Chromatography Forum

Here's a question.

Due to the ACN shortage - or just to save solvent in general - one approach is obviously to use a smaller ID column. But we know that when you use a 2 mm ID column on a standard HPLC system you usually have problems: mostly because the volume of the detection cell is too large (other extra-column effects can be controlled).

So my question is: what about 3 mm or 3.2 mm columns. Are these OK on "standard" HPLC systems or do we need to replace the detector cell?

Much Thanks in Advance.

It will depend on the quality of your LC system. If you have a relatively new system with small i.d. tubing and a relatively small flow cell (10 uL or less), then you can probably be successful. Remember you may also have to scale your injection volume with the flow rate. On many systems this should be an easy way to save 50% on solvent usage.

Try it and see. Then report back with your results.

I am helping a pharma company with the same issue. If the results are usefull I can let you know.

I have used 3mm columns on old (10+) systems without any serious problems.

I agree, and to be specific, "relatively recent" can be measured in many years.

I also have successfully used these columns on a 10-year-old Agilent 1100. The only change was to a semi-micro flow cell.

I also can also use this system for 3 um columns, and even the solid core fast columns. You can get pretty impressive results without investing in an expensive new system, ... if you know what you are doing. See me at Pittcon to see some examples.

Adam,

you are correct the most problematic issue is that of extra column volume effect, and the flow cell is one of them.
still 3 mm id column should show less problems then 2 mm columns.
simply run one such application you need on one of your HPLCs and see what you need to do in order to restore hardware performance.
if changes are necessary it will be of smaller tubing id betweent the sampler and the column and after the column to the UV.
also in most detectors you should be able to swap your analytical flow cell to a semi micro one which will add less dwell volume.

i think that there are out there charts of column volume to efficient preferable extra column volumes of system.
maybe i have on of those somewhere

the only time we made a change was to a smaller flow cell and tubing was using a 50 mm column 3 mm id 3 micron on a very old instrument that we scrapped a year after that.

We use 2.1mm id. columns as well as 3mm i.d. columns on 20-year HP/Agilent 1050 HPLC systems, works great. Halve your flow rate with 3mm columns of same length, run 1/4 original rate for 2.1mm columns. That keeps the linear velocity about the same. You may need to hold injection volume to 10 ul or less (we normally inject 5 ul).

But when I scale down from Agilent Zorbax XDB C18 (150mm*4.6mm 5um) column to Agilent Zorbax XDB C18 (50mm*2.1mm 1.8um RRHT) column. Lots of early peaks collapse together.

Condition for long column run:
70:30 MeOH:water (isocratic)
0.4ml/min flow rate (we use 1200 LC-MS, keep slow flow for ESI & APCI ionization)
semi-micro flow cell (5ul)
There are all baseline separation for 13 compounds.

When I switch to short column and modified these things:
70:30 MeOH:water (no change, isocratic)
0.208ml/min flow rate
DAD peakwidth set to fastest (40Hz)
all connecting tubings changed to small 0.12mm id. Keeping them as short as possible.
Use weaker-than-eluent system for sample separation (50:50 MeOH:H2O)

That's ok for late eluting compounds (e.g. Benzophenone, Toluene, Xylene, Naphthalene, etc)
But for ealry eluting compunds (e.g. Acetone, Phenol, 3,4-Dimethylbenzene, acetophenone, etc) They merge together.

When I change to 55:45 MeOH:H2O, they look better but still I can see some tailing for all peaks.

Is it the normal case when I switch from "classic" long column to "modern" short RRHT column? Must I change to solvent system to achieve the baseline separation?

Anthony: you either have still too much extra-column bandspreading on your instrument to run the smaller column to the same conditions, or your injection solvent is not correct. Also, you are probably running the RRT column at too low a flow rate, but - considering that your last peaks are OK - this is probably not the major problem.

Anthony - from what you state, you just didn't "change the diameter" - you changed length, particle size, etc. And when you drop to higher aqueous content in your mobile phase, be aware that if injecting your sample in organic solvent may cause issues, that you may need to decrease the injection amount as well. So with all these changes: optimizing the separation is your job, have fun.

Thanks experts!!!

May I give you more information in my experiment.

