High performance liquid chromatography column packings with deliberately broadened particle size distribution: relation between column performance and packing structure.

The effect of the addition of 25%, 50% and 75% (weight percent, wt%) of larger particles (resp. 3 and 5 µm) to a commercial batch of 1.9 µm particles has been investigated as an academic exercise to study the effects of particle size distribution on the kinetic performance of packed bed columns in a magnified way. Comparing the performance of the different mixtures in a kinetic plot, it could be irrefutably shown that the addition of larger particles to a commercial batch of small particles cannot be expected to lead to an improved kinetic performance. Whereas the addition of 25 wt% of larger particles still only has a minor negative effect, a significantly deteriorated performance is obtained when 50 or 75 wt% of larger particles are added. In this case, separation impedance number increases up to 200% were observed. Studying the packing structure through computational packing simulations, together with the experimental determination of the external porosity, helped in understanding the obtained results. This showed that small particles tend to settle in the flow-through pores surrounding the larger particles, leading to very high packing densities (external porosities as low as 32% were observed) and also negatively influencing the column permeability as well as the band broadening (because of the broadened flow-through pore size range).

Modelling the relation between the species retention factor and the C-term band broadening in pressure-driven and electrically driven flows through perfectly ordered 2-D chromatographic media.

The band broadening that can be expected in perfectly ordered cylindrical pillar arrays has been calculated for a wide range of intra-particle diffusion coefficients (D(sz)) and zone retention factors (0

Influence of pressure and temperature on the physico-chemical properties of mobile phase mixtures commonly used in high-performance liquid chromatography.

To fulfil the increasing demand for faster and more complex separations, modern HPLC separations are performed at ever higher pressures and temperatures. Under these operating conditions, it is no longer possible to safely assume the mobile phase fluid properties to be invariable of the governing pressures and temperatures, without this resulting in significantly deficient results. A detailed insight in the influence of pressure and temperature on the physico-chemical properties of the most commonly used liquid mobile phases: water-methanol and water-acetonitrile mixtures, therefore becomes very timely. Viscosity, isothermal compressibility and density were measured for pressures up to 1000 bar and temperatures up to 100 degrees C for the entire range of water-methanol and water-acetonitrile mixtures. The paper reports on two different viscosity values: apparent and real viscosities. The apparent viscosities represent the apparent flow resistance under high pressure referred to by the flow rates measured at atmospheric pressure. They are of great practical use, because the flow rates at atmospheric pressure are commonly stable and more easily measurable in a chromatographic setup. The real viscosities are those complying with the physical definition of viscosity and they are important from a fundamental point of view. By measuring the isothermal compressibility, the actual volumetric flow rates at elevated pressures and temperatures can be calculated. The viscosities corresponding to these flow rates are the real viscosities of the solvent under the given elevated pressure and temperature. The measurements agree very well with existing literature data, which mainly focus on pure water, methanol and acetonitrile and are only available for a limited range of temperatures and pressures. As a consequence, the physico-chemical properties reported on in this paper provide a significant extension to the range of data available, hereby providing useful data to practical as well as theoretical chromatographers investigating the limits of modern day HPLC.

Kinetic plot and particle size distribution analysis to discuss the performance limits of sub-2 microm and supra-2 microm particle columns.

To contribute to the current debate about the "ideal" particle size range (sub-2 microm vs. supra-2 microm), the present study compares the kinetic performance of some commercially available sub-2 microm and 3.5 microm particles used under quasi-adiabatic conditions via the kinetic plot method. Under the adopted assumption that viscous heating effects can be neglected (which is uncertain in a pressure range above 400 bar), the obtained kinetic plots show that, provided each particle size is used in a column with properly optimized length, the gain in separation speed that sub-2 microm particle columns might have over maximally performing 2.5 microm particle columns is very small. Sub-2 microm particle columns can only yield a gain in separation speed in the range of high-speed/low-resolution-separations (total time based on k=10 below 5 or 10 min). And even in this range, the actual gain that can be expected is only marginally small (only a few %). The present study hence suggests that the development and the use of particles in the 2-3 microm range should deserve more attention than it did in the past few years. However, to be competitive, this 2-3 microm material should be packed in relatively long columns, with a packing quality matching that of the current best performing 3.5 microm particle columns. The supra-2 microm particles should also be able to withstand the same pressures as the sub-2 microm particle material one is comparing it to.

Detailed characterisation of the flow resistance of commercial sub-2 micrometer reversed-phase columns.

