short life of bead packing on ZIC-HILIC and ZIC-cHILIC

[carlo.annaratone]

Hi all,

I am currently having a puzzling problem with ZIC-HILIC and ZIC-cHILIC columns.Merck has two application notes for this:

on the HILIC column they suggest running isocratic 70:30 ACN/ammonium acetate pH 4 100 mM
on the cHILIC they suggest isocratic 70:30 ACN/ammonium acetate pH 4.5 100 mM

We purchased one of each (100 A pore x 4.6 x 150 mm x 3.5 or 3 µm respectively), and one appropriate guard column for each (2.1 x 50 mm x 3.5 or 3 µm). For the first use, we primed our system with the solvents of the application note, connected the guard, flushed for 20 mins at 1 ml/min, connected the analytical column through the column coupler that comes with the guard, and ran a series of standards diluted in mobile phase at a flow of 2 ml/min. Mobile phase was prepared with freshly opened HPLC grade water and ACN from Biosolve and HPLC grade salt and acetic acid from Sigma. The pH was checked over an aliquot taken out of the bottle to prevent introducing contamination from the probe. Solvent were visually checked for dust or particulate and no further filtering was performed.

In both cases the first injection of AMP, ADP and ATP gave excellent peak shapes and the k' described in the application, but immediately following the weird problems started...

In the case of the HILIC column, peak shape started to deteriorate rapidly. The peak tailing became so bad that after 5 injection ATP was no longer detectable. Therefore we decided to rinse the column and the guard column separately, following the procedure described in the booklet. Rinse was 20 volumes of 1ml/min water, followed by 20 volumes NaCl 0.5 M, and finally 40 volumes water. This did not help, as afterwards the peak shape was even worse. We run the samples with the guard column, and without. The result was the same. We tried to reverse the flow (0.5 ml/min, 40 volumes) and then analyze the standards again in normal direction: the peak was split in two.

In the case of the cHILIC, not only peak shape degraded, but we had weird backpressure problems. Peak tailing started to show up at the second run, and between the second and the third run the pressure got over 350 bar and the system stopped. After stopping the pump, the pressure did not drop. Therefore we proceeded to disconnect the fitting at the inlet side of the analytical colum. The pressure relieved. I thought of a misalignement problem of the fitting, and changed the column coupler with a homemade one. Peak shape was still tailing, and after two runs the backpressure problem showed up again. We changed guard column and coupler, and peak tailing was still there. We ran the column without guard, and tailing persisted. Lookign accuratley at the chromatogram, it looks more like a broad shoulder after every peak rather than tailing.

If I look at the data I have, it leads to the conclusion of a clogged frit or dislodged bead packing. This leaves me perplexed however, as I cannot recognize any step in which I did something wrong. I normally run HPLC-TOF on RP columns and I know how to prevent solvent contamination. On the same system used for the ZIC-HILIC I regularly use 2.6 µm columns and never had a problem with clogging.

Has anyone else faced the same problem?

[Andy Alpert]

Questions:

1) What solvent were your samples dissolved in? Which salt, what concentration, what pH, how much ACN?

2) Did you really use a guard cartridge with an i.d. of 2.1 mm with a column with an i.d. of 4.6 mm, and at a flow rate of 2 ml/min?

Once you provide these answers, then we can proceed to the next step in troubleshooting.

[M_Farooq]

Once you have tried the approaches mentioned by Dr. Alpert, you can open the column inlet and see if the bed has collapsed for some reasons. Peak splitting or the appearance of shoulder on every peak indicates that somewhere in the system (from guard to the column end) there is a dead volume (sort of a mixing chamber effect). Note that it requires a very tiny level of settling to produce pretty bad peak shapes either at the guard column or the analytical column.

I am surprised that both columns deteriorated so rapidly. Is 100 mM the overall concentration of the buffer? We have seen silica actually dissolving in such systems (it was not ZIC HILIC though).

