Phthalates with Varian ion trap

[Ron]

Did the problem start before or after the trap was replaced? If it started after, then I would try a different trap to verify the new one is OK.

[rp18]

It started on the first day we got the (new) instrument...

[Peter Apps]

Measure the flows with an independent flow meter, not the onboard GC hardware. That way you troubleshoot possible malfunctions in the GC's measurement of flows.

Peter

[rp18]

I did already - no malfunction. Both septum purge flow and split flow are OK. It looks like the solvent migrates somewhere during the splitless minute and bleeds back during the run.



Anyone here with a 1177 injector who can post a splitless chromatogram with some solvent injected?

[WK]

RP18,
One thing I would say about my 3800 with 1079 is that I would never touch the pneumatics - it is impossible to figure out where anything is there (cf PE Autosystem). It is also interesting that the engineer's splitless injection for his test mix is:
Initial ON 20
0.00 OFF OFF
0.70 ON 50

I don't believe this is same as in manual. Also I think that even if the split is OFF then flow continues from the split vent - design I guess.

Regards
WK

[rp18]

The only thing that makes me wonder about the 1177 is the magnetic T-valve which is connected to the septum purge outlet AND central gas inlet. The splitless program you posted looks OK, 0.7 minutes splitless is a normal time.
If the split is OFF the flow is directed through a bypass, so there's always a flow at the split outlet.

[ChromatographyDude]

Here is a trick I've done in the past with my 1079 injector. Remove the analytical column, septum it off. Follow back the gas lines from the 1177 split where it connects into the circular flow connector disc. Loosen the screw holding the disc, and completely pull out the split line. Using a syringe and tubing, flush some ethanol or acetone through the split line and it will spray into the injector. Do this several times. Then, take a can of compressed air, put the tip against the gas line, and blow into it. This should help remove solvent from the line and remove any particle debris. Reconnect the line, use a heat gun if you have one to heat it up, then bakeout the injector. Worth a shot.

Do you only have an 1177 injector on the system? No 1079? I really think you have an injector related issue as I haven't seen this type of sensitivity issue on the Varian trap. How hot is your initial GC start temp? For toluene, I'm guessing it's close to 100? And your injector and transfer line are both at probably 250 or greater? What happens if you try this same injection using a pressure pulse?

Also, just wondering what type of liner you are using? As I'm sure you already know, the taper on the 1177 splitless liner should go up in the injector, not at the bottom like on Agilent injectors.

Good luck!!

[rp18]

Thanks CD, maybe I'll give it a try (warranty issues?). The 1177 has two split lines coming from the body, so if one is plugged, the solvent may choose the other way.

Do you only have an 1177 injector on the system? No 1079?

Only 1177.

I really think you have an injector related issue as I haven't seen this type of sensitivity issue on the Varian trap. How hot is your initial GC start temp? For toluene, I'm guessing it's close to 100?

Yes, 90 °C. It's not only toluene, it's more or less any solvent (hexane, cyclohexane, ...).

And your injector and transfer line are both at probably 250 or greater?

Yes, 280 °C

What happens if you try this same injection using a pressure pulse?

Same

Also, just wondering what type of liner you are using? As I'm sure you already know, the taper on the 1177 splitless liner should go up in the injector, not at the bottom like on Agilent injectors.

Tried several, correctly installed (the service people also always changed them...)

[rp18]

Final update:
according to Varian it is not possible to use the 431/220 for this purpose (PAH below 1 ppm, phthalates below 1ppm). They tried it in the Varian application lab with the same results - low sensitivity, bad results in splitless mode, long tailing solvent peak. They recommend another GCMS.

[AICMM]

rp18,

Couple of comments about your chromatogram, maybe too late. First, how much solvent are you shooting on what id of column? Second, it looks to me like your split valve never kicks on. The reason I say that is normally when it does, you see a sharp drop in the solvent peak back to baseline. You never achieve that which makes me think you never go split and just bleed solvent off during the whole run. Based on that, what are the spectral lines on the baseline out at 8 or 10 minutes? What does just the solvent peak look like (forget the target analytes for a bit) if you run the injection as split? Nice and sharp or tailing?

