I'm currently running residual solvent analyses with the following GC system:
GC: Agilent 5890
column: packed, Poropak Q (col. temp.: 100°C, iso)
injection: 2 µl, on-column. Injection chamber is kept at 240°C, septum purge 2ml/min.
Carrier gas: helium
Sample: water solutions of different solvents (ethanol, acetone, etc.)
I'm performing injections with an Agilent AS (6890 series), and I have a carryover problem which I can get rid of only performing repeated analysis of the sample solvent (water).
I don't have any idea about a possible solution. Injection chamber temperature seems to me high enough. Maybe changing injection dwell time?
Agilent adviced me to decrease injection volume, which I made (from 5 to 2 µl), but with no improvement.
I have to add that when I used to perform the same analysis with manual injections I didn't have such problem.
Has anybody experienced the same problem?
By Anonymous on Monday, September 30, 2002 - 07:57 am:
Just an idea: Try to decrease the injection temperature. Water occupies a very large volume when in the gaseous state and it seems to me you might experience back flashing because the large volume you inject probably is too much for the volume of the liner.
If you cannot reduce the injection volume, then I would suggest you reduce the inlet temperature because the volume of the vaporised water will be proportional to the inlet temperature and pressure.
By Anonymous on Monday, September 30, 2002 - 08:51 am:
make sure your autosampler is programmed to rinse the syringe several times (just like you would do manually), pump several times (just like you would do manually), then rinse with your water solvent (just like you would do manually). This sounds more like a programming and/or syring problem than Agilent's, because that autosampler basically copies what a human operator would do. When you say you get no carryover with manual injections, is that using the SAME syringe???
By Anonymous on Monday, September 30, 2002 - 09:41 am:
The Agilent autosampler unlike other autosamplers is not capable of a complete flush of the syringe, only a rinse using a vial of wash solvent. One cannot be certain that after using the 'wash solvent' it is not contaminated with solvents from the first cleaning of the syringe after the first injection.
What may be acceptable for a simple analysis of a concentrated solution may not be acceptable for a trace analysis of a solvent in a drug solution.
I won't suggest a more useful autosampler from another manufacturer but anyone who checks out the competition will see the practical benefits of a design other than Agilent's. Perhaps they will someday offer a better design, but that day has not yet arrived. One can hope.
Meanwhile I prefer other designs for my analytical work.
signed, anonymous for obvious reasons (note: I do not work for the competition)
By Anonymous on Monday, September 30, 2002 - 03:43 pm:
Anonymous 09:41am - your explanation doesn't account for seeing a difference with manual or automatic sampler use (assuming the same syringe is used).
By Anonymous on Tuesday, October 1, 2002 - 04:39 am:
In the autoinjector system is the syringe needle entering the column, or is the injection made into a chamber of glass liner? In the manual injection system is it a direct on-column injection? This could be the source of the difficulty. In either case I would reduce the injection volume to 1ul or even 0.5ul for the reasons stated by first ananymous.
By Anonymous on Tuesday, October 1, 2002 - 06:50 am:
Definately sounds like backflash due to too large an injection volume. I believe the reason you do not see the problem in manual injection is that the manual injection is much slower than the autoinjection, so the degree of backflash is reduced. No more than 1 uL, less if you can meet the detection limits using less.
p.s. to anonymous ripping Agilent washing method- I have used most of the commercially available autosamplers and Agilent is about average in terms of washing efficiency. If the problem was bad design this would be a general problem. I don't work for Agilent and have no great love for them, but syringe washing with a standard type syringe is something they do as well as anyone.
By Bruce Freeman on Wednesday, October 2, 2002 - 08:53 am:
I agree with some of the above responses, but let me clarify WHY they are correct:
When the autosampler rapidly injects a sample into a hot inlet, the sample may flash vaporize. (Or it may bounce around like water on a hot iron, but that wouldn't explain your problem.) When it does so, it suddenly occupies much more volume and may tend to back up into the cool upstream plumbing of the gas chromatograph -- leading to carry-over.
Reducing the injection volume decreases the vapor volume, reducing the problem. (Calculate the vapor volume of the injected water at the inlet temperature, and compare that to the available volume ahead of the column.)
A slower injection (as a manual injection usually is) can help by reducing the pressure pulse. Unfortunately, you can't control the injection speed on an HP5890.
Reducing the inlet temperature will slightly decrease the vapor volume, but may substantially reduce the rate of vaporization of the liquid. You don't need 100C to vaporize water in the inlet. 5 mcL of water placed in a petri dish will be gone in an hour because it evaporates into the air at RT. It's all a matter of the "relative humidity" of the carrier gas -- which is zero. So you can probably use an 80C inlet just as well as a 250C inlet.
