Newby 6890 GC problems

[laurimc]

I was given a GC that hasn't been used in 5+ years and told to figure out how to analyze HAAs on it... I have experience with just about every other instrument, but not GC! I'm having trouble with what I think is likely a mixture of problems...

My setup:
Agilent 6890 w/o autosampler
Running method E552.3 for HAAs using MtBE
Detector: Micro-ECD at 290C, make-up with N2 at 20mL/min
Column: DB-5.625 (30m x 250um, 0.25um film), He at 32cm/sec
Program: 40C for 10min, 2.5C/min to 65C, then 20C/min to 205C, post run hold 210C for 7min.
Injector: Split/Splitless in Splitless mode, 210C, currently have old 4mm single taper liner (ordered new 2mm straight to install next week per method recommendation) 45sec hold before purge at 30mL/min, 1uL hot injection

Maintenance so far:
Former employee left the injector, oven and detector on with no gas flow, so there were issues to start with. I did a bake-off of the uECD which brought baseline down to ~150Hz. I figure my column is probably shot from being heated up for months/years without any gas flow, so I expect I will need to change that out (have a brand new one in the cabinet... woohoo!). When doing the bake-off I noticed the 1/4" ferrule to the detector was pretty sad and replaced that afterward. I also found that the ferrules being used on the column were meant for 500um, not 250um, so there's probably some leaks.

Symptoms:
The baseline looks great when I do a run with no injection. It also looks good with an air injection, though the inlet seems to have trouble maintaining pressure during injection, even with new septum. Running a blank (straight MtBE w/o any extraction) is where it gets weird. No solvent peak at the front but I usually get a peak around 13-15min, sometimes only 600Hz, but often 9000Hz. Running a standard (20ug/L of multiple HAAs in MtBE) gives a few really broad and short peaks (50Hz or less) and that same big peak at 13-15min.

Plans:
Re-build the inlet with new liner and gold-plated seal. I think I may have a leak in there...
Replace column (with correct ferrules!). New ferrules could correct that possible leak...

I would love input from those GC veterans out there! Am I on the right track?

THANK YOU!

[James_Ball]

Definitely get the correct ferrules on there to seal it up, and most likely a new column would help also.

For capillary columns it is not so good to inject air, as oxygen degrades the column.

Maybe Jessie will chime in here later as he runs the HAAs here in the lab, he has a good setup for those, but you really need to get the instrument working well before fine tuning the method. I am trying to remember if that makeup flow is supposed to be 20ml/min or 60ml/min, been a while since I used the ECDs. Check the expansion volume of MTBE when injected to make sure you are not overloading the 2mm injection port liner. You will be ok at 210c as long as your head pressure stays above 14psi. Look on the Agilent site and see if you can find a download for their Pressure and Flow Calculator, it is a handy little tool for figuring if your inlet volume and flows are set correctly.

The mismatched ferrules are probably what is causing the pressure problems at injection, make sure those are good and tight with the new size and that the o-ring around the inlet liner is in good shape.

Sounds like you are headed in the right direction.

[Steve Reimer]

What you are seeing isn't surprising for a column that has been exposed to air for 5 years.

[laurimc]

Thanks everyone! I'll let you know how it goes when I receive my new supplies next week.

[laurimc]

OK, I'm going to need more help!

I replaced the column, inlet liner (I went with 4mm straight to accommodate the MtBE volume as suggested), and gold seal. I injected 2uL of methylene chloride headspace and got a nice straight peak with very minimal tailing, which tells me I'm leak free. Conditioned the column and did a blank run with no injection and it was beautiful - nice and flat, with a little rise at the end due to big temperature ramp.

I thought all was good, but when I inject 1uL MtBE it looks the exact same as before all of the maintenance! No solvent peak, or maybe just a hideous and small one, then a big peak at ~14min. Here's what it looks like:


Sorry for poor quality... had to take a picture with my cell.

Help!

[Peter Apps]

The shape of a peak tells you nothing about whether you have leaks or not - in fact a leak of the right size in the right place can take the tail off a peak - exactly like a split injection. So check you system with a leak seeker.

The ECD should not see your solvent, so getting a small solvent peak is exactly what you should expect. The peak at 14 min could be an electron capturing impurity in the solvent (check whether the manufacturers put a stabilizer in it) or a contaminant.

Peter

[asm]

MTBE is often stabilized with Butylhydroxytoluene.

