New to Chromatography - Manual Injection Technique

[MichaelStuart]

Thanks a bunch Peter.

*The liner is packed with the wool at Restek. They appear to be consistant with the amount, but the wool is also a bit loose in there.

** I am getting ready to do a service on the instrument. Septa, trim column, inspect liners, and inspect/clean split.. I'll be sure to compare liners while i'm in there.

*** I will also try spiking with an internal standard. As I am measuring 10ml of solvent, it seems to be easier to get it spot-on than when dealing with 1ul of aliquot.

(Silly question alert:) When measuring my solvent with the 10ml pipette, the solvent adheres to the wall of the pipette a bit creating a parabolic level. Do I measure from the top of the parabola against the graduations, or the bottom? I wish I paid closer attention in High School chemistry.

I also would like to to thank you for sharing your manual injection technique in the other thread. I found that giving the needle a little more time in the injection port gave it the time to vaporize the entire aliquot, both the contents of the needle, and the barrel. I believe this has helped a great deal with consistency. Thanks for the tip.

I do hope I have an opportunity to reciprocate some time. If you ever have a question regarding medical cannabis, I'm your guy

Mike

[Peter Apps]

** I am getting ready to do a service on the instrument. Septa, trim column, inspect liners, and inspect/clean split.. I'll be sure to compare liners while i'm in there. when you disconnect the column form the nlet have a look at how far they protrude through the nut and ferrule into the inlet

(Silly question alert:) When measuring my solvent with the 10ml pipette, the solvent adheres to the wall of the pipette a bit creating a parabolic level. Do I measure from the top of the parabola against the graduations, or the bottom? I wish I paid closer attention in High School chemistry. Go for the bottom of the parabola

I also would like to to thank you for sharing your manual injection technique in the other thread. I found that giving the needle a little more time in the injection port gave it the time to vaporize the entire aliquot, both the contents of the needle, and the barrel. I believe this has helped a great deal with consistency. Thanks for the tip. No problem, glad I could help

Peter

[MichaelStuart]

Hi Guys,

Monday and Tuesday, I had good reproducibility all day. Every time I shot my reference, as I did intermittently throughout the day to see how things were going, I was within a few percent of my known standard. Yesterday, back to all over the board. I have three analytes in my reference, CBD,THC,CBN.. 1000/1000/1000 (ug/ml).. After a few runs that provided unexpected results, and poor reproducibility, I shot the reference standard and it returned the results in the 1750/1780/1760 (ug/ml).

I am using air as my carrier gas. I use a compressor for the air, and notice that my regulator that came with the compressor is not consistent pressure. The first clue was when I came in and went to ignite my FID, and it wouldn't light. Hydrogen pressure was great, air tank had ample pressure. I then noticed the regulator was at zero (from the 40psi that it was set). I reset it to 40, did a couple runs.. looked down and it was at 60psi. I turned it down to 40 again. Throughout the day I monitored it and it seemed to fluctuate between 35 and 45.

I am suspecting that this could be contributing to my issue, and intend to get a better regulator. Any thoughts on the subject would be appreciated.

Also, what should I expect my reproducibility to be if I am extremely diligent in regards to sample introduction. I am very aware how plunger throw, paralaxing errors, bubbles, sample temp, etc. and go to great lengths to be consistent with my introductions. I have read that variability of 1% between shoots and 2% between chromatographers can be reduced to .1%, and .2% respectively with good consistent injections. I can't imagine that my sample introduction is so whack that I'm starting to feel good if I can shoot twice within 10%.

Warm regards,

Michael Stuart

[chromatographer1]

" I have read that variability of 1% between shoots and 2% between chromatographers can be reduced to .1%, and .2% respectively with good consistent injections. "

That is awful good for manual injections with a 10 mcL syringe.

1% reproducibility with manual injections is great.

With smaller syringes this can be improved. Even the best splitter will have 1-2% variability if you are doing split injections.

good luck,

Rod

[MichaelStuart]

Hi Rod,

Thanks for chiming in. I would throw a party if I could get single digit variability. I do have a split injection port. One thing I noticed is that when I shoot my standard which is 1000/1000/1000 and calibrate.. I can shoot it again and it can be 950/1030/1060 which seems to indicate that if it is seeing the first analyte as lower (5%) and the other two as higher (6%) that either there is something wrong with the machine, my method, or some variable is coming in to play during the run.

I think I'll focus on the regulator for now and see if that straightens things out.

Thanks again, as I really do appreciate all insights at this stage of my noobiness.

Mike

[Don_Hilton]

Take a look at integration marks and be sure the peaks are being integrated consistantly.

[MichaelStuart]

Don, Thank you for the suggestion. I will learn about it and, and do just that.

In the mean time, I replaced the compressor with the suspect faulty regulator that I was using for air. The second compressor I got was the same Craftsman 7 gal. 125 psi as the first and the regulator was spraying air out from the knob. Complete garbage. Replaced it with a 20 gal, different manufacturer and store and reconnected my supply plumbing. I just shot my primer, and my reference standard twice. I calibrated the GC to the first shot. The second hit it on the button. Well, sort of..