[carlo.annaratone]

Thank you all for the quick response! This forum is always very useful!

Samples were dissolved ACN/water 70:30. Regarding the guard column, I used the 2.1 mm because it's the only size that Merck provides.

Regarding the salt concentration, it is 100 mM in the water. Therefore afte mixing it would be approximately 30 mM.

I thought about opening the column, but I see a few problems here. ZIC-HILIC comes packed in PEEK columns (http://www.analytics-shop.com/media/import/ZIC.jpg) that have no grip points for wrenches once they are tightened. So I fear of making more harm than good....

[Andy Alpert]

There is no wrench flat machined into the body, but there is one on each end fitting. You could try to loosen the inlet end fitting by putting a wrench on each and and turning. Of course, you do run the risk that you will then loosen the wrong end fitting...

Why did you order a column with a PEEK body instead of stainless steel? If this column uses porous PEEK frits as well, then that might be the source of some of your problems with the backpressure.

Getting back to your response:

1) You should not use a guard cartridge with an i.d. of 2.1 mm with a 4.6-mm i.d. column, especially at a flow rate of 2 ml/min (which is about 10x faster than one usually uses with an i.d. of 2.1 mm). I am incredulous that Merck SeQuant doesn't offer any other option. I recommend that you not use a guard cartridge at all. That will avoid a lot of problems with backpressure. If you feel that you must have a guard cartridge, then buy a second regular column, open up one end and extrude the packing material, and have someone pack it into a guard cartridge of the correct i.d.

2) The higher the charge-to-mass ratio of an analyte is, the higher the concentration of salt needed in the mobile phase and sample solvent to insure that all of the ionizable groups share the same counterion. Otherwise, some of them may swap counterions during the course of the chromatography. Different ion-counterion pairs will differ in polarity, resulting in a variety of retention times in HILIC. Result: Skewed peaks or even multiple peaks for a single analyte. ADP has a pretty high charge-to-mass ratio, and ATP has a very high one. Make sure that your sample solvent contains the same salt that your starting mobile phase does, and at the same concentration.

3) Farooq's comments are cogent. It would not be implausible that both the ZIC-HILIC and the cZIC-HILIC columns developed voids quickly while in normal use, in view of your observations.

[carlo.annaratone]

To Andy: believe it or not, Merck is only offering the HILIC columns in PEEK, and the guard columns in 2.1 mm format. I do not think that the problem is the salt in the sample, because I have the same peak shape injecting a test mix made of toluene, uracil and cytosine. I think that the problem might be the high flow through the guard... something that the company should warn you about, because we wasted two good column in one week! Once the guys from SeQuant seemed to be around on this forum pretty often, but they have bought by Merck and things have changed.

[Andy Alpert]

If I may ask without being flamed for crass commercialism, have you considered trying a HILIC column from some more accommodating company?

[carlo.annaratone]

I thought about your company, but we decided to go for only two columns, and it was easier to sell to the boss two columns with a lot of literature. Maybe you can write me in PM

[Andy Alpert]

Send me an e-mail and we'll continue this conversation off-list.

[carlo.annaratone]

I sent you an email, waiting for your answer

[carlo.annaratone]

Just an update:

I contacted some other researchers that have used ZIC-HILIC columns. While no one had a catastrophic failure as me, all of them had issues with short lifetime. Therefore I have been suggested to use NUCLEODUR HILIC columns from Macherey Nagel.

[Patrik Appelblad]

Dear Carlo,
I have no problem to discuss this matter in this forum, however I wish you would have contacted our application support team before you started this thread.
I think it would have been more appropriate if you had contacted our company first before you went public with your comments.
Myself I used to post frequently at this forum until a couple of years ago. Nowadays I spend the time making sure we provide good support to our customers via our global support hotline - chromatography(at)merckgroup.com. That said I was informed earlier this week about this thread.

