Chromatography Forum

Running Varian 3800 GC coupled to 4000 ion trap mass spec. Column: Rxi-5ms 30m x 0.25mm x 0.25um. After cleaning trap components, clipping both end of column & 24 hr bake-out, system was clean & dry.

Ran short sequence of clean samples in MeCl2. 1 uL splitless injections with 0.5 min 20 psi press pulse. Injector liner is clean Siltek-coated single gooseneck w/o packing. Injection speed 5 ul/sec. (My mistake!!)

Ended sequence in Stdby mode with oven at 200C, injector at 200C with 25:1 split and Xfer at 250C.

2 days later in Stdby mode, air-H2O check indicates system is dry, but filament emission check scanning m/z 30-150 shows MeCl2 spectrum. Ion time only ~2 msec & RICs ~16000.

How to proceed ridding system of MeCl2? I've not previously messed with carrier gas lines - and don't know if that's the right thing to do to start.

Thanks for suggestions.

Before you strat dismantling the carrier gas lines, you first need to know where the MeCl2 is coming from, which needs systematic isolation of the components of the system. NB NB I am assuming here that you have checked that there are no bits of septum or other crud in the inlet liner.

First; go from split to splitless, change the split flow by a factor of ten, and increase and decrease the septum purge flow (you do have the septum purge open, right ??!!). If any of these make a difference the MeCl2 is coming from the inlet, or upstream of it.

If changing inlet conditions makes no difference, take the column out of the MS transfer line and cap the connection, pump down. If the MeCl2 is still there it is stuck in the MS (and maybe coming from upstream also). If it goes away it is coming from or through the column.

To isolate the inlet form the upstream plumbing you have to do some serious pipe work, defer that until you have tried everything else.

For what it is worth, my money is on no flow thorugh the septum purge to lead crud in the inlet by a short head.

Peter

Thanks Peter.
Checked septum flow 1st thing this am: 3.24 ml/min, which is appropriate for this 1177 injector. Left split at 200 over 3-day week-end (usually my standby split is 25 to conserve helium) and find no MeCl2 having done no injections over week-end.

Air-water check is: Base peak 28, Base amt 140, fixed ion time with no AGC is 300 usec, RIC 925. This is also usual for my system.

Filament emission check scanning 45-135, fixed ion time is 25msec, base peak is 91 (not MeCl2), base amt 231, and RIC is ~12000.

There are no ions >m/z 500. Trap is clean.
But in full mass range I see 91, 207, 281, 307, 341, 355, 429 in diminishing intensity with increasing m/z. Hmmmm......column bleed contributing to the higher than desirable RIC in filament emission check.



--> I need to get to the bottom of this MeCl2 retention. So, will proceed with injecting while changing inlet conditions as you suggested and will send update.

Again, thanks!

hi
are you working in internal configuration or in external?
I have 4000 ms with 3800 and I work with large volume injection (mecl2) in external.
The background of m/z 49 and 84 is always high. I'm suspecting about external gas damping line....
Recently we have tried some tests in internal, with damping gas lines unused.
Can mecl2 accumulates into this line.?

Although I was sure my inlet and liner were clean, I decided to follow Peter's advice because sometimes the things we think we know just ain't so....

I cleaned inlet and installed brand new metal Siltek single goose liner and did 12 hr bake-out of trap.

In morning, a check of the filament emission gave ion time 25 msec and RIC ~3000 --> clean and free of MeCl2.

Made one injection WITHOUT pressure pulse, then immediately rechecked filament emission --> ion time still 25 msec and RIC still ~3000.

Then, using exactly the same method EXCEPT added pressure pulse lasting 30 sec (which is same as my splitless time) at 20 psi. --> MeCl2 was still overwhelming >3 hours later.

So.... let me back up and recount some conversations with Varian and Restek.

During installation, my Varian engineer used a single gooseneck liner - but the gooseneck was up, not down. This was bothersome because, as you know, the 3800 does not have a removable gold seal at bottom (it is a closed inlet and must be swabbed out from the top). So, if the liner is upside down, then any septa particles and garbage just falls to the bottom of the inlet which cannot be disassembled and emptied (like the Agilent design).

So, I asked Varian trainer who said it was not a gosseneck, it was "a needle guide and should be up."

After coming back to my Lab, I called Restek who have never heard of the gooseneck being called a needle guide.

I put the goose down to keep crud out of the bottom of my inlet.

I ASSUMED that the increased pressure would focus the injection on the head of the column. What appears to be happening is that it backs into the damping gas line. And stays a long time....

This is my current guess: to meet higher sensitivity specs, the installation engineer inverted the goose and used a press pulse to eek out a bit more S/N.

