The evolution and current state of instrumentation for analytical supercritical fluid chromatography.

Supercritical fluid chromatography has contributed significantly to chiral method development and semi-preparative purification in drug discovery, where sensitivity was not an issue. Now, analytical scale SFC's have been validated using a multi-vendor, inter-lab study for quantitation of achiral trace pharmaceutical impurities using sub-2 µm particles, with gradient elution, and UV detection. This should significantly increase the penetration of SFC into both chiral and achiral QA/QC applications. However, there is still work to be done. Extra-column dispersion is no better than previous generations, yet the technique is often superficially called "ultra high performance" SFC (UHPSFC), simply because sub-2 µm packings are used. Dispersion is far too high for use with sub-2 µm particles, requiring extensive hardware modification to use such particles with high efficiency. However, the most common means of reducing system dispersion in ultra high performance liquid chromatography (UHPLC) results in distortion of kinetic performance in SFC. Vendors need to specify or provide standardized plumbing schemes that allow the full use of sub-2 µm fully porous, and sub-3 µm superficially porous particles, with reduced plate height, h ≈ 2 when k' ≈ 2. There is no consensus on how to best perform dynamic compressibility compensation, since each vendor uses different CO2 pump head temperatures, resulting in subtle differences in flow and composition between vendors.

Parallel chiral sub/supercritical fluid chromatography screening as a framework for accelerated purification of pharmaceutical targets.

Chiral sub/supercritical fluid chromatography (SFC) has established itself as one of the preferred techniques for enantioseparations at both analytical and preparative scale. Herein, we introduce a parallel multicolumn SFC screening for automated chiral method development in fast-paced settings. The practicality and speed advantages of this approach are illustrated with parallel screening of a diverse set of chiral molecules across ten columns with five different organic modifiers/CO2 based eluents enabling rapid identification of suitable enantioseparation conditions for accelerated purification of pharmaceutical targets. Rapid delivery turnarounds of pure enantiomers of less than 1 h from screening to target isolation are demonstrated illustrating the power of this approach.

Spice authentication by fully automated chemical analysis with integrated chemometrics.

Spices had a worldwide market value greater than $20 billion in 2020. The price per kilogram or ton is high for many spices often due to intensive labor and handling requirements in their preparation, with black pepper as one example. Spices are produced in many countries, in plots ranging from fractions of a hectare to multihectare mechanized operations. From harvest through production and to final market package, spices pass through several steps in the supply chain that provide ample opportunity for fraud. Contamination with harmful plant materials, ingredient substitutions, and adulteration in spices are frequently encountered. In addition, spices in specific geographic regions can have distinctive characteristics from the localized growing conditions giving rise to changes in terpene profiles and other flavor components, creating variation in market price based on country of origin. The research reported here details a system for automated chemical analysis of 170 terpenes in spices and other botanicals by gas chromatography (GC), through enhancements to a proven bacterial identification system. The system integrates pattern recognition searches of custom spice databases to identify country of origin of most spices, aiding in the authentication of spices and detection of fraudulent ingredients.

Ramifications and Insights on the Role of Water in Chiral Sub/Supercritical Fluid Chromatography.

More than 40 cosolvents have been used with carbon dioxide to alter its solvation strength. Among the most interesting systems is the subcritical/supercritical CO2/alkanol eluents. Using small amounts of water in CO2/MeOH is known to be beneficial in chiral subcritical/supercritical chromatography. However, the ramifications of introducing water as a cosolvent component is not entirely understood. In this work, we demonstrate important aspects of the CO2/MeOH/H2O system on nine chiral stationary phases with very different surface chemistries, encompassing derivatized polysaccharides, macrocyclic glycopeptides, iso-butylmercaptoquinine, isopropyl macrocyclic oligosaccharides, and π-electron acceptor/π-electron donor phases. A hydrophilicity scale has been shown to be useful in predicting if a given chiral column chemistry would show a significant enhancement in separation efficiency in the presence of water in the CO2/MeOH system. We demonstrate up to 8-fold enhancements in plate counts of chiral separations with a concomitant decrease in retention times, as predicted by the qualitative test. The same chiral analysis can now be completed in almost a third of the time with the addition of small amounts of water, thereby decreasing organic solvent consumption by a considerable amount. Hydrophobic stationary phases show a minimal increase in efficiency and decrease in analysis times and optimized separations show much larger reduced plate heights, compared to more hydrophilic stationary phases. Furthermore, the presence of water can alter the nature of the adsorption isotherm under nonlinear conditions. Small amounts of water can be used to tune nonlinear tailing peaks into fronting ones, significantly improving preparative enantiomeric separations.

