i.d. for c18 column 2.1 vs capilary?

[noam]

Hi,
i've recently developed a method for separation of several compounds using a c18 300A 250*4.6 column(UV detection).
I've tried using my column with formic acid (can't use TFA in M.S.) for an LCMS, with a splitter, whan F=200uL/min with elongated times and lower N.

I'm thinking about lowering the column i.d.(I think the low flow is my main problem, with 4.6 i.d.) and was wondering about the possible difference between ~2.1(narrow) and <1mm(capillary).
What's the down side of capillary, I'm mainly worried about its possible sensitivity to stuff like dead volume, fragility Etc. (using it would mean loosing the splitter, but the sample loss isn’t a problem).

[Mark Tracy]

You can usually modify a standard (4.6mm) system to operate with 2.1mm columns at a modest cost. Mostly things like replace tubing, buy a new UV flow cell, low-volume kits for autosampler and gradient pump. Fitting a conventional system for capillary LC is really not easy. For capillary LC, you will get better results from a system designed for that use.

Check the manufacturer's recommended optimum flow range for your MS interface, and choose a column for that flow. Most electrospray interfaces are optimized for flows around 0.2 mL/min. They may say that it can operate over a much wider range, but the signal will not be optimum over that whole range. For capillary work at <0.05 mL/min, you need a different interface, which may or may not be available for your MS.

[Vanquish]

Going to Microbore (or Narrowbore) 2.1mm columns is fairly simple, most HPLC systems will run 200ul/min OK. Going to MicroLC (1mm), CapillaryLC (300 / 180um ID), and then NanoLC (100 / 75um) present some more complex issues that need to be considered.

Looking at MicroLC, 1mm column would usually run at around 50ul/min. This would require a system set up specifically for this, either a dedicated system, or a standard HPLC with a splitter (pre-column). You don't specify what you are trying to separate, but I guess you are running this as a gradient system? If isocratic, the dwell time between the pump and injector is not relevent. If gradient, especially low pressure gradient, the dwell volume between the proportioning valve and the injector are critical. Take a typical LPG pump, the delay will be ~800ul. When used normally, say 1ml/min, that would result in a delay of less than 1 minute. So running a 2.1mm column, delay time would be ~4 minutes, and with 1mm columns, 16 minutes. And thats without any other tubing, and injector. Using a splitter means you can keep the pump flowrate higher, keeping these delay times minimised, within a reasonable level. The injector and loop volume need to be considered carefully with the smaller column, any loop volume is 'swept', and adds to the overall system delay. Using the correct ID tubing also makes major differences at this flow range, as well as using correct fittings.

As far as robustness is concerned, you should not be put off. The 1mm columns available are physically very strong, as long as you get your plumbing right, the sensitivity gains are well worth the effort. Your main consideration should be the amount of sample to want to put on the column. You want the best sensitivity you can get, without overloading the column. Going from 4.6 to 2.1 columns, the column capacity is reduced by a factor of approx. 8, and from 2.1 to 1mm, by the same factor. So 4.6 to 1mm results in a combined reduction in capacity by a factor of 64.

Feel free to contact me if you would like to discuss this further.

[noam]

hi, thanks for the fast answer.
Basically I'm suffering from lack of confidence problem.
I'm afraid the M.S. isn't mine to play with, so I'd be limited
With changing tubing etc.
The M.S. works with ~20uL/min, so I could just connect a capillary column, or use a 2.1 (or slightly smaller i.d.) with around 200uL/min(a flow that my splitter could handle).

The thing is, I have no experience with i.d. <4, and my main fear is the possible sensitivity to system effects(comparing 2i.d. to capillary columns).
Also I was wandering about extra fragility of the capillary columns.

[tom jupille]

If it's an isocratic method, the big issues are extra-column volume (including injection volume); the replies from Mark Tracy and Vanquish are accurate. If it's a gradient method, you may also have to scale the gradient time to keep selectivity constant, and you definitely will have to decrease the dwell volume (gradient delay volume). A splitter will take care of this, otherwise you will need appropriate hardware.

[noam]

Vanquish, yes I'm using a gradient and the delay is a problem-
If I've understood you, it's possible that I would need to add ~16min to my run time going from 4.6 to 1 i.d(that would rule out the capillary column, since it's a busy instrument).

Maybe I should try shorter column with smaller particles size?
Thank you, noam

[Vanquish]

Its not quite as simple as that. When talking about a 16 minute delay, I refer to the time it takes the gradient being formed at the proportioning valve, through the pumpheads, mixer and associated tubing. You then need to add the time it takes to get through the tubing to the injector, through the sample loop, then onto your column. So the total delay time will be longer. Most LPG pumps do not form gradient effectively at this low a flow, so will be neither accurate or reproducible. That is exactly why, at Dionex, we use a split flow system to run small ID columns. We can run a pump much faster and split the flow prior to the column. You state that your MS interface will handle ~20ul/min. To run a 2.1mm column, you would have to split post-column, with the net result being that you will lose sample, and therefore sensitivity. To run 1mm or smaller, you need either a dedicated system, or the correct pre-column splitter equipment. Just running a normal system at 50ul/min will not work effectively for the reasons I describe above. In addition to this, you will get excessive pulsation from the pump running so slowly, causing possible ionisation variations in the MS. Solvent mixing will also be ineffective, which leads to further loss of reproducibility regarding separation and resolution.

Do you have access to another MS interface more suited to 200ul/min? In my opinion, the easiest thing for you would be to use a 2.1mm column directly into the MS. You would still need to ensure that your system is plumbed with the correct tubing (mostly 0.005" ID), but this is the easiest solution.

[Mark Tracy]

The gradient delay won't be changed by changing the column; it is the delay in front of the column. You could use a splitter in two ways: in front of a 1 mm column or after a 4.6 mm column. Either way, the pump would be operating at 1 mL/min so its delay will be only 1 minute. Since sample size is no concern for you, a post-column splitter is the easiest way to do this because the performance requirements of the splitter are low. If you want to do pre-column splitting into a capillary column, my colleagues would be happy to sell you an excellent system; a do-it-yourself system is not for the faint of heart.

[noam]

Thank you,
I was just thinking that the extra delay would be another factor that would give me even longer run times when using capillary column.

My M.S. system is capable of handling roughly 20uL/min while my splitter could convert nicely ~400-200uL/min into 20uL/min.

Basically I'm thinking about using around 2 i.d., and since I would like to shorten my runs, I was wondering about the possibility of using smaller particles and shorter column (currently I'm using 250 length)-
Actually I have a new question, I've looked around and havn't found applications for proteins with particles smaller than 5u (I'm separating small proteins 5000-20000, without digestion),
Is there a clogging issue with smaller particles?

[Mark Tracy]

I've noticed that too. There are two reasons: proteins respond more to the gradient than the stationary phase, so unlike small molecules, the advantage of smaller particles is not as dramatic. The other is that protein samples are notorious for clogging columns; whether that is deserved is another question. It is easier to foul a small-particle column than a large-particle column. Dionex has a column that fits your description; search dionex.com for "Acclaim 300"

[Bryan Evans]

We also have a 3 um (300A) column for the separation
of proteins. Our Intrada WP-RP column can separate
proteins up to 300kDa. Let me know if you are interested:


http://www.imtakt.com/TecInfo/TI262E.pdf