filament problem or trap problem?

Discussions about GC-MS, LC-MS, LC-FTIR, and other "coupled" analytical techniques.

12 posts Page 1 of 1
I think this might be a trap problem. But it may also be related to this SISalloy+yttria coated filament I am using in the #2 position. This is a 6890N/5973inert system coupled with an Archon 4552 autosampler and Tekmar 3000 purge and trap unit using Vocarb 3000 trap.

I'm back worrying this bone again because after getting a very nice calibration on Jan 18, 2018; I found out that my certified 54 cpd 8260 standard did not actually have styrene in it. Its not been an issue for BTEX but I just got in a new set of standards to do another calibration. I figured I would do another tune on the SISalloy+yttria coated #2 filament and run a calibration. I used the Atune followed by ramping the ion focus and adjusting the 69/219 ratio to 60% which got my 50/69 ration right at 0.8%. I adjusted abundances to a 600K (I normally use a 500K) target for mass 69 and put on my calibration. I'm using 5uL of a 70 ppm internal standard mix in every 5mL of sample water. Inlet split ratio is 40:1. So, I end up with about 16 million counts on the Fluorobenzene 96 peak. With the 500K target, I was getting around 12-13 million counts.

So, now the problem. As concentration goes up above 160 ppb, in the calibration; everything from ethylbenzene onwards gets radically non linear as in peaks actually dropping in intensity to half or less of they values they should have. This is happening to the BFB surrogate but not the toluene-d8. The chlorobenzene-d5 seems to be stable but not the 1,4-dichlorobenzene-d4 which is also varying widely in intensity. The highest TVH(GRO) standard seems to cause loss of fluorobenzene dropping it from 16 million to 6 million counts [Edit]but not affecting should be... also affecting to a lesser extent[/Edit] the other surrogates or internal standards. My calibration curves are typically 10% or less RSD for a 20ppb to 480ppb calibration.

I have recently had several samples that had very high branched pentanes or benzene preceeding the fluorobenzene peak and fluorobenzene drops to half its normal value, while all the remaining internal standards and surrogates are within limits.

This sort of thing normally does not phaze my fluorobenzene recovery and so I have to ask, is this a trap problem with sample analytes tying up trap capacity so I get this weird behaviour. Or, is this some sort of quenching behaviour at the filament?

I've got two replacement traps on order but I wonder if anyone has actually seen this sort of thing happen. In 17 years of doing 8260 work, I've never seen an internal standard get wacked by analyte concentrations when they were only a few thousand ppb.
I looked at a sample from two weeks ago. My ethylbenzene had an area count of 15 million. It was flat on top across 3+scans. Calculated concentration was 5 times my high standard. I'd start by assuming detector saturation.
When a similar event happened to me, it was remedied by a source cleaning.
Do you use the 6mm draw out plate?
I just did a source cleaning and new filiments in mid January. I do use the 6 mm draw-out plate. What really seems odd is having a fluorobenzene or 4-bromofluorobenzene drop out so strongly when (e.g. TVH[GRO]) is high. This has never happened prior to installing this new filiment type. At least not in my normal operating/calibration range. I calibrate GRO from 1000 to 40,000 ppb and 8260 from 20 to 480 ppb. Peaks (BTEX/TVH) are typically linear to double the top calibration standard. Because diluted sample results tend to match the undiluted results. Also, even with BTEX components ranging up to say 4,000 ppb; therre was no quenching of internal standards or surrogates. I will try adding spike to the top of the trap and see if it still occurs; to implicate or rule out a trap problem.
You could try "burning" in the filament. Go to manual tune and set to scan, leave it for 2-3 hours.
I swapped in a different trap and duplicated the problem. A 40K ppb GRO standard suppresses fluorobenzene and all the other internal standards and surrogates to some extent. A 320 ppb 8260 standard reduces 4-BFB intensity in the same way (isopropylbenzene R.T. follows 0.05 min away) as before.

I swapped the original trap back in and while it was preconditioning, I got a good tune on Filiment #1 which is the standard rhenium 5973 filiment. I then ran a blank, a 40 ppb 8260, a 320 ppb 8260, and a 40K GRO. All of them came back with internal standards and surrogates (ISS) nicely tight and unaffected by analyte concentrations.

