20 years running purge and trap GCMS and I am stumped

[James_Ball]

I just fired this email off to a service tech but wanted to post a copy here and get any thoughts.

I am still having a weird problem with the system you worked on Monday. Pretty sure it isn’t in the purge and trap itself, but it is certainly odd. Attached are two blank runs, the first I ran today(Thursday), it is run on the soil side of the Archon because I had good data from that to compare it with. The second if from a run on Tuesday, it is the blank after the calibration and first Cal check ran. 1,4-Dichlorobenzene goes from 220586 area counts to 45819 area counts. I ran that soil curve and five samples that were pretty clean, then the MS/MSD. I noticed that the Chlorobenzene-d5 and 1,4-Dichlorobenzene-d4 were low in the MSD but didn’t think much of it because sometimes it will happen on a soil. The areas for 1,4-DCB-d4 were all within 15% of each other on the curve from low to high, all looked good until that MSD. Last night I ran a Water curve with eight standards, the first standard had DCB-d4 at about 250000 area counts and the high standard had it at about 950000 area counts. A 50ppb check standard passed internal standard recovery versus the 50ppb in the curve which was about 500000 counts, but all the blanks and samples that ran last night had about 175000 area counts for DCB.

When this falls out it is like flipping a switch, it works great to that point then just drops out on the next sample and goes crazy. I am seeing similar things on the Encon Evolution running drinking waters. The other day I cleaned out the MORT tube with a pipe cleaner and methanol and ran a curve and the 1,4-Dichlorobenzene-d4 had very stable area counts throughout the curve and 1,2,4-Trichlorobenzene gave a calibration from 0.0-25.0ppb that was less than 10%RSD using average response factors. By the next day recovery for 1,2,4-TCB had fallen and when it was recalibrated the curve became quadratic at best with a loss in sensitivity at low concentrations. The calibration for 1,2,4TCB had responses as follows: 0.5ppb=724 area counts 1.0ppb=707, 5.0ppb=6171, 10ppb=32704 and 25ppb=229698. From very linear to very curved in a matter of maybe 20 standard and blank runs.

I am decent at purge and trap but this one has me stumped. My next step is to try to see if it can be something from reagents or water or carrier gas, since it seems to happen to more than one instrument. Would you or anyone else there have ever seen a problem like this before and if so know what causes it? Any ideas where to start looking, I don’t even think now it is from bad samples since it is even happening to the instrument reserved for drinking water samples.


I have been fighting this one for weeks and still no luck. It is almost like something builds up in the instruments that selectively traps halogenated benzenes. I have also noticed that 1,4-Dichlorobezene-d4 seems to fall out faster than 1,2-Dichlorobenzene-d4. If you take that calibration above for 1,2,4TCB up to 50,100 and 200ppb the area counts at each level increases exponentially, the resulting curve is almost vertical at the high end.

Anyone out there ever seen something like this?

[Bigbear]

Yes
I ran a 5970 system for 15+ years with no problem, then changed to a 5973! Worked ok for a while, then the bottom fell out.
In my case it was the water "activating" the MS source.
Do your areas come back after pumping overnite? After you clean the source?
We had the Agilent water kit ( #G7022A) installed last Oct. Things have been very good so far. Havn't needed to clean the source for 9 months, and still using a calibration from Feb.

More details on your system would help.

[James_Ball]

Well today I bypassed the moisture control traps(MORT) on both instruments and got the response for the late eluters back. The problem seems to be we are getting some kind of contamination on the MORT and it is happening fairly quickly.

Now we have to figure out if it is coming from the water or the carrier gas. A theory I have rolling around in my head is this:

If we are getting some hydrocarbons coming in from the carrier gas, and those deposit a little each time on the MORT which is held at 50c during purge, then during bake they pyrolyze from the 260c temp and inert atmosphere, it leads to a build up of "charcoal" that adsorbs a certain amount of organics on the next purge cycle. With the small amount of "charcoal" on the trap it retains some of the late eluters but the early gasses pass through just like when breakthrough occurs in a regular trap. Also once you fill all the active sites the rest of the late eluter molecules can pass through unimpeded which is why you get the drastically quadratic calibration curves.

Like I say, just a theory, but I am going to try hooking a tank directly to each instrument instead of using the building wide gas lines for the Helium and see if it slows or prevents the contamination. I will report back the findings once that is done later just for everyone to know.

[01310040231]

James:

I have seen this problem on our OI systems from time to time. I found the cause of the problem was the water management device. I rinse the sample path about once a month with methanol to avoid changing the water management every time. I think the acid vapor form the preserved water samples destroy the water management. Even with the clean up, they last about a year for us.

[Yama001]

OI water management is an increasingly expensive pain, particularly since they make it impossible to bypass. I like some features of their equipment, but that fitting is not one of them.

On that subject, OI claims their WMGT reduces the water transfer to the GC to 18 uL or less. Am I mistaken here, or does that still seems to be an awful lot of water?

[James_Ball]

I haven't worked with OI for a long time, and back then they didn't have water management yet. One of the best for removing water ever was the moisture control unit on the Tekmar LSC2000 that used sub freezing peltier cooling, only problem was it also froze out the last few compounds.

I also think that even on the EST the acid may be playing a part in this, but I would think it would cause more problems with breakdown of compounds like Pentachloroethane into Tetrachloroethylene than a loss of 1,2,4-Trichlorobenzene.

What I am seeing when I have the moisture trap inline is loss of the last four analytes (1,2,4-trichlorobenzene, 1,2,3-trichlorobenzene, Naphthalene, and to a lesser extent Hexachlorobutadiene.) Also when I begin to lose these in the calibrations I see a lot of carryover of these compounds into the next two or three desorbs, naphthalene will even carryover through almost the entire 12 hour run. Once I bypassed the moisture trap the carryover was down to less than 1% on the first desorb following a 25ppb standard and none in the rest of the run.

