UPLC --- What's the Success Rate???

Discussions about HPLC, CE, TLC, SFC, and other "liquid phase" separation techniques.

13 posts Page 1 of 1
OK so let me clarify the question, because I'm sure it's not clear from the Subject Line. We all know that as we reduce the particle diameter, in an HPLC column, we should - in theory - get sharper peaks. And, of course, UPLC is all about building instruments that can "take the pressure" so as to allow the use of particle diameters even below 2 um.

But I think we all know, as well, that there are any number of 'Extra-column Factors' that can prevent one from obtaining the full efficiency (or 'peak sharpness') that we should get in theory.

So my interest here is to take a poll, asking the basic question: What extent of the full benefit have people generally observed? My sense is that when we go below 3 um we will typically not see the full, theoretical, benefit. In fact, my sense is that it may usually be closer to 50% of this. I have even had one or two cases where I went from 5 um to 3 um (to better separate some closely resolved peak), and in fact, the peaks were not any sharper with 3 um column (which, of course, they should be from theory).

Now one last thing: We are very interested to hear everyone's feedback. But, when responding, please steer clear of "Explanatory Answers" such as explaining how the connecting tubing should not be too long and should have a narrow ID, and the injection volume.......etc etc......I think most people know these things by now. What we are looking for is people's real world observations....."I went to a smaller diameter particle and......I got all the expected benefit......or none of it......or half of it......".

Thanks very much in advance.
10 years ago my company only had 2 UPLCs and the rest HPLC. Nowadays we have only 4 HPLCs and 25 UPLC/H-Class and the column particle size is 90% sub 2 micron, as in most methods use 1.7micron particles and typical column length is 50-100mm.

Pros: Time being the main one. Can run three times the amount of samples in only a fraction of the time.
Not that hard to develop and validate methods on and the column manager is a great tool.
All system start ups and priming can be done from console, very handy for end user.
Uses much less solvent, less waste and raw materials used.

Cons: Waters make the binary UPLCs a lot more awkward in terms of parts than it needs to be- try changing a check valve on the older models, you have all manners of wires and connections in your way, very user-unfriendly and gives them an excuse to keep coming in.
Parts on UPLC and H Class not as robust as old Alliance HPLCs- a lot of parts and if even one part is not washed/primed out properly etc then it can lead to an instrument failure.
Steep learning curve going from HPLC to UPLC.
I agree with Empowersbane. In addition I would add 2 pros to the list;

1. Shorter run time which means faster turn around for the lot (with ALL the benefits).

2. Use of less solvent than a regular HPLC.

There are some chromatographic separations that cannot be made with UPLC/UHPLC.
Hi Jimi,

How are we to define the "full benefit" of UHPLC (to remove controversy regarding the use of the marketing acronym "UPLC")?

The previous contributors to this thread make good, salient observations, I think, that have only a small amount to do only with the sharpness of peaks. That said, I agree completely with what EmpowersBane and HPLC Chemist both say--these are all benefits (and some drawbacks) that are relevant to UHPLC.

As noted in Chapter 1 of The HPLC Expert II: Find and Optimize the Benefits of your HPLC / UHPLC, "The fact is that the dead volume of every UHPLC system on the market is much too large for a 50mm× 2.1 mm, ≤2 μm column or smaller. Depending on the manufacturer, the loss of efficiency is around about 20–40%, and this is openly admitted by most manufacturers." This is certainly true for isocratic separations--and not so much for gradient separations where the large majority of the time peaks are sharp, anyway.

In principle are not narrow peaks obtained (gradient separation considered) when
a steep gradient program, a high start and also end % B, small particles, low eluent viscosity, and high temperature employed?

Just seems to me that there is much more to UHPLC than Sharpness of Peaks.
MattM
If you want the very highest resolution, small particles are just wrong. Go for a very long column of big particles. UHPLC is about rapid separation. For that it's great; it's revolutionised our lab. We used to have loads of 45 minute methods, now we have a lot of 5 minutes methods.
lmh wrote:
If you want the very highest resolution, small particles are just wrong. Go for a very long column of big particles. UHPLC is about rapid separation. For that it's great; it's revolutionised our lab. We used to have loads of 45 minute methods, now we have a lot of 5 minutes methods.


::Disagrees, especially when it comes to isocratic methods::

Ways to get around the issues that accompany extra large columns include core-shell technology, gradients.