So its not a trap problem. It seems to be a problem with the linear response of this particular rhenium-yttrium alloy + yttrium oxide coated filiment. There may be something about my istrurment hardware that is causing this. So, don't hold this against the vendor.
Did you reduce the emission current on the Yttria coated filament? I have read before that you need less current because the Yittria filaments throw off a lot more electrons than the standards one does. If normal is 35uA they try it at 15-20uA emission current and see if it is more linear.

It seems you have quite high responses on those calibrators.
The past is there to guide us into the future, not to dwell in.
James_Ball wrote:
Did you reduce the emission current on the Yttria coated filament? I have read before that you need less current because the Yittria filaments throw off a lot more electrons than the standards one does. If normal is 35uA they try it at 15-20uA emission current and see if it is more linear.

I think there may be an emission current problem that is caused by using the ATUNE hat trick. I tried an Atune and ramped Emission and EIEnergy both sets of curves are sequential and fairly close together on the low side and the shape for 69, 219 and 502 all resemble eachother. But after doing the hat trick, the left side still looks pretty close but the right side gets a pronounced shoulder (suggesting a second peak) and as mass increases the peak and shoulder merge and the peak more than doubles in width.

This shape change is happening in a region where the source is not operating but the distinct change suggests that something funny is going on.
James_Ball wrote:
It seems you have quite high responses on those calibrators.

Yes, I try to keep things in the mid-range. 4 million counts is about as low as I have used on the fluorobenzene and 19-20 million counts is the most Iv'e used. When the instrument was new, it was more sensitive and we had to use a higher split ratio. Since then, we have replaced the detector (probably twice), swapped in a used sideboard, rebuilt the EPC, and refurbished the turbo pump twice. The tune tends to stay stable longer and require less EV, when I tune to around 500K on the 69 rather than 600K (default Atune) or 800K (default BFB tune).

For the last calibration using the SISalloy filament at 1388EV Fluorobenzene is ~12 million counts. The ethylbenzene 20 ppb is 5 million counts and the 480 ppb is 100 million counts.

This new calbration using the standard rhenium filiment is giving ~12 million counts for fluorobenzene at 1342EV. The 20 ppb standard is giving 4.8 million counts fo ethylbenzene.

I have a lot of samples to run after this new calibration. But I will try reducing the emission current to see what happens, next month when things slow down.
I should add, I did a hydrogen source/inlet cleaning bake out overnight last weekend just before trying the calibration. That might have something to do with it. But I don't see how.
We calibrate water standards from 0.5ppb to 200ppb for 8260 and our Ethylbenzene 20ppb standard is running about 220K counts. That us using the RTE integrator. Low standard is at 8000 counts and 200ppb is at 2.2M counts.

I found that keeping the sensitivity lower increases the signal to noise ratio and gives a larger linear range.

When I tune to BFB the PFTBA is giving about 600K counts for mass 69. I use the BFB Target tune and then tweak it with the variable entrance lens offsets.
The past is there to guide us into the future, not to dwell in.
Hmmm, I'm using the Chemstation Integrator. I'll have to research what the difference would be. My working MDL's are...
0.36 ppb for MTBE
0.67 ppb for benzene
These below are raised from determined values at request of clients; as being sufficient...
3 ppb for toluene
1 ppb for ethylbenzene
2 ppb for m/p xylene
1 ppb for o-xylene
3 ppb for napthalene

Edit: OK, I see that the RTE integrator gives counts for each EIC that are 10 times lower than the Chemstation integrator gives. Also, Chemstation lets one set timed integration events such as I need for doing GRO. One link said that the Chemstation integrator was only semiquantitative wheras the RTE integrator is quantitative. But there was only one link that said that so I certainly hope that is incorrect.
I believe 10x is about the difference between the two, so we are near the same sensitivity I think.

I always use RTE simply because I have been using it for 25 years since back when it was actually ran on an RTE based mini computer so I know what the settings do. I tried Chemstation integrator but never could make it do what I wanted. Also you can save setting for each compound if you want and add them in the last page on the analyte table so if you need a really small peak for one analyte you can set a low peak area threshold and then set a large area threshold for another peak that you are not interested in at such low concentrations. Or if one peak tails and another is quite sharp, you can set individual integration parameters so that each is detected well, and it isn't time based but analyte based.
The past is there to guide us into the future, not to dwell in.
12 posts Page 1 of 1

Who is online

In total there are 14 users online :: 0 registered, 0 hidden and 14 guests (based on users active over the past 5 minutes)
Most users ever online was 234 on Tue Feb 06, 2018 7:33 am

Users browsing this forum: No registered users and 14 guests