I am sorry to hear your frustration about the perceived performance about the ZIC-HILIC and ZIC-cHILIC columns you purchased and used.
First observation/question I have is if you use a premixed isocratic mobile phase (as used in the two application notes you refer to) or used online mixing with a mobile phase A with 100% acetonitrile and a mobile phase B with 100% 100 mM ammonium acetate. If you have used the second approach you should be aware that precipitation can easily happen (depending on type of LC system and type of gradient mixing - low pressure vs high pressure) if you mix 100% acetonitrile and a high concentration buffer. The [u]total ionic strength in the isocratic mobile phase used in our application notes are 30 mM [/u]and this WILL not destroy any column material as suggested in another post.

Now what I want to focus on is the results you got, good initial results. Thus both columns was working as intended when u first started working with them and you were able to reproduce our application notes, but then you experience drift in retention, difference in peak shape and increasing pressure ending up at 350 bar (a pressure very hard to reach in HILIC mode unless you get precipitation and this is what I believe has happened. When you have precipitation the pressure build up is slow in the beginning and then is goes quickly as you described. This would also sometime lead to splitting of peaks.

You obviously read the column inlet with the regeneration protocol although the proper procedure is as follow
30 column volumes of Milli-Q (in this case this means 75 mL since column tube volume of a 150x4.6 is approx. 2.5 mL), and this removes all residual organic solvent
30 column volumes of 0.5 M NaCl (75 mL 0.5 M NaCl is a strong eluting solution that will elute all highly retained analytes and matrix)
30 column volumes of Milli-Q (another 75 mL to remove all sodium chloride before you apply solvent again....or you will have precipitation of sodium chloride in the column)

Andy's comments about samples dissolution and type of salt, what concentration, what pH, how much ACN is very valid.
I would add a question about what autosampler wash solution have you used. An ill-chosen was-solution can also be a reason for precipitation.
We provide recommendation in both our HILIC guide and in the column use leaflet that ship with all columns.

The ZIC-HILIC and ZIC-CHILIC columns we produce are with PEEK material and these are pressure stable to 350 bar so this is no problem, and as said before a pressure you very seldom would reach if columns are used properly. We use PEEK bodies and porous PEEK frits, instead of stainless steel of several reasons such as better biocompatibility and no active surfaces that would interact with delicate compounds such as phosphorylated analytes.

The reason why we use a 20x2.1 (not 50x2.1 mm as you mentioned) for our 4.6 mm id columns is simply of void volume reasons.
A guard should maximum provide 5% extra volume, and we have chosen to use a smaller id (this gives even lower volume) since the guards are packed with larger particles (it is 5 micron - not 3 micron) and all together this provide a lesser impact on reproducibility when changing the guards (after all a guard is a more expensive filter and must be changed repeatedly).Despite Andy claim that it is not advisable to use such configuration I think the opposite. A short guard column with a larger particle is only negligible contributing to the total system pressure even at 2 mL/min in HILIC mode.

Finally I just want to make a comments about your last posting about other researchers using these columns with negative or less good results but with no reference. Just name dropping and at the same time promoting another brand....
To all other readers of this post I would just suggest that you look at all papers that have been published with the use of ZIC-HILIC and ZIC-cHILIC columns. Perhaps this gives a more neutral comment. A column can last very long, and opposite is true also. It all depend on how we apply the experimental conditions and what type of samples we analyze. Thus it is impossible to give a "injection number guarantee".
If you really want to understand the difference among all HILIC columns commercially available, please study the work of professor Ikegami and others.
We all work in and for the scientific community, and it is free to everyone to use any product. We contribute with our experiences at forums like this and we should try to make as neutral comments as possible. Either as you, as researcher at a university or as me and Andy as representatives of column manufacturers.

I believe you got precipitation in both columns and this blocked the inlet frit. The wash did not resolve the problem, since the columns went over maximum pressure limit and maybe also the column bed got compressed leading to a void. The results from testing the column according to the certificate of analysis with toluene, uracil and cytosine indicate this.