I agree with the H-class comments but wish to add that the reason behind the issues with it is that it is trying to be all things to all liquid chromatographers.
The way to get more robustness and reliability is to buy HPLCs for "conventional" runs and buy UHPLCs for use with <3µ particles.

If all you have is a lower pressure system and want to trade a little capacity for much sharper peaks, try a narrower version of your current column. Not enough? couple this with a core-shell version of your stationary phase and/or smaller particles. This all assumes that 1) you've got all of your extra-column effect producing tubing issues sorted as best you can and 2) you're aware that when you change column diameter and particle size, your optimal flow rate will also change to maintain optimal linear velocity. 3) that you're also aware that with smaller stationary phase particles, you need to adjust your sample filters downward concomitantly.
Thanks,
DR
Image
@ DR,

I must agree with most of your disagrees--I'll clarify the point regarding isocratic methods by adding in; when using small ID, short (50mm or less in length) columns in combination with isocratic methods, up to 40% of the theoretically-possible peak efficiency is lost to a combination of column packing efficiency and UHPLC system dead volume. [Particularly with early-eluting peaks in the separation.] This was confirmed by Guichon and Gritti, with superficially-porous stationary phases/columns of varied IDs and lengths, a while ago.

And I'll note...this is a special case, as this is a time when gradient separations seem to be employed ca. 4 times out of five separation methods that are carried out--at least, where I've been.

This statement:
...the reason behind the issues with it (UHPLC) is that it is trying to be all things to all liquid chromatographers.
I agree with every day of the week and twice on Sunday.

EDIT: Regarding the packing of UHPLC columns,

http://www.chromatographyonline.com/und ... ntals-chal
MattM
DR, I also very much like core-shell particles. In fact nearly everything I use is a core-shell particle, which I regard as UHPLC-at-low-pressure (because core-shell also requires a reasonably optimised system if it's to create the best peaks, and because it's not either/or; there is no reason not to combine core-shell with high pressures if you want).

But my argument for long columns of big particles is a matter of physics, not my personal opinion. If you double the particle size you reduce the back pressure by a factor of 4, but the plate-number-per-length by only a factor of 2, which means that for any given pump, be it UHPLC or conventional HPLC, you can quadruple the maximum length of column that can be run and double the plate-number. Of course the root-2 improvement in resolution comes with a 4-fold increase in column-volume, reequilibration time is enormous, and runs get veeerrry slow (welcome to the wonderful world of proteomics...).

Core-shell particles are obviously lovely because they allow us one notch less pressure for the same plate count, but for a given core-shell material, achievable resolution varies with particle size in the same way as it would for a fully-porous material.

All of this argument depends on the assumption that the limiting factor is pump-pressure, not the optimum linear flow rate, but if you're pursuing the very highest resolution, you will want to look at the highest-plate-count-column that you can achieve with your pump (which means the longest, or best-particle-size, consistent with not deviating too far from the optimum point on the van Deemter curve).

Most of us, I suspect, want adequate resolution as quickly as possible, not best-possible-resolution irrespective of time available. The problem I now have with core-shell 2.6u particles and a UHPLC pump is that I can get decent resolution out of 2mm diameter columns and get my run-length down by increasing flow (modern columns cope with excessive linear velocities with little loss of resolution) - but now I'm finding I could easily use 0.8mL/min or above, while in LC-MS systems, most MS would really prefer to be working at not more than 0.6mL/min - so should I be risking 1mm diameter columns? Or will I regret the loss of resolution caused by going down to 200uL/min, and the stresses of tiny injection volumes. I really don't know.
lmh wrote:
DR, ...you can quadruple the maximum length of column that can be run and double the plate-number. Of course the root-2 improvement in resolution comes with a 4-fold increase in column-volume, reequilibration time is enormous, and runs get veeerrry slow (welcome to the wonderful world of proteomics...). ...Or will I regret the loss of resolution caused by going down to 200uL/min, and the stresses of tiny injection volumes. I really don't know.


I think the bit that raised my hackles initially was the implication that you were cutting run times by increasing particle size and column length. Apologies if I just misunderstood that portion of your original post.

I agree with your assessment of the physics of the situation as well as the time requirements.

Especially in isocratic situations, I suspect that at 200µL/min, you would miss resolution as the drawback to larger particles is that little portion of the resolution equation that addresses the band broadening due to tortuosity within the particles. It adds up, and would be potentiated by the (far) less than optimal linear velocity for larger particles, I think you'd be sunk.
Thanks,
DR
Image
Thanks! I think we're basically in agreement, I just wrote at such length that the message got lost.