Best regards.

[rp18]

Thanks AICMM,

- volume: tried 0,5 to 1 ul, even 0,1 to 2 ul
- column ID: 0,25 mm (MS standard column)
- I know how a normal splitless injection looks like
- split valve IS working, opens 1 min after splitless
- split injections works better (peak shape, solvent peak), except for even lower sensitivity

I really have years of experience with HP/Agilent and Shimadzu machines and I just can't believe the bad sensitivity of that new Varian GCMS...

[Peter Apps]

This gets odder and odder .

How do you know that the split valve opens ? Unless the cut off of the solvent peak is during the filament off time at the begining of the chromatogram I agree with AICMM that this looks like a permanently splitless injection. If a split injection gives a sharp solvent peak it implies that the solvent tail problem is with the inlet, not the MS.

Is the problem here with the lower limit of detection for the phthallates being too high, or the signal for the solvent being too high, giving a poor signal:noise even though the signal is reasonably high ? In other words, if you could put the phthallate peaks on a flat baseline, would they be big enough ?

Why would the apparent deficiencies on the MS be specific to phthallates and PCBs ?, they don't have much in common beyond an aromatic ring, and plety of other molecules have that, including your solvent toluene, for which the signal is higher than you want.

Peter

[rp18]

How do you know that the split valve opens ?

- without opening after 1 min it looks even worse (I tried).
- the EPC tells me it's open
- same problem on different GC's of this type (even the Varian people tried - they should know)

Unless the cut off of the solvent peak is during the filament off time at the begining of the chromatogram I agree with AICMM that this looks like a permanently splitless injection. If a split injection gives a sharp solvent peak it implies that the solvent tail problem is with the inlet, not the MS

- that's what I supposed the very first time I saw the chromatogram (see older posts)

Is the problem here with the lower limit of detection for the phthallates being too high, or the signal for the solvent being too high, giving a poor signal:noise even though the signal is reasonably high ? In other words, if you could put the phthallate peaks on a flat baseline, would they be big enough ?

- No. Flat baseline = split injection = too low sensitivity. But no tailing solvent peak.

Why would the apparent deficiencies on the MS be specific to phthallates and PCBs ?, they don't have much in common beyond an aromatic ring, and plety of other molecules have that, including your solvent toluene, for which the signal is higher than you want.

(PAH, not PCB)
So far the problem seems to be the injector, which is simply not designed or doesn't work for splitless injection with toluene solvent. And yes, that's weird, because it's really a standard application for GCMS.

[Peter Apps]

And what is it about toluene specifically that causes the problem ? - which in post 19 you say also occurs with hexane and cyclohexane "more or less any solvent".

Let me word more carefully: if you could put the peaks that you get with a splitless injection (as in the chromatogram that you posted) onto a flat baseline, would they be big enough ? You already told us that making the baseline flat by doing a split injection made the analyte peaks too small.

Earlier you said that split and septum purge flows are OK, but I don't recall seing what they are in ml/min (and the topic review doesn't scroll back that far so I can't check).

This is a brand new instrument that appears to have a problem with its inlet, in your place I would be pushing for the inlet and gas controls to be swapped out.

Peter

[rp18]

- the effect is more dramatic with toluene
- the peaks are way too small
- split flow: tried from 5 to 100 ml/min
- septum purge: tried from 1 to 5 ml/min
- the engine always PASSES the test with octafluornaphthalene

Again: even the specialists at Varian (with THEIR instruments) couldn't get the sensitivity needed. My 20 year old HP engine is decades better.

So: ANYONE here with a Varian 431/220 who could give it a try?
I still can't believe that a new, modern GC can't do more or less standard analysis.