Performance of 2.7 um vs 3.0/3.5 Columns as f(flowrate)?

[indium]

Has anybody done a comparison of 2.7 um fused core columns with similar phase 3.0/3.5 um columns in terms of separation of a real world mixture as a function of flow rate on a 400 bar LC system? Although I haven't done any formal systematic studies, my experience has been that an Ascentis Express 150 x 3.0 mm column did not perform significantly better than a Zorbax 150 x 3.0 mm, 3.5 um column at 40C on a 400 bar system from 0.4 to 0.8 mL/min. I cannot go much higher than 0.8 mL/min at 40C. Perhaps someone has done a study with shorter columns, e.g. Ascentis 100 x 3.0mm versus an Ascentis Express 100x3.0mm to a higher flow rate? I guess another way of putting this would be at what flow rate would the 2.7 um and 3.5 um column perfomance really start to diverge?

[Bryan Evans]

Hi indium -

Your results are on par with our findings. All the promotional material
for 2um, 2.7um, ect. (porous material) is done with isocratic elution -
using a neutral solute. Van Deemter and kinetic plots only tell a one part of the story.

In the real world - we deal with samples from all sorts of matrixes,
solutes of all different types (i.e. ionic, small, large, ect).
Gradient elution is often used for complex mixtures / matrixes.

When you put it all together, you find that what really matters is (in no particular order):

- column permeability (i.e. resulting system pressure)
- column selectivity (i.e. retention, spacing between bands, peak shape, ect.)
- column efficiency
- pH durability
- column lifetime
- batch to batch reproducibility

My suggestion: Look at the big picture - and then use the best column chemistry (whatever the particle size)
to solve your problem.

Little known fact: VD for a well packed 3um particle column is flat.

[Wayne Way]

As Bryan stated, there are many factors to consider when choosing a column.

Some parameters that may be keeping you from getting better performance on the Ascentis Express Fused-Core than the porous 3.5 micron include.

1) Injection volume too large
2) Injection Solvent too strong
3) System Dispersion too high
4) Electronics set too slow
5) Retention not optimized
6) Mobile phase conditions not optimized

First and foremost, the method developer has to decide (or know) what the method needs to do and optimize those parameters.

[danko]

Wayne Way,

I wonder why the parameters you mentioned, although in general, relevant enough, will not apply to the first column, whereas it seams to me you find them important in the context of the second column, mentioned by Indium. If they’re kept constant then they can be ignored in a comparative study. Don’t you think?
So, I think (maybe I’m wrong) it is a perfectly valid question, namely whether fused core based stationary phase performs much better than the traditional particle based column material.
I’m not particularly advocating for any of them, but I’m still interested in acquiring some impartial real-life experience of these particles’ advantages and disadvantages.

Your thoughts and suggestions are still of utmost importance – no doubt about it.

Best Regards

[Wayne Way]

The method on either of the two columns may not be optimized. My points were not specifically about just the Ascentis Express. But a poorly optimized method will effect the better performing column to a greater extent.

I play guitar. I have an amplifier that costs $2,000 (good amp). I also have practice amps that cost $175. I can notice much more fine detail in my playing and fine detail in the tone of the guitars through the good amp. Through the practice amp, the fine detail is blurred away.

Hope that helps.

[danko]

The initial post is more like this: You’re testing two different guitars with the same amp. So, can you hear a difference in the sound without changing anything else but the guitars?

BR

[Bryan Evans]

How about this:

We have 2 different guitars
(1) "High performance guitar"
(2) "Regular guitar"

The "Higher performance guitar" is usually played on the optimized (low dispersion) amp.
The "Regular Guitar" is usually played on the standard (large dead volume) amp.

The argument is that both guitars sound much better on the optimzed amp.

[danko]

Yes, but the question wasn't about the aplifiers (systems) but the guitars (columns).
I still remember what I leart in school: Don't vary more than one parameter at the time

BR

[Wayne Way]

Danko,

You are right about not changing more than one variable at a time. Also, you need to have a system of measurement that can detect the differences.

It's like me asking you to tell the difference between (Television) high definition programming and regular programming on a TV without hi def.

