honestly if I had a spare instrument I could use just for experimentation I would dig out my old jet separator I have stored away and see how that would work with N2 or even N2/H2 carrier/makeup combinations.

I am switching from focusing on Hydrogen carrier for MS to simply Helium alternatives, whatever they may be.

Oh and thanks everyone for the suggestions on what my interference is and for keeping the discussion alive

[gapeev]

Here is what I have on my mind regarding using N2 as carrier gas in GC-MS.

1. N2 will yield more than He, H2 ions in the ion source:
1.a because of its lower than He, H2 ionization energy
1.b larger diameter, also contributs to larger (~7x) ionization cross section

2. There will be more ion-molecule reactions involving nitrogen ions because:
2.a there are more nitrogen ions (e.g., N+, N2+, N3+ etc.)
2.b nitrogen ions are promiscuous

3. Space-charge will cause focusing and scattering issues, decreasing sensitivity

4. Have you ever replaced a filament? Why? Probably because metal evaporated, right?
Did you ever clean the ion source? Did you see some metallish-looking film? That used to be the filament, correct? The metal vapor condensed on the first surface that was not white-hot and did not go anywhere because it did not react with He or hydrogen ions.
Nitrogen ions will happily mate metal atoms begetting some other ions, they in turn will land onto quadrupole rods distorting field in the mass filter.

Does it make sense?

[Peter Apps]

Hi Alexei

Yes, it makes sense as an a priori argument, but does any of it make a enough difference in practise to make nitrogen impossible to use in GC-MS ?. And if the difficulties with helium supply continue, is there any prospect of benchtop MS designs changing to accommodate alternative carrier gasses ?

Peter

[James_Ball]

Here is what I have on my mind regarding using N2 as carrier gas in GC-MS.

1. N2 will yield more than He, H2 ions in the ion source:
1.a because of its lower than He, H2 ionization energy
1.b larger diameter, also contributs to larger (~7x) ionization cross section

2. There will be more ion-molecule reactions involving nitrogen ions because:
2.a there are more nitrogen ions (e.g., N+, N2+, N3+ etc.)
2.b nitrogen ions are promiscuous

3. Space-charge will cause focusing and scattering issues, decreasing sensitivity

4. Have you ever replaced a filament? Why? Probably because metal evaporated, right?
Did you ever clean the ion source? Did you see some metallish-looking film? That used to be the filament, correct? The metal vapor condensed on the first surface that was not white-hot and did not go anywhere because it did not react with He or hydrogen ions.
Nitrogen ions will happily mate metal atoms begetting some other ions, they in turn will land onto quadrupole rods distorting field in the mass filter.

Does it make sense?

For #1 if you reduce the amount of gas flowing into the ion source by 7x for N2 compared to the amount of He, how will that affect the difference in how the two gasses behave in the source comparatively?

For #2 what I have observed so far with the analysis of volatile organic compounds I am getting less reaction between carrier and analyte with N2 than with H2.

#3 I was prepared for because of the size difference of the molecule there is going to be more crowding of the molecules in the analyzer reducing the mean free path which reduces the sensitivity, but using lower flows seems to compensate. The electron multiplier needed only a few volts boost to bring abundances back on par with He.

For #4, how badly would the N-metal complex interact with quartz quadrupoles? I need to look again at the Agilent quads to see if the gold conductors are inside the quartz or on the surface, I can't remember exactly off hand. I do know that with the Agilent quads I have never had the problem of charging that is more common on units with metal quads when they become dirty.

Also what would be the advantage or disadvantage of using the Yttrium coated filaments? More or less interaction with the N2? They do operate at lower temperatures and you can reduce the current needed for ionization, but I have yet to try them to see what will happen( though I am planning to order a set soon).

[James_Ball]

Hi Alexei

Yes, it makes sense as an a priori argument, but does any of it make a enough difference in practise to make nitrogen impossible to use in GC-MS ?. And if the difficulties with helium supply continue, is there any prospect of benchtop MS designs changing to accommodate alternative carrier gasses ?