I am still leaning towards some type of hydrocarbon buildup on the trap, which when it is at room temp or slightly above begins to act as a chromatographic column, that is why the light molecules make it through to the trap but the heavies do not. As you saturate the column you get breakthrough which is why as you increase standard concentrations those last few compounds give exponentially higher responses. The trapping is also why there is so much carryover. Take that trap out of line and put a short piece of tubing to bypass it in the heated valve oven and you don't see the carryover or non-linear curve fits. I just can't yet figure out where the contamination is coming from that would cause such behavior. I wish I had a way to analyze the surface of one of those traps to see what is on there.

[Steve Reimer]

Is your water management heating properly?

[James_Ball]

Is your water management heating properly?

Everything seems to be working well, and it is happening on two different models on three instruments.

One of the units I have bypassed is starting to show a little of this same effect again, I think the tubing I used is actually being contaminated a little.

[01310040231]

We use 125:1 split with 0.5 minutes desorb. These settings, along with a clean water management has minimized water issues on our systems.

In regard to Naphthalene and heavy hydrocarbons; we have this issue on one of our systems. I have replaced all the sample path parts, but I still still a little carry over after CCV (10ppb). However, the second blank and all the runs after are clean. On the OI systems, there is a filter between the sparge tube and the 6 port valve to prevent foam entry into the sample path. I have replaced this part several times without seeing any improvement.

[James_Ball]

We use 125:1 split with 0.5 minutes desorb. These settings, along with a clean water management has minimized water issues on our systems.

In regard to Naphthalene and heavy hydrocarbons; we have this issue on one of our systems. I have replaced all the sample path parts, but I still still a little carry over after CCV (10ppb). However, the second blank and all the runs after are clean. On the OI systems, there is a filter between the sparge tube and the 6 port valve to prevent foam entry into the sample path. I have replaced this part several times without seeing any improvement.

On my waste water/soil instrument I am currently running 60:1 to 75:1 split with 1 minute desorb and it seems to be staying a little cleaner than the drinking water instrument that is running 40:1 split. Much more split on that one and I lose the sensitivity I need. When running the UCMR3 samples that require quantitation limits of 0.030ppb for for a couple compounds I sometimes need to run as low as 25:1 split while really cranking on the multiplier to see those well.


I did just notice something. What I normally use for my transfer line is the 0.53mmID Hydroguard MXT tubing from Restek, which is silcosteel coated tubing that has been deactivated to resist water, but I just looked at the spool I have here and the tag has it listed as MXT-1 0.53mmID 1um silcosteel tubing. The box had the proper label, but the tag on the tubing doesn't match. Of course if that was the problem then bypassing the moisture trap should not have an effect but it is just another thing to consider. I do remember a long time ago when people first started using the split injection port for volatiles that a service tech said he found that using something like this actually helped with water management since it somewhat focuses the analytes in a slug separate from the water being transferred. I may experiment with transfer line temps and see if that has any effects. (Also this is only installed on the waste water instrument so that would not be a factor on the drinking water one)

[Bigbear]

Please tell us more of the materials you use:
trap
column ( dimensions)
flow rate and split

[James_Ball]

Please tell us more of the materials you use:
trap
column ( dimensions)
flow rate and split

trap=VOCarb 3000
column= Rtx-VMS 40m, 0.18mmID 1um film
flow rate= 0.7ml/min He
split ratio= 30:1 up to 70:1 depending on method
GCMS= Agilent 5973 with Encon purge and trap Agilent 5975 with Encon Evolution purge and trap

5973 is used for waste water/soil analysis
5975 is used for drinking water by 524.2 and UCMR3 drinking water analysis by 524.3

Problem exists on both instruments. Currently the moisture traps tubes will be contaminated rather quickly, as in two analytical batches of either matrix. I saw it the last time running only DI water on the drinking water instrument.

Another thing I have discovered is that the GC analyst is seeing interferences in his blanks for pesticide and PCB analysis. When we run these blanks on the GCMS in the semivol section, we see small hits for what appears to be carboxylic acids. These could be coming from the new water system that was installed recently, but I can not attribute my problems to that since it happens when running one instrument on DI from our old system that doesn't show those interferences.

[Bigbear]

I still think you are putting too much water into your gc/ms. I use a 20M, 0.18 column flowing @ 0.5 cc/min and use a 150:1 split . I use a 1mm dia liner. I desorb for 3 min and, using the gas saver up the flow through my transfer line to 150 cc/min @ 1.5 min.

[James_Ball]

I still think you are putting too much water into your gc/ms. I use a 20M, 0.18 column flowing @ 0.5 cc/min and use a 150:1 split . I use a 1mm dia liner. I desorb for 3 min and, using the gas saver up the flow through my transfer line to 150 cc/min @ 1.5 min.

How much do you have to raise the EM Voltage to get low enough detection limits at a 150:1 split ratio?

I set my m/z 69 to 800,000-1,000,000 counts in manual tune and I can get down to about 0.2ppb at 50:1, sometimes need to drop to 40:1 to hit that. For the UCMR3 samples I have to hit 0.015ppb where I need to drop to 30:1 and add almost 300-400v to the EM volts which pushes my m/z 69 up to over 3,000,000 counts in tune.

[Bigbear]

I use the autotune voltages. I have the Agilent water management kit (G7022A) wich has a BFB_autotune that is running 1211V.
I only calibrate to 0.5ppb. I get about 170000 cts. for 5ppb fluorobenzene.