My main point was that for me personally, the major benefit of small particles/UHPLC (and also of core-shell) has been the shortening of run times rather than any change in resolution. If I can get similar resolution in 5 minutes (or less) to what I used to get in 30 minutes (or more), then I am already a very happy analyst.

It's easy to forget how dramatic this is, but in some situations it's the equivalent of a mass spec manufacturer turning up and saying "You know that triple quad we wanted to sell you for £210K? Well, it's now available for £35K."
OK so I am the Wise Guy that started this thread. And I appreciate all the activity. But to quote one comment above (or actually to paraphrase it) "There are many salient observations being made, but not all are focused on the sharpness of peaks" which I think is the real key.

And from the same comment above, the following statement was referenced, from a text book: "The fact is that the dead volume of every UHPLC system on the market is much too large for a 50mm× 2.1 mm, ≤2 μm column or smaller. And this is the key to what I am ultimately trying to get at. I know that some will say that modern UHPLC systems are designed to minimize these extra-column effects. And there is certainly validity to that argument. But when you think of all the things working against you: bandbroadening in the connecting tubing, detector, fittings, secondary interaction with the column, radial temperature gradients (heck, Dave McCalley did a whole series of papers showing that one generally sees overloading at much lower levels than what we would usually expect -- for charged analytes). When you consider all of this stuff that is working against you, I just find it hard to believe that one will observe all the expected benefit. And when I see instrument companies do demos (in other words, take a conventional HPLC method from a customer, and make a UHPLC method) invariably I find that they've changed other things like making the gradient faster.

So any additional thought on this? The underlined statement being the real key. Bottom line: when you reduce dp, are you peak widths narrowing as much as you would hope. I tend to doubt it. But I'm happy to be proven wrong.
Phenomenex tested it in the context of their solid-core columns, showing plate-count (which is directly related to peak-width) as a function of particle-size for conventional fully porous and solid-core (Kinetex) particles in the range 1.7u to 10u. I saw the data in a talk by one of their reps a few years ago and have no idea where to find it now - you could ask your local Phenomenex rep if they can find it. They did the test on some real-world LC systems, although they were happy to admit they'd chosen top-end analytical systems.

The upshot was that small particles did give improved resolution, with solid-core consistently behaving like particles one size smaller, until they reached the situation of 1.7u solid core, where the benefit was small. They ascribed this to exactly the problem you describe: that the extra-column effects in even the best systems at the time of testing were negating the theoretical benefit of a particle that should give resolution for a sub 1.7u particle size. This is pretty extreme: I personally know only one person who's attempted to use a 1.7u solid core column.

But I'm utterly convinced from personal experience that 1.7u fully porous columns such as Waters Acquity genuinely outperform 2.6-3u columns on real-world systems (Acquity or Nexera), and my feeling is that 1.7u Acquity fully porous has similar resolution to a 2.6u Kinetex solid core (for which reason I tend to go with 2.6u solid-core; why use a higher pressure than you have to?).

As I say, though, I generally don't aim to use them to get narrower peak-widths. I use them to get the same peak-widths but with shorter run-times, which is directly equivalent because it frequently involves a shorter column or running the column at a less optimal flow-rate, so I need the small/solid-core particles to keep the plate-count.
I think Jimi may be, at least partially, right. I have also seen cases in our lab where going to a smaller particle size did not give us sharper peaks, even when we were still well above 2 um.

I think to get the full theoretical benefit of sub-2 um is probably not common. Any number of factors have to be well controlled.

Please don't misunderstand me. I am not throwing cold water on UHPLC and saying it has no value. It absolutely does. Rather I am just saying that one (I think) won't usually get all the benefit that the text books would suggest.
13 posts Page 1 of 1

Who is online

In total there is 1 user online :: 0 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 1117 on Mon Jan 31, 2022 2:50 pm

Users browsing this forum: No registered users and 1 guest

Latest Blog Posts from Separation Science

Separation Science offers free learning from the experts covering methods, applications, webinars, eSeminars, videos, tutorials for users of liquid chromatography, gas chromatography, mass spectrometry, sample preparation and related analytical techniques.

Subscribe to our eNewsletter with daily, weekly or monthly updates: Food & Beverage, Environmental, (Bio)Pharmaceutical, Bioclinical, Liquid Chromatography, Gas Chromatography and Mass Spectrometry.

Liquid Chromatography

Gas Chromatography

Mass Spectrometry