Problems with Kinetex C18, 1.7 µm

[Mattias]

Mattias,

I'm sure you are aware of this but even a UHPLC system must be optimized in order to realize the improvement in performance provided by the highest effieciency columns available today. The QC results for the columns in question provide a good estimate of the performance under realistic conditions. These results are generated on standard UHPLC instruments (>100 instruments) so users should not have a problem reproducing these results when using similar equipment. If you are unable to reproduce the QC results then the system must be optimized. Unfortunately, many UHPLC system are not capable of showing the true performance of these ultra-high effieciency sorbents without careful optimization.

The QC department where the method has problems (that could be related to column batch variations, we don't know), also has a Waters Acquity instrument. I have changed the capillaries to and from the column, but I have not done more than that on their instrument.

The trooble shooting point towards the columns, but I feel that it is very strange that I every column that I buy and start to use works fine (for at least 5-6 batches of columns). QC have now two batches of columns with the same problem (I guess I should ask them to send one of their new columns to me).

Back to the instrument: are there other things that need optimisation, or that could make peaks splitted? Their Acquity is almost new, and my instrument is from 2006. It appears as I generally get higher plate counts than QC.

[Hollow]

Back to the instrument: are there other things that need optimisation, or that could make peaks splitted? Their Acquity is almost new, and my instrument is from 2006. It appears as I generally get higher plate counts than QC.

any bad connection, introducing some voids, could generate peak split.
We once had a demo Acquity that showed typical peak splitting. Later, we were told, that after they had reconnected the whole flow path, the peak splitting was gone...
But voids wouldn't generate higher backpressures...

Maybe have a close inspection of the columns in- & outlets.
If there is some plastic debris, e.g. from the closing fitting (threading). this might generate high backpressure and symptons like a blocked inlet frit (splitting).
Once we had a HPLC column that was fully blocked by some unknown plastic part... knowbody knew where it came from...

Regarding the variation of the particle diameter, I would make sure to keep the L/dp ratio contstant and make some comparative studies.
By this you're not going to change anything on the selectivity nor do you reduce the (theoretical) plate count of the column. (this was something that Uwe Neue suggested for the revision of USP chapter 621 but revision's still pending).

I would prefer this rather than to switch to a new stationary phase of the same particle size. I guess it should be easier to argue as it is still the same chemistry.

If going from fully porous to core-shell, I would also show that there are know overlading effects on the core-shell, due to the smaller surface on these particles.

[Klaus I.]

I have not found any trends in the certificates!

Just to summarize your observations of the back-pressure:
[a] you have noticed that the backpressure of NEW(!) columns would be have a large range from 7000 to 9000 psi
the 9000 psi columns give usually a poor peak shape
[c] you have observed no trends in the certificates (what do you mean with this, since the factory test-mixtures are very simple, you will in many cases not noticed difficulties, even when they are really exists).

Does your different pressure observations corresponds to the noted back-pressure during factory testing?

[Mattias]

I have not found any trends in the certificates!

Just to summarize your observations of the back-pressure:
[a] you have noticed that the backpressure of NEW(!) columns would be have a large range from 7000 to 9000 psi
the 9000 psi columns give usually a poor peak shape
[c] you have observed no trends in the certificates (what do you mean with this, since the factory test-mixtures are very simple, you will in many cases not noticed difficulties, even when they are really exists).

Does your different pressure observations corresponds to the noted back-pressure during factory testing?

These are the observations at the QC department, and your summary is correct. I never have any problems with the columns in my lab (R&D).

Update: I sent two columns to QC which I proved worked very well in my lab. They install these columns and just barely pass the resolution demand of 1.5. I got resolution of 3.0.
We check the plate count of the main peak, and their value is less than 50% of mine.

So now it is proven that this has nothing to do with columns after all. Why is their Acquity giving so much bad broadening and sometimes peak splitting, when mine doesn't?.

I have asked them to check the volume settings in the sample manager (loop, needle, syringe). I guess if they are wrong, anything can happen? These values should be set by the Waters technician, but anyone can make mistakes.

What I have done is to optimise the capillaries and fittings, and also exchange the wash solvents to something that is weaker than the mobile phase. What else can be wrong with the QC instrument? They get the right retention times, and the pressure ripple is very low = pump is working fine.

[yangz00g]

Then the problem may be related to system dead volume, a common error cause for LC method transfer, especially UPLC

[carls]

Have you verified that the wash solvent lines are correct (i.e. not reversed)? Trace them thru the degasser to the syringes

Is the volume of the weak wash solvent set to at least 600uL.

Have the needle and loop volumes been characterized recently?

How do the peak areas compare between instruments?

[Mattias]

Have you verified that the wash solvent lines are correct (i.e. not reversed)? Trace them thru the degasser to the syringes

Is the volume of the weak wash solvent set to at least 600uL.

Have the needle and loop volumes been characterized recently?

