Mobile Phase Filtration

[Bruce Hamilton]

Hi Victor,

I tend to ultrasonic degas all premixed solvent mixtures, whether containing buffer or ion-pair salts or not. So, even if I vacuum filter a mobile phase with dissolved salts, I will also ultrasonicate it.

The reason for this is, unlike Hans, I don't have a vacuum system that could hold 1 litre Duran/Schott bottles, and I only have one of those bottles that is vacuum rated ( much more expensive than standard ones ).

If I filter through the Millipore 47mm all-glass system, I still have to transfer the filtrate to the 1L mobile phase reservoir, so quickly putting it into an ultrasonic bath removes at least some freshly-dissolved gases.

I agree that helium sparging is the best, especially for systems without an in-line degasser and using the same mobile phase repeatedly, eg on ion-chromatographs.

The main reason I prefer ultrasonication is that there are no sinters or vacuum lines that could contaminate the mobile phase. I have a 100 watt ultrasonic bath that really shakes those gases out, and can hold two 1 liter Duran bottles.

The ultrasonic bath was fairly cheap ( about US$150 ), and I think they are mainly used for helping to sterilize the implements of torture used by dentists. I recommend a high power unit, but be prepared for it to shatter autosampler vials if you also use it to help sample dissolution.

The "holes in aluminium foil" test gives a reasonable indication of the energy available.

With regard to single solvent mobile phases, I don't bother too much, unless I believe they will be an issue - such as with methanol-water gradients, and some normal phase mixtures. I rely on the instrument in-line degasser for those. But I should note that I have my Agilent 1100 degasser set for continuous operation, rather than the default mode.

The mobile phases I find giving the most pump problems are normal phase mixtures and aqueous-solvent ion-pair mixtures, which are isocratic.

The normal phase mixtures tend to give me the most problems with bubbles forming in some older UV detectors, but modern systems seem to have eliminated most detector bubbles..

With regard to filtering HPLC solvents, I assume they are OK from most reputable manufacturers, but may perform some confirmatory tests on new suppliers.

The easiest way to check for really bad contamination is to accurately preweigh two filters, stack them together, filter the solvent through them, allow them to air dry and reweigh. You should examine the upper surface of both using a binocular microscope. The lower filter also enables you to easily correct for weighing/drying variations without getting into a really fussy methodology....

Please keep having fun,

Bruce Hamilton

[Russ]

For what it's worth, I have seen colored "threads" on the filter paper after filtering freshly opened HPLC solvents. They were readily visible to the eye. Don't know what else is there, if anything, but I thought it was strange since the solvents are supposed to be filtered at the manufacturer. Have seen this from a couple of vendors / suppliers.

[danko]

Hi Russ,

Your observation might have another explanation. If the filter contains some impurities that are invisible when dry, these impurities can get colored upon contact with a given liquid. Well known effects of the kind are color shift with pH change, or change of solvent type (f. ex. from polar solvent to non-polar one) etc.
I’m not trying to convince you that your observation isn’t valid, but it’s good to keep other possibilities in mind.
It might even be possible that the filter manufacturer has built in this as an extra feature on purpose, so that one could distinguish used filters from unused.

Best Regards

[HW Mueller]

Bruce, there is no pressure difference due to pulling a vacuum on a container which is placed, opened, into a dessicator. I use mostly the 1 L bottles in which solvents are sold, one does not need a vacuum bottle for this.

On He sparging: I am sure that if anybody claims that it is best that they mean it´s best for their application. As mentioned before, He doesn´t degas at all, it regasses. You sweep air out, but now you have He dissolved in there. It just happens that a combination of He concentration and possibly its physical nature causes it to be less of a peoblem than air.
I have also mentioned this before: In UV check of solvents He is almost as big a problem as air or N2. To do these checks I use an emptied prep column with an SSI needle valve plumbed onto one end. I draw ln the solvent with the column in a vertical position and the needle valve on the bottom. Then the valve is closed and different times of vacuum pulled on the content. For the tests N2 pressure is applied the valve end connected to the UV detector, etc. . . . I seem to remember the data with water: Saturated with N2, air, or He the baseline was about 30 000 µV higher than after +20 min of vacuum (directly pulled on the solvent by closing the needle valve and attaching the othsr column end to a vacuum pump). The noise of the baseline was less when the water was saturated with He than with N2, as an example. He seems to have a lower tendency to degas. But remember that the reduction of light through the cell was about the same for all gases. (The useful scale goes to ~ 10^6 µV).

It should be mentioned that one sees what Uwe mentioned: The strongest drop of baseline is obtained right after a short application of vacuum, but a slower decrease in baseline continues with longer application of vacuum.

[Bruce Hamilton]

Bruce, there is no pressure difference due to pulling a vacuum on a container which is placed, opened, into a dessicator. I use mostly the 1 L bottles in which solvents are sold, one does not need a vacuum bottle for this.

Hans. Thank you for that. I should have been more explicit. I don't have a desiccator that can conveniently handle conventional 1L Duran/Schott bottles, and I only have one bottle that can handle the alternative method of degassing vacuum applied directly to the bottle without being inside a desiccator.

With regard to solvent UV absorption when gases were present, a few years ago I tested a UV-grade solvent with nitrogen sparging + ultrasonication for a local distributor who had a customer complaint. I also tried He, and at the test wavelengths 190 - 240 I found that both gave similar data - considerably quicker than just using vacuum, which I think also gave a similar result.