The programming comes into the TV with different qualities, but the TV only broadcasts in the low quality.

[mbicking]

I hate to retreat from this exciting guitar discussion back to chromatography ...

I have used the Express columns and think they are in fact more efficient than the standard 3 um particles, on an equal length basis. Plate counts were higher and reduced plate heights were lower than on a standard column. But, as Wayne mentioned, I had to pay attention to operating parameters.

In my case, I increased the flow to keep the reduced velocity the same for the smaller particles. I also had to change the "Peak Width" setting on my (Agilent) detector, to make sure that the detector response time did not distort the peaks. The system also had a smaller flow cell (5 uL) than the standard.

I will see if I have any apples-to-apples (or in this case, nicer guitar to nice guitar) comparisons and post them here.

[mbicking]

I was able to find the comparison data fairly quickly and summarize below. The data are for one flow rate only, but given the flat nature of VD curves for these particle sizes I do not expect any significant changes vs. flow rate. However, instrumental issues can be a factor, as noted above. Experiments were performed on the same system on the same day using the same standard (at 20 ug/mL, 25 C).

Columns: Ascentis Silica, 4.6 X 50 mm, 3 um vs. Ascentis Express HILIC, 4.6 X 75 mm, 2.7 um (fused core).

Mobile phase: 80/20 acetonitrile/10 mM ammonium formate

Flow rate: 1.67 vs. 1.85 mL/min.

Injection volume: 5 vs. 2 uL

Detector peak width: >0.03 vs. >0.01 min. (Agilent 1100 diode array at 210 nm)

Col. Name, RT, Plates, Plates/m
3 um Nicotinic Acid, 0.721, 4054, 81080
2.7 um Nicotinic Acid, 0.665, 10887, 145160

3 um Creatinine, 1.046, 5451, 109020
2.7 um Creatinine, 0.921, 13180, 175733

3 um Protriptylene, 1.334, 3944, 78880
2.7 um Protriptylene, 1.087, 13415, 178867

3 um Procainamide, 2.961, 5780, 115600
2.7 um Procainamide, 2.549, 12655, 168733




I have less information for a C18 column, but the trends appeared to be the same. To me it looks like there is an advantage to using the fused core materials vs. a standard 3 um particle. If you have a fast detector, you can push the flow even more. I can send the ppt file to anyone who is interested; this is from a Pitton 2009 poster.

[danko]

So, it was possible to compare these two types of particles after all

Merlin, beautiful data. Thanks for sharing.
Have an idea of what co-elutes (or rather elutes on the tail of) with the peak at approx.1.05 min (first chrom.)? You think it’s actual peak/compound that’s partially separated or…….?

Best Regards

[mbicking]

Danko:
Not sure about the shoulder. Would have to look at more data - could be impurity or other artifact. I remember some issues during the project but it's been a while. I don't think it is a simple peak shape problem.

Also, this was in HILIC mode and this may be a bit more favorable to the fused core design, since some people think that the HILIC partitioning is a little slower. However, as I mentioned, the C18 version did appear to be significantly better as well, but retention and peak shape were not good for these analytes.

[Uwe Neue]

I am not terribly impressed. This is an apples to organges - no, grapefruit - comparison. A 5 cm column with 3 micron particles compared to a 50% longer columns with the smaller particles. This does not tell you anything about the true performance of superficially porous particles...

Don't misunderstand me - in a proper comparison of superficially porous 2.7 micron particles and fully porous 3 micron particles the superficially porous stuff should win, but this example here is not a proper comparison...

[mbicking]

Well, Uwe, I don't work for a column manufacturer so I don't have infinite access to any column I want in any configuration I want. I have to rely on what people are willing to send to me for evaluation and joint research. Or, I have to buy it with my own company's money (which has considerably less than Waters does).

Yes, I wish I had a better matched example, but it was what I had on hand and could put together in a short time to answer the question. If you read the table you will see that I did include numbers for plates per meter, which should normalize the comparison. Would you rather I include reduced plate height? The results will be the same. As always, you are welcome to provide your own data if you don't believe mine.

While you're at it, why not complain that I used an old LC system (which is what I can afford) rather than one of those shiny new UPLC systems that you can use for free from your employer.