Chromatography Forum

There's always a possibility than you are forming a vacuum in the vial

Hi Bruce,

If I understand you correctly, you’re suggesting that the vacuum is created upon the penetration of septum by the needle. I am aware of this effect, but it is likely to diminish by each penetration/injection and not increase – as a matter of fact it is most likely to disappear following the first penetration (some people call it “first penetration injection errorâ€

As I said, the experiment ( slit the septum ) will decide whether vacuum is an issue in this case. Another way is to inject from a different vial of the same solution.

The vacuum in the vial may cause less to be fed into the injector. If the vial vacuum keeps increasing ( which in can if the needle/septa seal is hermetic ), the quantity injected each time will get smaller.

The vacuum can either cause the sample loop to take less, or, as the needle rises, it passes though vapour space and a small volume could be backflushed from the needle.

A simple experiment would determine if the concept is relevant in this case. It's certainly relevant of low vapour pressure, large injection volume situations.

With regard to viscosity, the experiment of changing to a lower viscosity solvent ( preferably the mobile phase ) would also ascertain if the effect is relevant in this case.

Autosamplers are not normally designed for samples with viscosities greater than water. I just think it's a bad idea to try and use solvents that have significantly-different rheology to the mobile phase for quantitative HPLC.

I favour experimentation to speculation.

Bruce Hamilton

You do need to use an autosampler vial of the type that is recommended by the instrument manufacturer.

Vials can have septa that are teflon only (of varying thicknesses), pre-slit teflon and teflon with a silicon backing (and there are others).

I have seen instances of the vacuum effect that caused injection reproducibility errors. Switching to the correct vial type solves the problem. Generally, it is the vials with a teflon/silicon septa that can cause the vacuum effect.

You can get the vials directly from the instrument manufacturer or from a third party supplier. But be aware, I have seen instances where vials from a third party supplier have not performed properly for a particular instrument even though it is the recommend vial (it may be a quality control issue with the vial). It is sometimes a trial-and-error approach to find a suitable vial.

As to the matching of injection solvent to mobile phase, the solvent characteristics to consider are: pH, polarity, ionic strength and viscosity (maybe others too). The injection volume is also a consideration; solvent differnces may be less important with smaller injection volumes.

Regards,
Dan

i would do one change right now.

inject twice
do not flush
but wait 30 minutes worth of the flush
reinject
if the area is like 1st injection then the problem is in the vial
if the area is smaller again then the flushing is what makes the difference, so the column is part of the problem

Dear moonchips:
Use Acetone instead of Caffeine to test your system and vial (Dilute 1:1000 or more). You will know if your vial&septum is the culprit. I believe that you are using GC vial with screw cap? Then you need to change to LC vials (use wide-neck) and caps with pre-cut septa.

You did not state how many components in your matrix. Is this a new method you try to develop? Can you state the RT of components? Do you achieve baseline-resolution for all peaks? Could you provide a pic of your chromatogram? Any change in peak symmetry if you inject 10 or 20 times without your 'flush'?

Just to clarify a misconception. I don't disagree about the need to ensure the vials are compatible, but they wouldn't be first on my list, however unsplit septa may create issues in some situations.

The vials moonchips uses should be suitable for the Agilent 1100 - they are not just GC vials. I've used them, and at one stage Agilent may have been selling NI Target Vials. I stopped using them because they can sometimes shatter during ultrasonication, but that issue can affect many vial brands.

"Target DP (dual purpose) vials are designed like crimp-top vials with the convenience of screw-thread vials. They are compatible with Agilent 7673 and 1100 series robotic arm samplers. Vials are made from Type I, Class A borosilicate glass. Target I-D* vials have a ceramic write-on ID patch with graduations at 0.5, 1.0, and 1.5mL."

My perception is still that the mismatch between sample solvent and mobile phase is a probable cause, whether viscosity, solubility differences ( may explain some compounds are OK, and others not, from the same injection ), or one of the other properties Dan listed.

I'd also want to review and modify any method that requires frequent flushes with a different solvent during a sequence.