Effect of density on kinetic performance in supercritical fluid chromatography with methanol modified carbon dioxide.

In a relatively recent reevaluation of the van Deemter Equation, Guiochon concluded that the mass transfer resistance in the mobile phase is independent of the retention factor. In the process he showed reduced plate heights ≈ 2 for a nearly unretained peak (k = 0.4) in high performance liquid chromatography (HPLC). In the present work, using supercritical fluid chromatography (SFC), efficiency was measured at various pressures, densities, and modifier concentrations. The highest efficiency, with a reduce plate height of hr = 1.63, was recorded with the lowest retention factor (k < 0.8). This is an extremely low hr for totally porous particles, at very low k, and appears to support Guiochon's analysis. The density of methanol/carbon dioxide mixtures were calculated using the REFPROP program from the National Institute of Standards and Technology (NIST) over a wide range of pressures and % methanol. The density of higher methanol concentrations (>20%), commonly used in SFC, was found to be lower than the density of lower concentrations (<20%). At low methanol concentrations, density varies widely with pressure. However, at high methanol concentrations there is very little change in density, and very little change in retention with pressure. With increasing modifier concentration, density decreases, while viscosity increases (ΔP increases). The pump and back pressure regulator (BPR) pressures are not necessarily good indicators of pressures or densities in the column. At high flow rates the extra-column pressure drop (ΔP) can be much larger than the column ΔP and can be unevenly distributed in front of and behind the column. In one extreme the ΔP after the column was 3 times higher (105 bar) than the actual column ΔP (32 bar).

Preliminary kinetic evaluation of an immobilized polysaccharide sub-2µm column using a low dispersion supercritical fluid chromatograph.

The performance of a 3×50mm, 1.6µm dp column with an immobilized polysaccharide stationary phase (ChiralPak IA-U) was evaluated for efficiency, and pressure drop, with respect to flow rate and modifier concentration using supercritical fluid chromatography (SFC). This appears to be the first such report using such a column in SFC. A unique low dispersion (ultra-high performance) SFC was used for the evaluation. The minimum reduced plate height of 2.78, indicates that the maximum efficiency was similar to or better than coated polysaccharide columns. Selectivity was different from ChiralPak AD, with the same chiral selector, as reported by many others. At high flows and high methanol concentrations, pump pressures sometimes approached 600bar. With 5% methanol, pressure vs. flow rate was non-linear suggesting turbulent flow in the connector tubing. The optimum flow rate (Fopt) at 40% methanol was ≈0.8mL/min, where the column efficiency was highest. At 5% methanol, Fopt increased to ≈1.6mL/min, but efficiency degraded noticeably. The differences in Fopt suggests that the solute diffusion coefficients are a strong function of modifier concentration. Several sub-1min separations, including a 7.5s separation, are presented.

Characterizing pressure issues due to turbulent flow in tubing, in ultra-fast chiral supercritical fluid chromatography at up to 580bar.

It has been widely suggested that the outlet pressure be changed to maintain constant density ("isopycnic" conditions) when comparing the kinetic performance of different columns in supercritical fluid chromatography (SFC). However, at high flow rates, flow in the tubing is turbulent, causing large extra-column pressure drops that limit options for changing outlet pressure. Some of these pressure drops occur before and some after the column, obscuring the actual column inlet and outlet pressures. In this work, a 4.6×100mm, 1.8µm R,R-Whelk-O1 column was used with low dispersion LD (120µm) plumbing to generate sub-1min chiral separations. However, the optimum, or near optimum, flow rate was 5mL-min-1, producing a system pressure of 580bar (with 40% methanol, outlet pressure 120bar). Both the flow rate and pump pressure required were near the limits of the instrument, and significantly exceeded the capability of many other SFC's. Extra-column pressure drops (ΔPec) were as high as 200bar, caused mostly by turbulent flow in the tubing. The ΔPec increased by more than the square of the flow rate. Reynolds Numbers (Re) were calculated for tubing as a function of flow rate between 100 and 400bar and 5-20% methanol in CO2, and 40°-60°C. This represents the most extensive analysis of turbulence in tubing in the SFC literature. Flow in 120µm ID tubing was calculated to be laminar below 1.0mL-min-1, mostly transitional up to 2.5mL-min-1 and virtually always turbulent at 3mL-min-1 and higher. Flow in 170µm tubing is turbulent at lower flows but generates half the ΔPec due to the lower mobile phase linear velocity. The results suggest that, while sub-minute chromatograms are easily generated, 4.6mm columns are not very user friendly for use with sub-2µm packings. The high flow rates required just to reach optimum result in high ΔPec generated by the tubing, causing uncertainty in the true column inlet, outlet, and average column pressure/density. When comparing kinetic performance of columns with different dimensions, the pressure drops in the tubing must be considered.