Peter

This is the ongoing problem that will need to be addressed. I have not been able to test one, but I have heard that the Leco Pegasus doesn't have as much trouble passing tunes as the Agilent because the source is more open and gives less interaction with H2. This may also be the case with N2, but I doubt anyone has ever tested it.

Of course what most people are trying this on so far are quad based instruments like the Agilent models, but has anyone out there tried either N2 or H2 on TOF or Ion Trap instruments? If you have I hope you will contribute your experiences with us. I must admit I would be like a kid on Christmas morning if I was in a lab with one of each type just sitting around to play with right now

[Peter Apps]

My guess would be that the instruments that work with harder vacuums (TOFS and quads with long ion paths) would be more robust to carrier gas type than those that have softer vacuums like ion traps. This assumes similar source geometries of course. I wonder what the prospects might be of exchangeable sources for different carriers - if going to CI can be done with a source swap, why not a different carrier gas ?

Peter

[James_Ball]

I have proposed to a few people a modification to the Agilent source where you would drill vent holes in it to allow the carrier to more easily escape. I was thinking opposite side from where the carrier enters, about 45degrees above and below the axis that the carrier is directed, which is also 45 degrees down from where the filaments are located. Similar in theory to a jet separator, the lighter carrier gas should exit these ports while the heavier analytes that do not get ionized on the first pass through the ion beams would impact on the opposite side of the source without escaping the ion chamber immediately.

If I had my own metal lathe I would probably be experimenting with this, but I can't make my own source bodies and I can't afford to start drilling into the ones we have given what the cost. I am beginning to wish we hadn't sold the 5971 I had been using as a test system.

[benhutcherson]

Digging up this old post I came across as I'm getting my 5973 nearly dialed in on H2, but still getting high 96 on BFB(which I understand is really difficult to fix because of 95 getting protonated).

James, did you get any further with N2?

As a little bit of a side note too-I'm not sure how common GC ion traps are any more(kind of a shame-they have their ups and downs but I always liked that the Saturns I've run in the past were so much less picky about vacuum and did appreciate the small full scan sensitivity boost on them) and I don't understand some of the "black magic" going on in them. Varian was very specific on the Saturn traps that you must use He, as whatever physic magic is happening depends on the presence of He in the trap for proper operation. The Finnigan LCQ Duo I ran briefly required a helium source and used a small capillary to "bleed" a tiny bit of He into the trap.

I could not really even get the Varian Saturn 2000 I tried to do much of anything with H2-it would just give me grass on a tune. Changing back to He but doing absolutely nothing else to the instrument would give me a beautiful tune. Vac readings were all still good(Varian didn't skimp on vacuum on these-they have a big diffusion pump and the normal config is backed by a DS102 rough pump, a bigger pump than the E2M2 that HP shipped with the 5971). They also expect you to get a lot more hands on, as PFTBA-or rather FC-43 in Varian literature-is controlled by a needle valve and there's a procedure for opening it the "correct" amount(I'm fuzzy as it's been a few years since I've used one). Dipping the RF coil is a routine operation too and IIRC something that you're supposed to do before a tune.

I don't know if there COULD be a way to make it work. I remember Varian tech support was always very helpful, but they were nearly a decade gone when I was working on this. A lot of the former Varian folks have now retired, and what knowledge is out there and still working in the field is scattered across Agilent, Bruker, and now Scion, and I'm not if anyone kept the Saturn alive after the buy-out.

[lmh]

Some physicist can correct me, in fact I hope they will. I can't remember the details for the Saturn GC-MS systems, but my understanding of those old traps was that they needed a fairly carefully-controlled pressure of damping-gas in the trap. Ions are held in the trap by an RF alternating-voltage field. The field is accelerating the ions; too much acceleration, they hit the electrodes and are lost. But too little, they drift off and get lost, so it's a balancing act. A damping gas means you can put in more voltage without the ions going into enormous orbits and hitting electrodes, but still with good retention in the trap. There should be the same problem in any quadrupole, but since ions tended to spend a long time in a trap compared to a normal quadrupole ion filter, it was more important that they have really good stability.
So a big change in the pressure or nature of the gas in the trap would require a corresponding, well-planned change in the electric field and possibly in the geometry of the trap.