Please keep having fun,

Bruce Hamilton

I'd also want to review and modify any method that requires frequent flushes with a different solvent during a sequence.

This is a very important and I would even say the only one adequate attitude. Nobody wants to work with unexplainable procedures. In this particular situation the first step should be; making absolutely sure that the recovery is decreasing systematically and whether it is a continuing tendency or it just changes direction in a non predictable manner.
Another important action should be to find out whether or not something elutes under the flushing step/s.

Troubleshooting is: Creating a system out of chaos!

Best Regards

moonchips,

what is the retention?

did you ever inject blank /diluent after flush?

is it possible that your flush increases the area? so 14 is the normal .

ym3142
the retention is 10 min
yes, I injected blank

each day, with a fresh system, peak area is 19-20, and then it decreases with multiple injections of standards till at the end of the day, it is 14. So I believe normal should be 19.

unmgvar,
I tried runing 70% IPA blank between BKF standard injections before (about 35 min run), and it did not affect peak areas.
Only flushing with some buffer under a gradient program can stop peak area decreasing and bring peak area back.

Alfred88
You did not state how many components in your matrix. Is this a new method you try to develop? Can you state the RT of components? Do you achieve baseline-resolution for all peaks? Could you provide a pic of your chromatogram? Any change in peak symmetry if you inject 10 or 20 times without your 'flush'?

9 standards in my matrix.
it is a new method
RTs range from 7.5 to 19 mins
yes, baseline resolution. worst resolution is 2
did not try peak symmetry test. what is it going to affect?

about vacuum or flushing issue, I tried injection without cap and comparing it with "with cap". before I run the flush program, peak area keep decreasing, and cap did not affect.

ym3142
each day, with a fresh system, peak area is 19-20, and then it decreases with multiple injections of standards till at the end of the day, it is 14. So I believe normal should be 19.

With a fresh system, you mean fresh mobile phase?

If you make a bottle mobile phase (A), inject a few standards from the same vial (A), your area decreases. What if you use the same standard solution (same volumetric flask), but in another vial (B)? And a freshly prepared standard in vial (C)
And if you make mobile phase fresh (B) and inject again vial (A)?

I would suggest doing the following:
Make mobile phase A, make a standard in vial A, inject a few times, within a couple of hours (4-5) this should be enough to see the decrease I think. Then fill a new vial (B) and inject. Also prepare a new standard solution and inject this vial (C).
Now make fresh mobile phase (B) and inject vial A, B and C.
Do the whole test within 1 day, and you can see if it's vial related, stability of the solution, mobile phase related, or something else.


Just my suggestion.....

Bart

moonchips this is your system:

column: Thermo Hypersil ODS C18, 5 µm, 120 Å 150 mm*4.6 mm

mobile phase: water/ACN

Gradient:
Time water ACN
0 50 50
11 0 100
25 0 100
30 50 50
35 50 50

flushing with 20 mM KH2PO4 (80): ACN (20) to do the flushing for about 30 min in a gradient mode:

Time (min) % Water 20 mM KH2PO4 % ACN
0 0 20 80
12 0 20 80
17 80 20 0
20 50 50 0
30 50 50 0

dilute of sample is 70% IPA

9 standards in matrix.
it is a new method
RTs range from 7.5 to 19 mins
worst resolution 2.

and the peak area of BFK is going down over time. flushing solves the problem

peak area loss can be divided into 3 main possibilities:
1. less sample is getting out of the vial. most of the time septa are the cause of that. this does not look like the case
2. your sample is degrading. again not the case since from the same vial after flush the peak area is back to normal and no new peaks are seen in the chromatogram.
3. your sample is getting "lost" inside that black hole we call the column. most probably the reason.

2 things are coming to mind looking at your system.
A. you say 9 standards in your matrix? is the standard injection made of 9 compounds or only BFK? is BFK the only one getting smaller?
B. when you flush do you collect data? if so then do a flush run prior to any STD run. run several times, even inject bigger amounts. Reflush, any peaks showing up anywhere on the chromatogram of the new flush that are not in the first flush?

if you answer what i expect you will have to do some adjustments.