Using glass wool in the inlet can be a useful alternative. Liquid will hit the glass wool, cool it quickly, and stay there as it evaporates. Hence, glass wool simulates a cooler inlet, without using a cooler inlet.
Carry-over might come from a syringe. Changing the syringe "frequently" is useful to prevent blow-by (past the piston) which will increase as the syringe ages. Look at your syringe. If there is a darkening of the barrel (metal worn off the piston) about 1/4 inch from the top end (opposite end from the needle) then throw away the syringe and replace it. (Syringes are disposable items, like septa.)
Best of luck.
By Anonymous on Wednesday, October 2, 2002 - 09:52 am:
From anon 9:41
to anon Mon sep 30
the difference is manually cleaning the syringe to the less than perfect washing procedure Agilent offers. Manual cleaning is usually cleaner - better.
to anon Tues Oct 1
I would not argue (or disagree) with your statements. Be that as it may, for general purpose work Agilent's autosampler is OK, but that is faint praise from me. The average car is not what I prefer to drive when I am on a slick or on a dangerous road. I want the best. I have to use Agilent GCs routinely, but I don't have to like it, especially when I have used much better equipment, but each manufacturer has his strengths and weaknesses. The Autosampler from Agilent (HP) I consider a weakness, not a strength. But I am capable of discussing other weaknesses from other manufacturers, just start the thread ! :-) I am ready. I have a justification to complain. I am a customer.
It is my money.
Anon 9-30-02 9:41 am
By New anonymous on Wednesday, October 2, 2002 - 10:03 am:
For what it's worth, I was always shown to rinse syringes manually with solvent, then rinse with sample, then pump in the sample, then pull back to the volume mark then inject. So Agilent's way matches well what I'd do manually.
By Rodney George on Wednesday, October 2, 2002 - 05:47 pm:
A common problem with direct injections of drugs or other compounds that are undergoing residual solvent analysis is the deposition of these non-volatiles at the head of the column. If not removed periodically they can capture analytes of interest from stds and other solvent containing sample solutions. These trapped solvents may only be eluted after an injection of clean carrier solvent, hence, a carryover, and irreproducible solvent peaks.
Clean the glass and remove the first 6 inches or so of the porous polymer beads of the column routinely to minimize this problem.
Water injected into a 240°C injection area is a bit extreme. You can do quite well running at 180°C max. Inject slowly (1µL per sec) to reduce the reverse flashing of vaporized steam and the deposition of drug(or other compound) into the upstream areas of your GC carrier system.
Feel free to call Supelco Technical service (a toll free call or email) for additional help.
Senior Research Chemist
Supelco Gas Separations Research
By Rodney George on Wednesday, October 2, 2002 - 05:52 pm:
You can use 5µL of carrier solvent if your injection liner (or column) ID is 4mm, your flow of carrier gas is flast enough(40-80cc/min typically), and your injection rate is slow enough to avoid flashback (with porous polymer bead packed columns or 0.53mm ID columns).
But this is rarely necessary, unless you are doing low ppm analysis and your conditions are just right for the analysis.
By labcat on Thursday, October 3, 2002 - 02:51 am:
Thank you for the several useful advices I received, which allowed me to give the proper direction to my experiments!
I have to add that in my first message I forgot to mention that I was pretty sure it wasn't a syringe problem, because the carryover was there even when changing the syringe between two injections. Anyway, syringe is rinsed with clean solvent 5 times before and 5 times after the injection, which is the same I was used to do when working manually.
In the last two days I ran a series of experiments changing some parameters and I discovered that:
1) The most important factor to reduce carryover is the speed of the injection. I can choose between fast and slow injection (Bruce: luckily we have a 6890 AS!), and with slow injection the "residual" peak can be less than half than with fast one. I guess that what I believed to be a "fast" manual injection wasn't that fast!
2) almost as important as the the previous is the injection chamber temperature: decreasing from 240°C to 180°C is very useful (I didn't try lower temperatures).
3) some significant decrease I had also with a post-injection dwell time of 6 seconds (I am not very sure about interpretation of this finding: disturbance of the vapour cloud due to withdrawal of the syringe?)
4) to my surprise, changing injection volume from 1 to 3 µl has no effect
5) also initial temperature of the column has no effect (60 and 100°C).
I think that vapourisation rate seems to be the critical parameter to control, at least when working with this kind of solvents. And above all, I learned a lot from this discussion!
Posting is currently disabled in this topic. Contact your discussion moderator for more information.