IIRC, Helium solubility in solvents is normally lower than oxygen and nitrogen, so I think "degassing" is appropriate term. I think there's Japanese? authored free article on the WWW that shows the change in ghost peaks when solvents are degassed.

With regard to fibres, some manufacturers used to ship dangerous goods internationally in large containers, and then decant into glass bottles locally. I'm not sure if that is still the case, but I also recall some "0.45um filtered" HPLC grade acetonitrile which had visible particles and 5L turned a 47mm 0.5 fluoropore filter light grey with some visible particles present.

I never received a formal response when I gave the agent the filter and one pristine filter after 5L from another supplier, but was advised that it was due to rectified problems at a decantation station. A retest several months later still gave a grey colour, so I switched suppliers.

Please keep having fun,

Bruce Hamilton

[lmh]

A minor hijack:

can I ask how people generally filter? I've always used a cellulose nitrate filter (0.2u) in a glass filter support/side-arm flask set-up. But someone has just trashed our filter support, and I can only find a nylon/plastic bottle-top filter (in pristine unopened condition) as a temporary fix while I wait for new. I remember trying samples through various syringe-tip filters years ago, and checking the background ions in LC-MS. Some were frightening.

I'd appreciate others' experiences...

(Also any hints on what to do to the person who trashed our filter support, if I ever find out who it was)

[Bruce Hamilton]

I think the Millipore catalog has a good table showing compatibilities of filters and housings of fabricated filters with solvents.

I use the Millipore 47mm all glass filtration unit with a large (50?mm) Sartorius Midisart 2000 0.2um PTFE fabricated filter on the side arm going to the vacuum source. The filter is to stop vacuum tubing debris.

I use three types of the Millipore 47 mm membrane filters.
Mixed cellulose esters for Water, IPA, alkanes hydrocarbons
Hydrophilic PVDF ( Durapore ) for aromatic hydrocarbons, Dioxane, Esters,
ammonia solutions, carboxylic acids, nitrile - water mixtures
PTFE ( Fluoropore ) for chlorinated HCs, Ketones, DMSO, THF and other nasties.

I prefer the unlaminated version of PTFE, so the PE substrate can't swell, but I think it's only available these days as 0.5um.

In general, the hydrophilic PVDF is also used for most water-solvent mixtures, unless the PTFE inertness is obviously needed. I seldom bother to prewet hydrophobic filters, just allowing the mobile phase to diffuse through before applying vacuum.

I keep using the filter until flow starts to slow, which can be several months these days because the number of mobile phases requiring filtration has greatly reduced. I also dedicate a filter to each type of ion-pairing mobile phase.

After each use I rinse the all-glass system and filter with clean solvent then Milli-Q water and then small quantity of solvent before using the vacuum to dry the filter and support before storing.

As for your colleague, the best option would be to swap some hydrophobic filters randomly into his mixed ester or hydrophilic stocks.

Please keep having fun,

Bruce Hamilton

[mbicking]

Also any hints on what to do to the person who trashed our filter support, if I ever find out who it was

I think this calls for having a little fun!

Let's show the world that we chemists have a sense of humor too!

How about some suggestions of some harmless practical jokes that would be appropriate fun for the guilty party?

[lmh]

Thanks for that very full description; very helpful! Yes, suggestions for filter-abusing co-workers much appreciated too.

[rhaefe]

Bruce,

is there any reason why THF and dioxane, Ketones and Esters are in different "filter groups"?

Thanks,

Robert

[Bruce Hamilton]

is there any reason why THF and dioxane, Ketones and Esters are in different "filter groups"?

Hi Robert,

My premature senility?

The list was abstracted from a brief list on my computer to remind me which to use, just based on experience. The incosistencies are usually based on my experience of pure solvents flow slowing on clean filters, suggesting swelling.

The Millipore listing is also dynamic. My 1980s catalogue gives different compatibility results to my 2003 catalogue - the older catalogue suggests the filters are less affected by some solvents. I'd recommend using the latest.

My 2003 Catalogue suggests:-

Dioxane = PTFE Generally recommended, PVDF = Recommended
THF = PTFE Generally recommended, PVDF = Not recommended
Acetone = PTFE Recommended, PVDF = Not recommended
MEK = PTFE Recommended, PVDF = Not recommended
EtOAc = PTFE Recommended, PVDF = Recommended
BuOAc = PTFE = Generally recommended, PVDF = Test.

For example, the older edition has:-
Dioxane PTFE = Recommended, PVDF = Not Recommended.
EtOAc = PTFE = Recommended, PVDF = Not Recommnded
BuOAc = PTFE = Recommended, PVDF = Recommended

Generally, I would be filtering solvents like dioxane as minor components in a premixed mobile phase, so would use filter based on most aggressive or major component.

When I have time, I'll have a peek online to see what the latest catalogue says..

Please keep having fun,

Bruce Hamilton

[Russ]

Your observation might have another explanation. If the filter contains some impurities that are invisible when dry, these impurities can get colored upon contact with a given liquid. Well known effects of the kind are color shift with pH change, or change of solvent type (f. ex. from polar solvent to non-polar one) etc.

Interesting. I wasn't aware of that. However, I do not always see the fibers even when using the same package of filters and same organic solvent (usually acetonitrile or methanol). The fibers also appear to be "loose" on the filter and not embedded in it.