Autosampler Problem - Precission

[syx]

Dear Members,
I have an odd problem with autosampler. A validated method was used to determine the released active substance in dissolution samples. On Tuesday the problem was not occurred. Next days, a series of injections from same vial gave great variability (%RSD not met the criteria – it was over 2.0%).
Different speed of syringe movement was tested and the problem did not move away.
Changing septum or without septum (the solvent was aqueous buffer solution without organics – so, we believed that the evaporation was minimum) was also cannot eliminate the trouble.
Then we checked the performance of autosampler and found nothing wrong.
We were back to system suitability test using standard solution and the problem came again.
I asked my analyst to increase injection volume from 20 μL to 30 μL. I assumed that greater injected sample may reduce variability, but I never predicted that the correction was very extreme: %RSD reduced to less than 1.0%! I am so amazed and do not know how to explain this phenomenon. Could anyone give me possible cause of the problem?

Note: Isopropanol was used as needle rinsing/washing solution.

Best regards,
SYX

[tom jupille]

How big is the syringe in the autosampler?

[syx]

I think it is not too big; it has about 9 cm length and 0.6 cm outer diameter. The volume is 100 uL.

[tom jupille]

Okay, that eliminates the hypothesis that you were down around the lower limit of the syringe range.

[andy]

Is your autosampler used primarily to make 20 uL injections? Sometimes the syringes get a worn spot when use to make a lot of the same size injections. When this happens, the plunger has more play and can effect the injection reproducibility.

[bert]

What kind of autosampler are you using? Is the autosamples equipped with an injection loop and what is the volume of this loop?

regards Bert

[syx]

It has 100 uL loop.

Btw, Mr. Tom, how could we open the locked post? I want to continue my old post but it is locked. Thank you.

[bert]

We had quite some problems with our Alliance HT 2790 when doeing partial loop injections. Now we always try to do full loops. So I am still curiuos which autosampler you are using.

[syx]

Hitachi LaChrom Elite L2200

[unmgvar]

Syx

Since you are using a L-2200 i might be able to help you.

i have a few questions first:

is your needle down speed slow? if no, then set it to slow.
this sometimes helps with the L-2200 eventhou we find no logical explanation for it. it mostly helps if your sample is being cooled. this is how i solved an EP application of calcitriol which was showing RSD problems

after you finish your set of injections do you wash the system with water?
if you have an high buffer concentration in your mobile phase then you should have either a low flow or wash the system straight away with water. the sample loop is part of the path of the mobile phase. leaving it dirty leads many time to RSD problems.

you say that you used IPA for needle rinsing, why?
as a thumb rule for the L-2200 you get best result when your washing soluton is either 1:1 water:MeOH or water:ACN.
the way that the needle is washed in the L-2200 is only in the outside.
the washing solution never mets the sample in the system (except one tiny moment in the wash port). you should use what ever solution gives you best dealing power with getting reed of the left over of sample solution in the outside of the needle. try to use 100% water instead of IPA.

the inside part of the needle is washed by the mobile phase itself.

in the software how many wash port strokes have you set for the wash port wash after the injection?
for 100ul syringe set it to 3 minimum.

what have you set for the needle wash time?
you should keep it at 1 seconds. this is the only place that the wash solvent meets the sample:
at the surface of the tip of the needle. the longer your wash time then the more diffusion occurs at the surface between the washing solution and the sample. never set to 0 sec (high carry over) but also be careful if you feel that you need to have it moe then 1 sec. this might cause a dramatic fall in RSD.


my guess is that because your sample is dissolved in buffer you have small micro deposit of salt in the outside of your needle maybe after a while even in the injector port. this leads to some type of sample loss. the moment you use an higher injection volume you compensate against un even rate of sample loss between injections and therefore your RSD is better.

[tom jupille]

To syx:

Threads are locked 30 days after the last post. This is done to avoid a problem we had on the "old" Forum with people appending new questions to old threads.

If you e-mail me privately with the topic and date I can unlock the thread.

[syx]

is your needle down speed slow?
>> I will try to do this. I just have tried to reduce syringe movement speed but not needle down speed.

after you finish your set of injections do you wash the system with water?
>> I always wash the buffered system with water (in combination with organic solvent).

you say that you used IPA for needle rinsing, why?
>> I take IPA based on LC Troubleshooting article in LCGC Europe-November 2001 about Attacking Carryover Problems. It stated that IPA makes an excellent wash solvent for many applications and proves to be more effective for removing contaminants than methanol and acetonitrile.
In other article, LCGC Europe-June 2001 there is a statement:
“The primary function of a flush solvent is to remove residual sample from portions of the plumbing that are unswept by the mobile phase. As a result, the wash solvent seldom, if ever, is injected. For this reason, analysts can select flush solvents for their solvating characteristics rather than their compatibility with mobile phases. Of course, a wash solvent should be miscible with the mobile phase, but it doesn’t need to match it directly. For example, workers can generally obtain better removal of sample residues by using a solvent stronger than a mobile phase. Thus, if you were using a 50:50 acetonitrile-water mobile phase, you might use 100% acetonitrile for a wash solvent. If the solvent isn’t strong enough, then residual sample might not be flushed from the system and carryover can occur.â€

[unmgvar]

Syx,
i was asking about the wash solvent because you yourself said that your sample solvent is made of a buffer solution:

Changing septum or without septum (the solvent was aqueous buffer solution without organics – so, we believed that the evaporation was minimum) was also cannot eliminate the trouble.

could it be possible that the IPA cause small precipitaton of the buffer inside the wash port and on the outside of the needle (which will "pollute" the injector port as well)?

that is why I suggested that you would change the washing solution to one that handles the buffer better.

in the L-2200 case as i said the wash solvent is only responsible for washing the outside of the needle. all the plumbing is swept by the mobile phase. the washing solution doesn't take care of that anymore.

[syx]

Unmgvar, thank you for the advise. I will try to change the washing solvent to aqueous solution, but still stronger than mobile phase to avoid any precipitation of the active substance.

[Mark Tracy]

Perhaps your autosampler is not the problem. If your detector is set to a wavelength that is on the steep slope of the UV spectrum of your base, you will also see poor RSD and linearity. The well-behaved component would be detected near an absorption maximum. Have you tried wavelength switching or a multi-wavelength detector?