Sample solvent influence on injection volume (Acquity)

[Mattias]

Hi,

My latest findings regarding the injector of the Acquity autosampler is disturbing enough to earn an own post. This could affect many people.

Situation 1. Partial loop, needle overfill injection (50 µl loop, 250 µl syringe). Injection volume 20 µl.
Injection from a 1 ml vial "as is": main peak area 3670000
Addition of 200 µl ACN to the same vial: main peak area: 3520000

This caused a lot of turbulence since it clearly indicated that we had a adsoption problem of the product. We should see about 83% of the area in injection 2, but we saw 96%. Acetonitrile increases the response.

Situation 2. Full loop injection (50 µl loop, 250 µl syringe)
Injection from a 1 ml vial "as is": main peak area 9010000
Addition of 200 µl ACN to the same vial: main peak area: 7780000

Now this is more how it should look! I cannot see any other explaination than that the Acquity in partial loop mode injects a larger volume when the sample contains acetonitrile, compared to the normal sample. In full loop, the sample solvent makes no difference. (I have repeated these experiements many times, with the same result)

Since it is often is not possible (or desired) to prefectly match the solvent composition of the standard and samples, my conclusion is that full loop must always be used if the Acquity is to be used for quantitative analysis.

[AA]

The composition of the weak wash can significantly affect this result. What are you using? Also, depending on your sample a different needle material (SS, peakSil) might give you a better result. inally, the lower volume needles (15 ul) can also help somtimes in these cases.

[Mattias]

I have now got the confirmation from Waters in Denmark that partial loop injection should not be used for quantitative analysis on Acquity. There was no further explaination.

[AA]

I would not fully agree with the answer from Waters Denmark.

There are 2 partial loop modes in Acquity, partial loop with needle overfill, which is more than adequate for quantitaition. The other partial loop mode does not have the precision that most users would want, its major benifit is speed.

By the way, you deserve some level of explination from you local Waters support team. It is certainly not good enough for them to simply state "dont use it".

[Flappytango]

Mattias,

i dont have time for a full reply at the moment but..

the injection omdes operate very differenly. there a couple of documents to look at that describe this from waters.



UPLC SM injection Mode Card

and

[ACQUITY UPLC System] – Injection Modes

ill try to follow up later

[Flappytango]

What you describe sounds more like phenomena that could be experienced in PLPA modes. I am surprised that waters would say you cannot use PLNO for quantitation. This goes against all of their documentation.

well this might be helpful anyway..

the link to the injection card

http://www.waters.com/waters/support.ht ... &type=USRM

I could not find a link for the other document which has the best description of how the modes impact the injected sample with regards to sample sensing, dispersion effects, volumes drawn and injected, and how the instrument determines where to "cut" the sample. i believe it was part of a waters presentation. i can email it to you if you like.

One thing i would suggest is to calibrate your Bubble Sensor. I was observing very large differences between injection modes and this helped.

Keep in mind that you do not get the same amount of sample actually injected with the different injection modes so comparing area between the different injection modes is convoluted. See the below image. Of couse this does not explain what you observed.

http://s1190.photobucket.com/albums/z45 ... figure.jpg

[carls]

Mattias' primary observation was that small changes in sample viscosity greatly affected injection accuracy when using partial loop with needle overfill. Accuracy was unaffected with full loop.

This might indicate an inherent problem with the "complex" injection system used by Waters.

Waters' quick reference card states "when changing weak-wash solvents, characterize the needle and loop volume".

Why?

Does this suggest a viscosity effect?

As Mattias stated, this could be a major problem since the viscocity of samples and standards are rarely identical.

[Mattias]

The confirmation from Waters came during an informal discussion with one of the field engineers. I do not have it in writing.

The data speaks for it self I think, and I have also confirmed the observation on the other Acquity system that we have in our department. This seems to be a general problem. It would be nice if any of the Acquity operators here in this forum could repeat the experiment with any molecule at hand.

Regarding the other questions: The bubble sensor is newly calibrated (during the loop change). I run 10% acetonitrile in water as weak wash. The method runs at 20% acetonitrile. Both our systems have no special configuration - they are the same as when they left the factory.

[Peter Apps]

A universal problem with partial loop fill, whether it be done by a good old fashioned six-port valve and syringe, or a fancy automated system, is that the liquid being pushed into the loop does not form a plug with a square front. Instead it has a roughly parabolic front profile that matches the velocity profile for streamlined flow through a tube. Consequently some of the introduced liquid (sample) will escape from the loop before a loop volume of liquid has been introduced. The loop then contains less sample than was introduced. The shape of the velocity profile, and hence of the leading edge of the sample plug depends on the viscosities of the liquid already in the loop and the liquid being introduced.

Peter

[Flappytango]

Mattias,

I see that your email is available i will send you the waters document i mentioned. I cannot find it currently on thier website.

It discusses the laminar flow and dispersion that Peter Apps is referring to. It also states that recovery can be viscosity dependant as carls questioned.

its is a good read but they have not put a doc# on it...

an excerpt
"Disadvantages: weak wash solvent, mobile phase and sample are co-injected,
and thus sample peak recoveries are method dependant. Recovery depends
upon:
(a) Sample viscosity and draw rate of the syringe.
(b) The ratio of the sample volume vs. loop volume (stay within 10 to 50% of
the nominal loop volume).
(c) The ability of the Sample Manager fluidics and firmware to create an
accurate “mapâ€

[AA]

The excerpts that Flappytango refers to only apply to the partial loop mode, not the partial loop needle overfill mode.

[Mattias]

I really would like the opinion of the Waters representatives in this forum about this. I have a feeling that this is common knowledge within Waters.

QA in my company have started an investigation of how this has affected our bioanalytical, QC and stability results since we started to use Acquity. Hopefully not so much since the methods are validated, but it is disturbing.

It certainly made me look like a fool, ringing the bell about product adsorption to management.

[Flappytango]

The excerpts that Flappytango refers to only apply to the partial loop mode, not the partial loop needle overfill mode.

true i should have stated that.

Overall i feel it is important to understand the operation of all the injection modes and how they impact the composition of the delivered sample. Reading all of waters information is a start.

Mattias for your particular issue, have you directly contacted waters to get some clairification on how what you experience is possible? They should forward this to their technical experts.

Also you might try the waters Aquity UPLC community forum. The waters reps are very active there.