The pressure-drop characteristics of six sub-2 microm columns of three different manufacturers but with the same surface chemistry (C(18)) have been studied using the recently introduced total pore blocking method for the determination of the external porosity and by using scanning electron microscopy pictures to measure the actual size of the particles. Having used the Sauter-mean particle size to correctly account for the particle size spread, and having corrected for the influence of the intra-particle porosity, it was found that all columns yielded Kozeny-Carman constant (f(KC)) values close to 180, in agreement with the theory. This agreement could subsequently be used to quantify how the different system parameters such as mean particle size, packing density and intra-particle porosity (all tending to vary significantly from manufacturer to manufacturer) each contributes to the observed total bed permeability. Small (upward) deviations from the f(KC)=180-value could be correlated to a larger width of the particle size distribution, and more notably to the existence of a high size ratio of the largest to the smallest particles present in the particle batches.

Experimental investigation of the band broadening originating from the top and bottom walls in micromachined nonporous pillar array columns.

We report on the experimental investigation of the effect of the top and bottom wall plates in micromachined nonporous pillar array columns. It has been found that their presence yields an additional c-term type of band broadening that can make up a significant fraction of the total band broadening (at least if considering nonporous pillars and a nonretained tracer). Their presence also induces a clear (downward) shift of the optimal velocity. These observations are, however in excellent quantitative agreement with the theoretical expectations obtained from a computational fluid dynamics study. The presently obtained experimental results, hence, demonstrate that the employed high aspect ratio Bosch etching process can be used to fabricate micromachined pillar arrays that are sufficiently refined to achieve the theoretical performance limit.

Relation between the particle size distribution and the kinetic performance of packed columns. Application to a commercial sub-2 microm particle material.

To study the influence of the particle size distribution (PSD), we measured the chromatographic performance of a series of sub-2 microm particle high performance liquid chromatography (HPLC) columns packed with four different particle mixtures having a purposely imposed different size distribution. Using the reduced kinetic plot representation by plotting the separation impedance (E(0)) versus the plate number ratio (N(opt)/N), the different columns could be classified according to their chromatographic performance without the need to specify a mean particle diameter or a molecular diffusion coefficient, as is needed in the classical reduced plate height and flow resistance analysis. The present analysis shows that it is not so much the width or span of the particle size distribution, but rather the presence of fines that greatly determines the chromatographic performance of particulate columns.

Domain size-induced heterogeneity as performance limitation of small-domain monolithic columns and other LC support types.

We have computed the band broadening and the flow resistance in a series of apparently self-similar porous LC support structures, all having the same mean geometric ratios and external porosity, but with a decreasing scale and disturbed by a scale-independent variance on the size and position of the porous solid zone elements. The study shows in a general and qualitative way that each type of LC support that is produced using a manufacturing process displaying a fixed (i.e., domain-size independent) variance on the size and position of the produced solid zone elements will eventually encounter a limit beyond which a further reduction of the domain size can no longer be expected to yield a significant gain in separation speed. This is currently observed in practice for silica monoliths and could also compromise the performance of photolithographically etched columns.

A first principles explanation for the experimentally observed increase in A-term band broadening in small domain silica monoliths and other chromatographic supports.

The present computational study illustrates how the existence of a residual lower limit on the variance of the skeleton and through-pore size of monolithic columns can be expected to severely compromise the possibility to prepare well-performing small domain monolithic columns. Adopting rather conservative estimates for the minimal standard deviation on the pore and the skeleton size (0.2 and 0.04 microm, respectively), the presented calculations show that, if such a fixed lower limit on the size variance exists, it will be impossible to decrease the A-term band broadening below a given critical value, no matter how small the domain size is made. From a given critical domain size value on, any attempt to further decrease the domain size without being able to co-reduce the size variance can be expected to be counterproductive and leads to an increase instead of to a further decrease of the plate heights.

Importance and reduction of the sidewall-induced band-broadening effect in pressure-driven microfabricated columns.

The influence of the detailed design of the sidewall region upon the over-all band-broadening in microfabricated packed-bed or collocated monolithic support structure (COMOSS) columns has been investigated using computational fluid dynamics (CFD) simulation techniques. It is shown that, under unretained solute conditions, very small structural variations of the order of only 5% of the particle diameter can give rise to a 4-fold increase of the band-broadening. A comprehensive study has been made to quantify this effect as a function of the fluid velocity, the particle diameter, the channel widths, and of course, the sidewall region design. Because the sidewall effect can be fully attributed to a mismatch between the flow rates in the column center and in the sidewall region, it is fortunately also quite straightforward to avoid it. A very simple design, yielding band-broadening values identical to that of a hypothetical sidewall-less column for all possible values of the flow velocity, the particle diameter, or the channel width is proposed.