My ESTD is comprised of 3 analytes. Where I calibrated to 1000/1000/1000, the following shot was 1009/971/940. Is this normal, or indicative of an issue with my machine, method, or anything else? Granted, I feel like throwing a party getting results that are within single digit deviation, however I am left needing to reconcile how one could measure "deviation of 1% in reproducibility between shoots, and 2% between different chromatographers" if the inherent deviation in the chromatograph is at least 6%. I am open to the possibility that I am not grasping all the principles, and there is a way that my lack of experience shooting is contributing to the variability.. some analytes higher while others in the same shoot are lower. I wish I had the vocabulary to be more succinct. Thank you for your contributions and patience.

Mike

[Peter Apps]

Hi Mike

It sounds as if you have some problems with your setup.

You absolutely must not use air as the carrier gas - it will destroy the stationary phase in the column.

You need air for the FID, but the air has to have a constant flow rate, and it has to be clean - especially it must not contain oil vapour. From the brand name I suspect that the compressor you are using is designed for paint spraying etc and unless you have a train of purifiers the air that it puts out will not be anywhere near clean enough to run an FID.

Gas chromatography labs typically use pure (99.999%) instrument air in cylinders, or zero air generators. You will be able to get the cylinders from the same place as you get the hydrogen. What grade of hydrogen are you using by the way ?, and do you have nitrogen as a make-up gas for the FID ?

Peter

[MichaelStuart]

Hi Peter,

The problem is really with my understanding of the setup and terminology. I misspoke. UHP Hydrogen is my carrier gas, and fuel for the FID. I do have a UHP N2 cylinder, but thought it was just for the ECD. N2 is the makeup gas for both? I have been out of nitrogen as my 310 cuft cylinder ran out in 10 days. I shut down the ECD before the N2 ran out and until I have a chance to leak test all the plumbing.

I was told by the fellow that installed the system that I can still run the FID for the potency screening without the N2.

What would the effect of being out of one's makeup gas have on testing?

I also addressed the need for moisture traps, filters, etc. on the air with the fellow and he said in the hundreds he's installed it was never necessary. Yesterday I was at my gas suppliers, called the installer of my system telling him I wanted to switch over to a cylinder for my air as I was suspecting the regulator on the compressor as a problem point, and his response to me was "you're crazy, the regulators on compressors should provide constant enough flow when set to 40psi." My FID signal at rest is a constant 10.0. My flow rate on the air is 350ml/m, so I calculated that I would need a refill delivered every 17 days if I go with a cylinder. The compressor is an oli-less system.

I also had read that I should not leak-test with a soap solution, as it is possible for the Venturi effect to pull soap in to the system. I addressed this with my technician who said.. "no, it's no problem".

Thanks for your patience as I learn my setup, and build my vocabulary. I am looking forward to knowing this setup inside and out.

Mike

[MichaelStuart]

I was informed that I am not using a makeup gas for the FID. My technician informed me that for the analytes I am looking for, and the program I am running, it is not necessary to use a makeup gas. He also let me know that if I look at the 'at-rest' FID signal, and it's consistent, the air quality for the sensitivity I am dialed in for is not an issue. At least, that's what I took from the conversation. I may have misunderstood some if not all of what he said, but I will continue to edify myself on these matters with hopes that the day will come where I am qualified to make these decisions for myself.

Again, I really do appreciate the time you guys have spent helping me learn these things. I do look forward to passing the favor on when I am equipped to do so.

Mike

[Peter Apps]

Hello Mike

You are safe to use UHP hydrogen as carrier gas.

As long as you do not need very low limits of detection, and as long as the tip of the column is inside the FID flame jet you do not need nitrogen as makeup.

BUT: I am worried about the compressed air. Even an oil-free compressor sucks in room air and compresses it - so the output is no cleaner than room air. Moisture and oil droplet traps are necessary but not sufficient - you need to take out organic vapours as well, which needs a charcoal trap that will have to be replaced quite frequently. Depending on your GC model 40 psi may not be sufficient inlet pressure for the onboard air flow controls.

A stable baseline is a good sign that there are no major problems, but a change in sensitivity of the FID can only be seen when it actually has something to detect, so you can have a stable baseline, but eratic responses to the same quantity of analyte in a peak.

The problem with liquid leak detectors in GC systems is not due to the venturi effect (although in principal it could occur) but to diffusive creep along the walls of the leak channel. If you are going to use liquids them do not use the kind that heating and air conditioning engineers use that make pretty bubbles when they are squirted on a leak. Mix isopropanol and water 1:1 and drip that on the connections - at least if some of that gets into the system it will evaporate in a few days rather than in a few months. Obviously you have to cool all the parts of the GC down for liquid leak seekers to work.

Has you technician mentioned that you need to drain the compressor tank quite frequently ?

