I hear mixed messages about filtering solvents. Some of my colleagues filter all solvents used in their LC systems, whereas others seem to filter the mobile phase only part of the time. Shouldn't you filter everything?

There is no harm in filtering every mobile phase you use, but you may not be gaining much by your efforts. Most workers filter solvents using a vacuum filter and a 0.5 or 0.45 mm porosity, 47 mm diameter membrane filter. At first glance, one would think that all solvents would benefit from filtration, but this is not necessarily the case. Look at the specifications for the HPLC-grade methanol or acetonitrile component of the mobile phase. Either on the bottle or in the accompanying literature, you'll find that these solvents were filtered through a 0.22 mm filter prior to bottling. This porosity filter usually is considered a "biological" filter used to remove any bacterial contaminants from the solvent in addition to any particulate matter. Most of us use lab-generated HPLC-grade water from a multi-stage water purification system. Usually the last step in water preparation is passage through a 0.22 mm filter. So you can see that solvents that nominally are clean at the 0.22 mm level won't gain anything by passage through a 0.45 mm filter. This assumes, of course, that your glassware is dust-free.

What about mobile phase additives? In my experience, the HPLC-grade additives also are filtered at the 0.22 mm level during preparation. For this reason, the Standard Operating Procedure (SOP) in my laboratory does not require filtration of any mobile phase which contains only HPLC-grade liquids. Solids are a different story. Consider,for example, the salts used to make buffers. The specifications may indicate that 0.22 mm filtration was part of the manufacturing process, but two other potential problems exist.First, buffer salts sometimes have solubility problems, especially with acetonitrile as an organic solvent. Second, although the buffers are chemically pure, they may contain bits of foreign matter. For example, if you look inside the lid of the buffer bottle, you may notice that the liner is scratched where it rubs on therim of the bottle. Bits of Teflon from the liner can drop into the bottle and end up in the mobile phase. Trace amounts of particulate material can foul check valves, causing pressure and flow problems. For this reason, the same SOP that allows us to skip filtration for HPLC-grade liquids requires filtration whenever solids are added to the mobile phase.

What's the best way to filter the solvents? I prefer the membrane filters mentioned above with 0.45 mm porosity. Teflon membranes are difficult to wet and filter very slowly. Pre-wetting Teflon with a few drops of pure organic solvent (methanol or acetonitrile) may help speed filtration. I prefer one of the alternate membrane materials designed for filtering mobile phases. Consult the manufacturer's literature to determine solvent compatibility before you first use a particular filter type.

I have seen advertisements for solvent filters that incorporate reversed-phase particles or carbon in the filter membrane. These are purported to provide additional mobile phase cleanup by adsorbing trace organic impurities. This sounds like a logical improvement to mobile phase preparation, but I do not have any experience with such products. Please add your experiences with these materials to this conversation.

Of course, filtration with a vacuum filter apparatus also accomplishes degassing. Improperly degassed mobile phases can cause post-column bubble formation, depending on the experiment. So, if you decide to not filter anymore, be aware that you might still need to degas, which can be done in a variety of ways.

Also, as pointed out by John Dolan in his reply, you need to ensure the compatability of your filter with your mobile phase. I have occaisionally experienced contamination in ion chromatography mobile phases when an "equivalent" filter to the one specified in a validated method was used, and no one had check the suitability to our method.

I work for a large chemical company with over 50 LC's and in 10 years I haven't seen anyone "externally" filter anything but obviously paticulate laden samples. We use catridge precolumn in-line filters from Upchurch #A315(or equivalent). Occasionally they need to be changed as evidenced by higher then normal back pressure. I have never seen a column fail from particulate contamination nor is such a failure likely with the precolumn filter in use. We use B&J solvents, MilliQ water and a variety of buffer salts. If filtering was necessary we would experience frequent in-line filter failure, but this is not the case. It seems to me that solvent filtering must be a legacy from the dawn of LC when solvents were less pure.

I am all for filtration when using solids in buffer preparation, however with a vacuum filter apparatus you not only degass your mobile phase, but also tend to remove organic solvents by evaporation. We used to filtrate phosphatebuffers with TBA and 30% v/v methanol... not any more since we noticed that the actual methanol content depended on the time of filtration!