HPLC Pump selection for RI Detector

[h.piatkowski@comalc.com]

Hi folks,

I'm looking at purchasing an Agilent 1100 Unit with an RI detector since I do fermentation carbohydrate analysis. The question deals with the pump. I have a system currently with an isocratic pump and it works fairly well.. but the baseline does have small bumps in it which I'm assuming are due to the nature of the pump. If I was to get a binary or quaternary pump would this smooth out the baseline? Is it worth the investement? The binary is $10,000 more than that isocratic..

Also, if anyone is doing fermentation analysis I would love to chat with you on your setup and how well it works.

Thanks for any ideas..

- Hubert Piatkowski

[Kostas Petritis]

I would suggest at least a binary pump for the flexibility it offers in method development. Even if you have to use it isocratically with your RI detector remember that you will have to wash in general as you are analyzing a complex matrice.

Maybe I am missing something but I do not think that the small bumps you observe have nothing to do with the specifics of a isocratic vs. a gradient pump. I guess that the bumps comes from pump pulsation that can be corrected.

[h.piatkowski@comalc.com]

Maybe I am missing something but I do not think that the small bumps you observe have nothing to do with the specifics of a isocratic vs. a gradient pump. I guess that the bumps comes from pump pulsation that can be corrected.

Thanks for the post..

I was always under the impression that because of the design of the isocratic pump, there would always be small pulsations and as such, you would see some of these as baseline fluctuations and little speedbumps on my printout.

We are not getting a large amount of these on our current system (we have an 1100 with an isocratic pump now) and as such I would find it hard to justify the $ 10K price increase for a binary pump if this is the only benefit I will see.

In regards to gradients and solvent mixing, as far as I know, I can't do that with an RI detector because of the nature of the beast, and since I'm working with carbohydrates, I'm kind of stuck with RI for now.

Does that make sense?

- Hubert

[Mark Tracy]

A quaternary pump is much less costly than a binary because it uses a proportioning valve instead of a second pump mechanism. It also offers the flexibility of changing mobile phases for washing, etc. Besides, a good LC pump will last ten years. Will it still be doing carbohydrates in 2015? If you are reasonably sure that you will wear out this pump doing carbohydrates, you probably can get by with the isocratic model.

On the subject of the small bumps in the baseline, are they regularly spaced? If so, does it match the cycle of your pump?

[h.piatkowski@comalc.com]

A quaternary pump is much less costly than a binary because it uses a proportioning valve instead of a second pump mechanism..... Will it still be doing carbohydrates in 2015? If you are reasonably sure that you will wear out this pump doing carbohydrates, you probably can get by with the isocratic model.

On the subject of the small bumps in the baseline, are they regularly spaced? If so, does it match the cycle of your pump?

Thanks for the response Mark,

That is an interesting line.. did you mean an isocratic pump is cheaper and not a quaternary as you stated.. I would think that quaternary would be the most expensive.

In regards to the testing.. this is what we do.. the HPLC analysis are on fermentation mash from an ethanol plant, so as long as there is need for ethanol this is what we will be testing.. hopefully past 2015 !!!

In regards to the bumps.. they don't show up all the time but when they do, they are not usually regurarly spaced.. and as such it would probably indicate a different problem I think, any idea what this could be?

Thanks again for all the help, I'm fairly new to HPLC and am learning as much as I can, so I can make the right decision on the next purchase.

- Hubert

[Mark Tracy]

The isocratic is the least expensive option. The binary pump could be one of two designs, but most commonly (and Agilent does it this way) binary pumps have two separate high-pressure pump mechanisms and combine the flows. This is the so-called "high-pressure mixing" gradient pump design. This design is the most expensive. A quaternary design (usually) uses a single pump mechanism with a 4-way proportioning valve at its inlet. This is the "low-pressure mixing" gradient pump design, and is somewhat more expensive than the isocratic pump, but much less expensive than the high-pressure mixing style. Some vendors offer all three options.

