5975C reproducibility problem

[Daniel_Lampiasi]

We are using a 7890A coupled with a 5975C mass spec for volatile analisys. I have noticed that after many injections our signal loses a lot of intensity. It´s not reproducible, but as we are using ISTD compounds the software can correct for these deviations. The problem is that after too many injections the signal is really affected. I will post a link to the images so you can see the lost i´m describing.

The signal after the equipment stays idle for some hours: http://imgur.com/v3Lju9M,Em0FRgS,Ti6FxLD#0
The signal after some injections: http://imgur.com/v3Lju9M,Em0FRgS,Ti6FxLD#1
The signal after many injections: http://imgur.com/v3Lju9M,Em0FRgS,Ti6FxLD#2 (On this image you can see that sometimes the sinal is even null; not even the gas peak is detected)

I always thought that this was indicating the need for cleaning the ion source, but after i cleaned last time, i´m noticing these lost in signal much faster than before. I even tried to replace the filament with a brand new, but after 20 or 30 injection the problem was visible again.

We are using CombiPal for headspace analisys.

Thanks in advance!

Daniel

[thohry]

The sample matrix may be not clean
Have you checked the inlet liner ?

[Peter Apps]

What are your headspace conditions, and what are you analysing ?. High incubation temperatures with aqueous samples will put a lot of water vapour onto the column and into the MS, which might degrade signal:noise.

Peter

[Daniel_Lampiasi]

The sample matrix may be not clean
Have you checked the inlet liner ?

But I always inject clean water (mostly). And the problem always occur after a great amount of injections.

What are your headspace conditions, and what are you analysing ?. High incubation temperatures with aqueous samples will put a lot of water vapour onto the column and into the MS, which might degrade signal:noise.

Peter

It seems to be a reasonable explanation.. I analyse volatile compouns in water (most of them are for human consumption). I use 20 mL vials containing 15 mL of water with 4g of sodium chloride (method 8260C). Here are the details:

Injection Volume: 1000.00 ul
Syringe Size: 2.5ml-HS
Incubation Temperature (°C): 85
Incubation Time (s): 600
Syringe Temperature (°C): 85
Agitator Speed (rpm): 500
Fill Speed (μl/s): 250
Fill Strokes (): 0
Pullup Delay (ms): 1000
Injection Speed (μl/s): 1000
Pre Inject Delay (ms): 500
Post Inject Delay (ms): 500
Flush Time (s): 360
GC Runtime (s): 1980

Front SS Inlet He
Mode Split
Heater On 160 °C
Pressure On 23.193 psi
Total Flow On 19.5 mL/min
Septum Purge Flow On 3 mL/min
Gas Saver On 20 mL/min After 2 min
Split Ratio 10 :1
Split Flow 15 mL/min

I will perform a water/air check before and after a long sequence to confirm.

Thanks thohry and Peter for the insight!

[James_Ball]

I run mostly purge and trap for volatiles by 8260, 624 and 524. If I make very long runs, such as when I start the instrument on Monday and it runs most of the week, sometimes 200-300 samples non stop, I will see a similar loss in sensitivity. My internal standard recovery can drop by 50% which triggers a recalibration. If I let the instrument sit over a long weekend, then most of the sensitivity returns. The water vapor builds up in the pump oil mostly and causes the loss in sensitivity.

I do have the inline moisture traps with molecular sieve placed right on the inlet to the rough pump( the same ones they used to put on the 5971 ) and that helps a lot.

[mdevay]

Couple things that may help:

Lower your EMV the next time you calibrate. Overloading the HED/EM can lead to decreased sensitivity and is most pronounced when comparing your CCV/LFB samples with blanks or clean samples. CCVs will have enhanced IS/SS responses, blanks and clean samples will have lower responses. This leads to a characteristic zig-zag shaped response chart when you run alternating CCVs and blanks, with an overall downward trend. Lowering EMV can mitigate this somewhat.

Replace your filament/switch filaments. As the filament heats up, it starts to deform. For long runs, the longer it stays hot, the more it deforms. After many runs, the amount of time it takes for the filament to deform to the point where it's bad for analysis becomes faster and faster. A deformed filament doesn't align properly with the aperture in the source, leading to fewer ions being produced.

Inlet maintenance. Change your liner, gold seal, ferrules, and trim a good chunk off the front of the column. The inlet is the most likely part of the instrument to have active sites. I'm assuming that you run a CCV before you run any samples - this tends to swamp active sites so that its effect is minimized. As the run continues, those actives sites/adsorption sites start to adsorb more and more of your IS/SS assuming you are running fairly clean samples compared to your CCV. Yeah, its headspace analysis, and yeah, you wouldn't expect your inlet to get gunked up too quickly, but you'd be surprised, especially if you have the least little bit of oxygen or oxidizing compounds in your lines.

It typically takes a long time for a source to get dirty to the point of detriment for VOC analyses (unless you run with a leak in the detector). We clean the source apprx once every 6 months, switch or replace filaments after every three months, and do inlet maintenance every two months. Recalibrations tend to coincide with inlet maintenance, although they are sometimes more frequent, particularly with older instruments whose lenses are starting to get a little foggy. We're typically running 12 hours a day, 5 days a week on most instruments. Even at 6 months of relatively heavy usage, the source tends to be fairly clean - most of the time only a slight golden hue can be observed, sometimes with a small black spot on the ion focus.