I have been working with the ZIC-HILIC product family since introduction in 2002 in different stages and in different roles, and if you have any questions further in this matter we can certainly discuss here or you can contact me directly at patrik.appelblad(at)merckgroup.com.

[Gerhard Kratz]

Excellent comment by Patrick. It is always recommended to contact in first line the manufacturer and get some advice from the technical support group.

[carlo.annaratone]

Hi Patrick,

thank you for your response. First of all I must say that I chose the two ZIC columns for method development because of the large amount of literature available.

I purchased in total 2 ZIC-HILIC columns (1 x 4.6 mm and 1 x 2.1 mm) and 1 ZIC-cHILIC and all three of them failed before they could deliver results. On the system on which I ran the 2.1 mm column I run without problems many different RP columns, and on the system on which I ran the 4.6 mm column I regularly run 2.6µm RP columns and 3.5µm amino column in HILIC and IEX mode.

The 2.1 mm column failed because of a defective guard column frit that leaked silica slurry into the main column frit. For this accident I contacted the Merck technical support through the local dealer here in Belgium, and I got fully refunded. However it took more than four months to get a replacement (complaint sent to local dealer on january 17, package arrived around April 25)

Regarding the 4.6 mm columns: before looking for technical assistance, I checked for solubility/precipitation problems (the eluents were mixed online on a low pressure mixer, but I tried the mixing on a small amount of solvent in a beaker, no precipitation or cloudiness detected even after one hour). The Agilent system we use also has no needle wash thus there should be no issue with needle wash precipitation. I will add as well that we used the same solvents we use for 2.6µm particles, and never had any clogging with these.
Then I tried to contact directly Merck through the chromatography support web form on the site. The complaint was sent on April 24 (I have the confirmation email) and never followed up. I learn only now that there is an email address for chromatography support I can write to, to me it is very difficult to navigate and find what one is looking on the huge Merck website.

Therefore I posted on this forum because I was unsatisfied of the technical support from Merck, and I wanted to hear from "the crowd" if someone else had my same experience and managed to solve the problem. I also contacted four researchers who had used ZIC HILIC guard columns: two of the responses mentioned problems with selectivity change after a couple of months of storage (under the conditions specified on the instructions), one of them shared my impression of a slow-moving customer service, and the remaining two mentioned similar performance and better lifetime with another brand. All these people managed to get the column working long enough to get the results on a journal, but definitely were not satisfied of long-term performance. I do know that column lifetime is not infinite, but I would expect column performance to deteriorate more gradually with storage and sample injections.

Just yesterday I opened the two 4.6 mm columns, and I found a 1.5 mm void at the head of the column. Thus it seems that the peak splitting was caused by collapse of the bead packing. I do not know if it happened because of my fault or because or manufacturing defect. However as already said I ran the same solvents and buffers (same concentrations) on an amino column in HILIC mode without any damage.

I understand that I might have been just one very unlucky customer, but for the price I paid and from a company with the good reputation of Merck I expect another level of quality control and technical support.

It has never been my intention to negatively advertise any manufacturer, but I think that my negative experience has to be shared on a forum. This kind of issues never find their place in the discussion paragraph of scientific papers, yet reliability of hardware and quality of technical support are paramount to advancement of scientific research. I must admit that I am very disappointed of the technical service from Merck. While, of course, I am happy that you answered me on this forum, I believe that my complaint had to be handled earlier and through more appropriate channels.

[richiekichi]

Hi Carlo,

I have experienced similar issues with Waters Atlantis HILIC columns. Horrific peak tailing and increasing backpressure after a relatively short time.

I won't go into details but I was also using an application note provided by the manufacturer although running this method on an Agilent 1100.

This method method now runs on a Waters Acquity system with an Acquity HILIC column. It runs very well.

These things happen.