In your specific case, isocratic is probably sufficient.

Irregular bumps in the baseline can come from many sources.
- Dirt washing off the column after a previous injection (common)
- Variation in room temperature
- Air bubbles in the pump
- Air bubbles or particles in the detector flow-cell
- An unreliable checkvalve in the pump
- Problems with the electricity or grounding (rare)
- Radio frequency interference (rare)
- Worn out lamp in the detector
- Leaks

[Kostas Petritis]

Mark gave a comprehensive reply on the irregular bumps issue.

About the choice of the pump, it will depend on how many injections you can make isocratically before you have to clean the column. With an isocratic pump you will have to change the solvent, purge, clean the column, change the solvent again, purge, equilibrate, do your analysis. If you can make several injections without a need to clean your column (i.e. no retention time shifts, no baseline drifts etc.) then you can buy an isocratic pump (assuming that you will do the same analysis with the same detector -maybe you will consider to upgrade to an ELSD-). Otherwise a binary or quaternary pump would be a better choice.

[Victor]

Mark,

You are wrong about the Agilent in that they make a high pressure and low pressure (quarternary) mixing system; maybe they do use the jargon "binary" pump to refer to their high pressure mixing system but of course you can use their quaternary pump for binary solvent mixtures and gradients. If I remember, there is not much price difference between the two systems. This may reflect sharp business practice, rather than scientific reality, but it makes little difference to the purchaser. Maybe this is not true for other manufacturers or different designs.

There are many quite complex arguments for/against high/low pressure mixing systems. Those that you see in textbooks may have become very dated by new developments e.g. MAYBE the old argument that a high pressure mixing system could not deliver 99% A 1% B accurately no longer holds (requires pump B to deliver 0.01 mls/min for 1 ml/min total flow). It would be nice to see these arguments again.

I presume from what you say that Dionex weighed up the arguments and decided on LP mixing for their applications. However, this is not in my opinion a universal truth. Maybe I got it wrong-maybe Dionex has a HP mixing system also?

[JesseG]

The isocratic and quatanary pumps for Agailent are the same platform , they add a valve and you can have an iso upgraded to a quat by Agilent ( I have 2 and have done this with one unit). The binary is a completly different design.
I also use a RI with my Agilent systems and have had numerous baseline issues. One problem comes from the micro-degasser. The other problem with your baseline may just be environmental.

Hope this helps.

[Mark Tracy]

Victor,
If you want to talk about what Dionex offers, instead of guessing, please check our website.

[h.piatkowski@comalc.com]

Thanks again for all the info.. this forum is awesome..

One more quick question... I use 0.01 N H2SO4 as my eluant using an RI detector, no gradiants, no mixing, etc.. nice and simple...

What's the best way to degass the eluant? I currently filter it through a vacum filter and after that it goes to the Agilent 1100 degasser and that's all.. should I do something else? What works to remove the baseline bumps?

Thanks for all the help everyone..

- Hubert

[Mark Tracy]

Your degassing should be sufficient.

One thing you can check is the % ripple. Your pump should have <2% ripple, and is capable of <0.5%. You can also collect pump pressure as data (just like a detector signal) and look for spikes or dips.

[h.piatkowski@comalc.com]

Your degassing should be sufficient.

One thing you can check is the % ripple. Your pump should have <2% ripple, and is capable of <0.5%. You can also collect pump pressure as data (just like a detector signal) and look for spikes or dips.

How would someone inexperianced in ChemStation do something like that.. sounds like good info to have a look at..??

- Hubert

[Mark Tracy]

I'm sorry, but its been 5 years since I last used Agilent HPLC systems, and so much has changed that I can't give specific advice. Your best bet is to spend an afternoon with the operator's manuals in hand. There are some ChemStation wizards on this forum that will be happy to help you, or you can ask your question in the Data Systems section.

[h.piatkowski@comalc.com]

Got an answer from JesseG, it's right on the main real-time plot.. just had to choose it..

- Hubert