[AICMM]

laurimc,

First, along the lines of Peter I thought perhaps a phthalate but you are sitting pretty cold for a long time (40 for 10 then 10 more to 65) so that's out.

My suggestion at this point is to shoot a stock of the HAA's. Remember the HAA's will not respond as well as something like the pesticides. If higher concentration yields peaks more or less where they should be, next step would be to look at sensitivity. Have to have chromatography before sensitivity.

Best regards,

AICMM

[laurimc]

My suggestion at this point is to shoot a stock of the HAA's. Remember the HAA's will not respond as well as something like the pesticides. If higher concentration yields peaks more or less where they should be, next step would be to look at sensitivity. Have to have chromatography before sensitivity.

Thanks, AICMM. I tried injecting HAAs at 20ug/L today (should be the middle of my future curve) and got absolutely no response - curve looked the same as my blank. So I obviously have a leak somewhere in the inlet. I just took it apart and put it back together with a fresh cut on the column, so I'll see what it looks like tomorrow.

I also noticed that when I inject 1uL my inlet pressure shoots up and then right back down. 1uL MtBE with the inlet I have should only be ~50% capacity, so I'm not sure what that means... I switched to 0.5uL and that worked fine, but I'm thinking this is an indicator of something wrong with my inlet. Again, I'll see how it looks tomorrow when it's warmed up again.

Thanks!

[Peter Apps]

Hi Lauri

You really really really !!! need to check for leaks with a leak seeker - just pulling things apart and putting them back together is just as likely to create leaks as to make them go away.

Peter

[James_Ball]

My suggestion at this point is to shoot a stock of the HAA's. Remember the HAA's will not respond as well as something like the pesticides. If higher concentration yields peaks more or less where they should be, next step would be to look at sensitivity. Have to have chromatography before sensitivity.

Thanks, AICMM. I tried injecting HAAs at 20ug/L today (should be the middle of my future curve) and got absolutely no response - curve looked the same as my blank. So I obviously have a leak somewhere in the inlet. I just took it apart and put it back together with a fresh cut on the column, so I'll see what it looks like tomorrow.

I also noticed that when I inject 1uL my inlet pressure shoots up and then right back down. 1uL MtBE with the inlet I have should only be ~50% capacity, so I'm not sure what that means... I switched to 0.5uL and that worked fine, but I'm thinking this is an indicator of something wrong with my inlet. Again, I'll see how it looks tomorrow when it's warmed up again.

Thanks!

Just to be sure, were these the methylated HAA's?

A pulse in the inlet pressure is not unheard of at injection, but how large was the pulse?

Do you have some methanol or hexane you can inject to see if you get the same peak at ~14 minutes? That will tell you if maybe your solvent is knocking something off the injection port or column as the solvent passes through on injection. Another thing to consider is taking off the split line(the 1/8 inch copper one with the brass swagelok fitting on the injection port) and cleaning or replacing with more copper tubing. If those get gunked up all kinds of weird things can happen. Come to think of it, it may not be copper on the 5890 it may just be 1/8" at the port reducing down to 1/16" stainless, but either way if you can remove and clean that line it may help.

[laurimc]

One victory, one big laugh at myself.

Victory: I found the leak at the inlet and got it taken care of and the pulse in pressure is gone. I wish I had a leak detector, but I had to do it using the EPC... which is cool and worked well.

Laughing at myself: My standard isn't methylated. Of course my chromatogram looks the same as the blank!

As for the ~14 minutes peak, I injected hexane a few days ago and it didn't appear, so it's definitely in my MtBE. Guess I'll have to get a fresh stock of that!

Thanks for your help everyone. We'll see how this goes once I do the extraction!

[laurimc]

I've got peaks! They are pretty ugly peaks, but I'm happy to see them there. I'm going to wait to tweak things until I have my new MtBE, but any suggestions on getting my peaks to be less broad?

Thanks!

[Steve Reimer]

Once methylated your peaks shouldn't be broad. If you still see blobs something is still wrong. I usually advise on this method to check your extract prior to analysis with pH paper. If it indicates acid you haven't fully neutralized the extract and injecting it will destroy your column. It's fast and easy and much cheaper than a new column (or in my case two columns -$900).
How does the trichloropropane peak look?

[Bigbear]

You may need to up your make up gas flow. I use 30 cc/min with a 0.32 column flowing at about 2 cc/min.