Tubings shipped with the instrument is 0.17mm id (green). I changed to smaller one (0.12mm id, red), the one from injector port to column, and column end connected directly to the flow cell. (We do not have thermostat compartment)

But I didn't change the tubing from needle seat to injection port yet (it is 0.17mm id, 3cm tube). Also I am not sure if the tubing from pump head to injection port (or autosampler) will affect the result.

Injection solvent is 50:50 MeOH:H2O (sample compounds dissolved in this system). I know that if I use 100% MeOH to disslove them and run in 70:30 MeOH:H2O, problems will occur in early eluting peaks because the injection solvent is stronger than the eluent solvent system.

Injection volume is reduced from 5ul to 1ul. I tried to increase flow rate from 0.208 to 0.30, or even 0.35ml/min. Patterns are similar (late eluting peaks ok but early peaks cannot resolve well). We cannot increase further because pressure goes up to 350bar (Agilent 1200 Quaternary Pump limit is 400bar).

Maybe I can extract the chromatogram to you all next week (because the computer is down on Black Friday)

The simplest way to know if extra-column effects are the issue is to disconnect the column and inject your sample, maybe after dilution, if necessary, at the same flow rate with the same mobile phase as in your analysis. If the peak that you are getting is of a substantial width compared to the ones in your chromatogram, you know that you have problems from extra-column effects. Then you can systematically look, what the major contributor to this effect is by replacing and shortening tubings etc. You may also end up with the conclusion that the instrument may not be suitable for what you want to do, and at least more substantial parts of the instrument need to be changed.

I am surprised to hear everyone saying that 3 mm ID columns and even 2 mm ID columns should be fine on standard HPLC systems. It seems to me that one of the main reasons UPLC is not commonly used in GMP labs is because of the potential issues with extra column effects. A warped fitting or a void that develops in the column, and the chromatography deteriorates. And running narrow bore columns on standard LC systems would have more potential problems.

I think if you run 2 mm ID columns on standard HPLC's it will often be problematic - unless a smaller volume detection cell is installed and you don't have too much dead volume in the system (the other issues are more easily controlled by the analyst: length and ID of connecting tubing, injection volume, etc).

With 3 mm ID columns I think you are in a safer range - which is what prompted my original question - but still some caution is needed.

Hi Adam,

Going from 4.6 to 3 mm column shouldn’t be a problem at all, provided the chromatographic equipment is of a reasonable configuration.
I’ve done it several times - achieving eluent reduction (at least halving the usage) and even improving the plate count. And I’m talking about pretty standard equipment with a flow-cell of 10 μL and so on. When it comes to 2.1 mm columns there is another factor to be taken into account as well. And that is the backpressure. So a newer/modern hardware configuration might be the way forward if 2 mm columns are the ultimate goal.
Regarding the “UPLC in GMP labsâ€

Just saying that we've been using 2.1mm i.d. columns on Agilent 1050 and 1100 for at least 15 years, and our manufacturing locations do this on different instruments, don't report any issues.

Here are the chromatogram:
Compond (increasing RT): Phenol, 4-Nitrochlorobenzene, Toluene, Naphthalene

Upper: Agilent Eclipse XDB-C18 4.6x150mm, 5um column (65:35 MeOH:H2O, flow:0.4ml/min, inj vol: 5ul)
Lower: Agilent Eclipse XDB-C18 2.1x50mm, 1.8um column (65:35 MeOH:H2O, flow: 0.21ml/min, inj vol: 1ul)



The chromatogram are silimilar to the LC column performance report shipped with the column.


When I add a bit more compounds (especially early eluting ones). I got the following 2 chromatograms:
Compound: Acetone, Phenol, Acteophenone, 3,4-Dimethylphenol, 3,5-Dimethylphenol, Propiophenone, 1,3-Dimethoxybenzene, 4-Nitrochlrobenzene, Benzophenone, Toluene, alpha-chloro-p-xylene, Naphthalene, Xylene

Upper: 70:30 MeOH:H2O, flow: 1ml/min, inj vol: 5ul
Lower: 70:30 MeOH:H2O, flow: 0.21ml/min, inj vol: 1ul



1. All tubings are reduced to small (0.12mm id)
2. Small volume flow cell (5ul) is used.
3. Injection solvet: 50:50 MeOH:H2O
4. Hardware: Agilent 1200 Quat Pump, DAD 40Hz, Standard Autosampler

Is the instrument not suitable for fast-LC column? (still dead volume somewhere?). Note also that there are small tailing on every peaks in the last chromatogram.