How do the peak areas compare between instruments?

Hi,

I have told them to put 10% acetonitrile in water in all wash lines (just to be safe). Cannot guarantee that they actually did it.

They characterised the needle and loop volumes today, and they corresponded well to the settings. The only thing that can have an impact is that their 20 µl loop is actually 25 µl (so they inject larger volume than I do, since our loop is very close to 20 µl). But I wonder if these 5 µl can make any difference? I will test that tomorrow (we use full loop for this method).

I have a question about the needle volume: does this volume affect the actual injection volume? I have noticed that I get a 20-30% higher peak area if I inject 20 µl full loop, compared to 20 µl partial loop needle overfill (using a 100 µl loop).

We get comparable peak areas, but we use different detectors (PDA at R&D and TUV at QC). The cell volumes are identical (500 µl).

[carls]

I would install and test an Acquity C18 50x2.1mm column (preferrable new, at least one of known good conditon) and run Water's QC test mix with the conditions/analyte (acenaphthene?) shown on the C of A for the column. Your results should be be within +/-10-15% of the value shown there for tr, N and As. If your results are significantly different then its likely time for a service call.

[KM-USA]

Particle size can be reduced by up to 50% according to USP chromatography chapter. Doesn't say about increasing.

Do a risk assessment. As long as resolution is still intact there is no reason to fully revalidate. Same phase, same everything...makes NO sense to revalidate if they know ANY chemistry.
Or adjust parameters so chromatography is exactly the same.

ACK. As long as the chemistry (stationary phase) stays the same, you should get away with a small amount of work. I also think you should do a risk assessment stating that by changing particle size only the resolution of the chromatographic method should change. Then redo the selectivity part of the validation to show that 1) there's still enough resolution and 2) selectivity of the columns actually is equal (i.e. no retention time changes - wouldn't be the first time that different particle sizes of the same column actually gave slightly different selectivities ).
Explaining this matter to your regulatory people will pe a pain in the a**, though

Side questions (partially):

(1) so as to "allowed USP parameters" which USP <621> states, can one arbitrarily make such "allowed" changes as long as the change is documented, or does some vague requirements of validation need to be performed? If work needs to be done, I wonder why such "allowed changes" would even be detailed, if revalidation studies were required regardless. In other words (like much of USP/ICH) I'm left confused.

(2) concerning such allowed changes: would those be applicable only to USP-NF monograph substances, or also to consumer products using USP actives, whose staff has developed and cGMP-validated their own finished product API assays, which contain small amounts of API in a finished OTC product.

Thanks.

[tom jupille]

can one arbitrarily make such "allowed" changes as long as the change is documented

So long as you meet the system suitability requirements (in other words, the system suitability defines the method).

would those be applicable only to USP-NF monograph substances,

Yes

or also to consumer products using USP actives, whose staff has developed and cGMP-validated their own finished product API assays, which contain small amounts of API in a finished OTC product.

No. Although the FDA has a similar set of default adjustment limits that apply when they run methods submitted by industry (search the FDA web site for document ORA LAB4.5.4). If you wanted to implement those same limits, it would be a good idea to incorporate them into an SOP (citing that FDA document as justificat.ion)

[KM-USA]

can one arbitrarily make such "allowed" changes as long as the change is documented

So long as you meet the system suitability requirements (in other words, the system suitability defines the method).

would those be applicable only to USP-NF monograph substances,

Yes

So can one simply move from (example) 5 micron 250mm x 4.6mm i.d. column to same chemistry particle and 3 micron 150mm x 3.0mm i.d. to modernize the chromatography?

Or can changes only be made if a system does not meet system suitability?

[tom jupille]

So long as you meet system suitability and your parameters are within the allowed default range relative to the published method, you're good. Tom Layloff said it best years ago when he was at the FDA: "What part of 'the system is suitable' is unclear?"

[Mattias]

Now the Waters technician has spent two days on the QC instrument. Nothing strange found, the instrument works as it should. But still it cannot run my method... (that works on two Acquity and one Ultimate system in my lab).

I have told them 100 times to check what they use as needle wash, but I never really get any answers. I probably need to spend two days travelling just to see what is actually in these wash lines. It is the only thing that can explain that all peaks are splitted (even the really small ones).

[DR]

BTW: The usable (allowed) maximum pressure of Kinetex 2.6 µm is specified to 600 bar (8700 psi). Only for columns with an ID of 2.1mm, Phenomenex have raised the maximum allowed pressure to 1000 bar (1500 psi).

Correction, 1000BAR=14,504psi (let's not have a bunch of students asking us what is wrong with their pumps )

[Fri_nick1]

I never had good experiences when trying kinetex column in our Shimazu Nexera system. I had problems with peak area consistencies, bad peak shape , etc. There seems problems with the loading and packing of the column. I did not waste my time for further investigation. The zorbax sub 2 um seems to be working better for us.