Peter

[MichaelStuart]

Hi Peter,

Again, thanks so much for sharing your expertise on this. I think I will add moisture traps and filters to the air for peace of mind. I have only integrated 3 analytes in my report, and looking to quantify analytes in the neighborhood of 1000 ul/ml in all determinations. As these are relatively large amounts, is this perhaps why my technician in not too concerned with small amounts of organic vapour etc in the FID, or perhaps he is just sloppy.

I am excited to have my auto-sampler tower(s) coming. It will rule out human error, although I still have a difficult time believing that my injection technique would account for the wide variability. I do think I was seeing some deviation due to discrimination, as the reference that I am shooting has 3 analytes in there, each 1000ul/ml. When i shot it the signal showed each subsequent peak lower than the last. When I tried a hot needle injection technique, it started showing them as the same height in my online signal plot. Again, I am not experienced enough to know if my logic is sound.

My tech did not mention the draining of the compressor tank, however I have worked with compressors before and purge and drain my compressor daily.

I have my column tip within the specified 23-25mm (from the base of the column nut). I think it is set to 24mm.

What would you recommend for the inlet pressure? I do think I need to get a flow meter if I am going to be sure this is set up properly. Should I invest in a digital one, or would a bubble meter suffice?

Thanks again Peter


Mike

[Peter Apps]

Hi Peter,

Again, thanks so much for sharing your expertise on this. I think I will add moisture traps and filters to the air for peace of mind. I have only integrated 3 analytes in my report, and looking to quantify analytes in the neighborhood of 1000 ul/ml I think that you probably have 1000 ug/ml not 1000ul/ml; 1000ul = 1 ml so the latter is a pure substance. 1000 ug/ml is 1000 ng in a 1 ul injection, so a split of 20:1 will give 500 ng on the column (sorry I don't recall your split ratio) which is more than enough to overload the column - do you see the peaks rising more slowly than they fall to give a kind of sharks fin profile ? in all determinations. As these are relatively large amounts, is this perhaps why my technician in not too concerned with small amounts of organic vapour etc in the FID, or perhaps he is just sloppy.

I am excited to have my auto-sampler tower(s) coming. It will rule out human error, although I still have a difficult time believing that my injection technique would account for the wide variability. I do think I was seeing some deviation due to discrimination, as the reference that I am shooting has 3 analytes in there, each 1000ul/ml. When i shot it the signal showed each subsequent peak lower than the last. When I tried a hot needle injection technique, it started showing them as the same height in my online signal plot. Again, I am not experienced enough to know if my logic is sound.Sound as a bell ! - having established that you have inlet discrimination with a manual fast cold needle injection you need to set the autosampler to give you a hot needle injection also

My tech did not mention the draining of the compressor tank, however I have worked with compressors before and purge and drain my compressor daily.

I have my column tip within the specified 23-25mm (from the base of the column nut). I think it is set to 24mm. It's a long time since I used an Agilent GC, but that sounds shorter than I remember - if it's in the manual it should be OK.

What would you recommend for the inlet pressure? With hydrogen as carrier you need to have the gas flowing through the column with a linear velocity of 50 cm/s. To measure linear velocity you inject some gas from a cigarette lighter, time how long the peak takes to appear and then do the arithmetic with the column length - with a 30 m column you are looking for a 1 min holdup. Adjust the inlet pressure up or down according to whether the lighter gas comes through too slowly or too quickly. Start with 7 psi. I do think I need to get a flow meter if I am going to be sure this is set up properly. Should I invest in a digital one, or would a bubble meter suffice? A flow meter is essential for checking split, septum purge and detector gas flows, not so useful for column flow. Digital flow meters are a bit quicker to use but I prefer the good old fashioned bubble meter - they work with all gasses, they never need calibrating, and the batteries never go flat

Thanks again Peter No problem


Mike

[MichaelStuart]

Hi Peter,

Yes, it is in deed 1000 ug/ml..

do you see the peaks rising more slowly than they fall to give a kind of sharks fin profile ?No, peak shape is looking pretty good now that I have my purge set up properly.

having established that you have inlet discrimination with a manual fast cold needle injection you need to set the autosampler to give you a hot needle injection also. Thanks so much for the tip. When my auto sampler arrives again, I will make certain to set it up for hot needle injection.

My auto-sampler did arrive. However unfortunate, I do suspect it was vandalized by the fed-ex courier. I was told by an employee that he was "very angry" when he delivered the box. I asked how angry.."very". The sender sent it to the wrong suite number. I opened the box to find the tower cover and chassis cracked. I installed it anyway to find the syringe carriage was also damaged rendering the tower unusable. after hearing of the courier's anger I went back to inspect the box in the location that would have had to taken the impact to cause that damage and literally found a boot print. Filed a claim with FedEx.. new tower on the way. I'm about 40 days in to owning this GC and am really chomping at the bit to get this thing up and running. In the mean time I will try to learn more about the subtleties of this machine.

As always, much gratitude for your willingness to share your experience.

Warm regards,

Mike

[MichaelStuart]

A quick question for anyone who has an opinion on the subject... How much more difficult, if at all, is HPLC compared to GC to operate?