So... yesterday I installed a brand new glass Siltek double goose liner and overnight bake-out. This morning filament emission check gives ion time 25msec and RIC ~3000 --> clean and dry. I'll repeat comparison using no press pulse and with press pulse having goose at top and bottom, then post again.

Just as an fyi, here are my conditions:
Injector: 275C; Initial - split 50, 0.01 min - split off, 0.5 min - split 200, 3 min - split 50, press pulse 20 psi for 0.5 min; constant flow 1.0 ml/min He;
1 uL injection volume; injection speed 0.1 ul/sec

Oven: 100C for 0.5 min, ramp 10C/min to 250C.

MS: internal mode, 10 uA ionization current, 10000 target TIC

Column: Rxi-1ms, 30m x 0.25mm x 0.25 um

This morning i have performed some tests. My 3800 has two injector, 1177 and 1079 (ptv).
Usually I use the second one, for 20uL injection.
Now the system is ok: there is no noticeable background of mecl2.
This morning I have tried to couple my column into 1177, used in past for splitless injection and..surprise!
Mecl2 rise at incredible level!
Just disconnect and reconnect to 1079, bake 1 hour, and now the system is ok again.
So, i think there is a problem with 1177 split line and EFC control.
I told my opinion to service engineer, and next week we will try change "pencil filter" into the efc.
I'll let you know news.
Thanks for your "report"
bye

I would work backwards from the injector and assume you have backflash products upstream in your injector. I would clean the lines from your EPC to the injection port thoroughly with nitric acid in methanol and rinse thoroughly with methanol. I would also clean the area under and around your septum. I believe the methylene chloride is being retained and slowly released by deposits upstream in your injector.

Of course your trap in the split liner could be blocked but you should have been able to determine that before now (I did not look at all the previous posts in this thread.)

good luck,

Rod

hi rod
the strange fact is that 1177 has 2 months of life, and only few injection for test.
Anyway, i'll investigate...
bye

There is a general problem with sharp pressure pulses on any GC system - as soon as the EPC/EFC cuts the pressure there is back flow from the high pressure in the inlet towards the point where split and inlet flows diverge. This contaminates tubing, connections and valves upstream of the inlet.

If you have to use a pressure pulse (and IMHO they get put into applications notes just because EFC makes them possible) you need to taper the pressure off slowly so that the EFC keeps feeding gas to the inlet. Opening the split before the end of the pulse will increase the velocity of flow into and through (instead of past) the inlet, and reduce problems with back flow.

For reasons lost in the mists of history Varian liners have always had the constriction at the top, HP and now Agilent at the bottom. WIth a single constriction in the liner you can choose which undesirable part of the inlet you sample contacts; the hot bare metal at the bottom, or the hot bare metal at the top. For splitless injections you need a liner with two constrictions. Why do they call them goosenecks ?

Peter

Peter,

Good post and good points.

Concerning the injector with few injections:

There may be contamination left over from manufacturing that just happened to be there and to contribute to the problem.

Dirt happens sometimes to the best.

Clean it up and it should go away.

wishing you the best,

Rod

Thanks to everyone who posted replies. It is much appreciated!

As an update, the MeCl2 remains still an hour later using the double gooseneck. (Ion times in filament check remain only 2.8 msec even an hour after a single injection with press pulse. They were 25 msec immediately after a run with no press pulse.) --> double gooseneck is not sufficient to prevent the backflash.

So.... Peter, as soon as it's cleared out I will try again - taking your advice about opening the split while the press is high, then end the press pulse. How long should the split be open while press is high before ending the pulse?

I'm beginning to have serious reservations about the purported advantages of the press pulse. How do others feel about it? Is it really that effective loading more onto the column? Really advantageous if using only a 1 uL injection?

Also, is it true that splitless injections require a restriction at top and bottom? Geez - look at any catalog for liners... If so, then please explain why as I really don't know why.

I measured the septum purge (3.2 ml/min). Haven't measured split flow, but will do it promptly.

If I can possibly eek by without purging the lines with HNO3/MeOH, I'm going to try. The test will be if I get back to low ions times and no discernable backgrd ions, yes?

Another ting to look out for - make sure that you have a wide bore liner for splitless injections. With a narrow bore liner the volume of solvent vapour is too big for the liner to contain, and it gets int oplaces that you do not want.

Before rinsing lines with anything I would try gently heating them with a hot air gun with gas flowing through.

All the manufacturers have data that show increases in detectability with pressure pulsed injections, but the timing and pressure in the pulse are additional sources of variability - especially in relation to when the split opens, so what you gain in detectability you might lose in repeatability. To increase detectability, large volume injections to a PTV are a far better way to go because they avoid the problems caused by flash evaporation of volatile solvents.

Peter