Instrumental Idiosyncrasies Affecting the Performance of Ultrafast Chiral and Achiral Sub/Supercritical Fluid Chromatography.

It is widely accepted that column technology is ahead of existing chromatographic instruments. The chromatographic output may not reflect the true picture of the peak profile inside the column. The instrumental optimization parameters become far more important when peaks elute in a few seconds. In this work, the low viscosity advantage of the supercritical/subcritical CO2 is coupled with the high efficiency of narrow particle size distribution silica. Using short efficient columns and high flow rates (up to 19 mL/min), separations on the order of a few seconds are demonstrated. In the domain of ultrafast supercritical fluid chromatography (SFC), unexpected results are seen which are absent in ultrafast liquid chromatography. These effects arise due to the compressible nature of the mobile phase and detector idiosyncrasies to eliminate back-pressure regulator noise. We demonstrate unusual connection tubing effects with 50, 75, 127, 254, and 500 µm tubings and show the complex relation of dead time, retention time, efficiency, and optimum velocity with the tubing diameter (via column outlet pressure). Fourier analysis at different back-pressure regulator (BPR) settings shows that some instruments have very specific noise frequencies originating from the BPR, and those specific frequencies vanish under certain conditions. The performance of embedded digital filters, namely, moving average, numerically simulated low pass RC, and Gaussian kernels, is compared. This work also demonstrates, using a simple derivative test, that some instruments employ interpolation techniques while sampling at "true" low frequencies to avoid picking up high frequency noise. Researchers engaged in ultrafast chromatography need to be aware of the instrumental nuances and optimization procedures for achieving ultrafast chiral or achiral separations in SFC mode.

Kinetic performance of a 50mm long 1.8µm chiral column in supercritical fluid chromatography.

Reduced plate heights (hr) of <2 were observed for the first time during the chiral separation of enantiomers, on sub-2µm particles with supercritical fluid chromatography (SFC). The enantiomers of trans-stilbene oxide, were separated on a 4.6×50mm, 1.8µm R,R-Whelk-O1 column, with hr as low as 1.93. The plumbing of a commercial SFC instrument was modified to create a low dispersion version. Without the modification performance was considerably worse. vanDeemter like plots of reduced plate height vs. flow rate, for trans-stilbene oxide, indicate that the optimum flow varied with% modifier. On a 4.6×250mm, 5µm R,R- Whelk-O1 column, the optimum flow was >4mL/min for 5% methanol in CO2, decreasing to <2mL/min for 40% methanol (more than a factor of 2). For a 4.6×50mm column packed with 1.8µm particles the optimum appeared to be near, or >5mL/min with 2.5%, 5%, and 10% methanol, decreasing to between 3 and 3.5mL/min at 40% methanol. This is the first time such shifts have been characterized. Since the solutes were the same in all cases, the differences are likely due to changes in solute diffusion coefficients caused by changes in modifier concentration, and pressure. Pump pressure requirements sometimes exceeded 500bar. It is shown that a 5mL/min flow rate is inadequate for use with 1.8µm particles in a 4.6mm ID column format. Instead, it is suggested to decrease the ID of the column to 3mm, where the optimum flow rates are on the order of 2mL/min with decreased tubing variance. Nevertheless, a number of sub-1min chromatograms are presented.

Instrument modifications that produced reduced plate heights <2 with sub-2 µm particles and 95% of theoretical efficiency at k=2 in supercritical fluid chromatography.

The concept of peak fidelity was shown to be helpful in modeling tubing and detector cell dimensions. Connection tubing and flow cell variances were modeled to determine appropriate internal ID's, lengths, and volumes. A low dispersion plumbing configuration, based on these calculations, was assembled to replace the standard plumbing and produced the reported results. The modifications made were straightforward using commercially available parts. The full theoretical efficiency of a 3×100 mm column packed with 1.8 µm totally porous particles was achieved for the first time in supercritical fluid chromatography (SFC). Peak fidelity of >0.95 was maintained to below k=2. A reduced plate height as low as 1.87 was measured. Thus, true "ultra high performance" SFC was achieved, with the results a major improvement from all previous SFC reports. Since there were no efficiency losses, none could be attributed to thermal gradients caused by the expansion of the fluid over large pressure drops, under the conditions used. Similarly, changes in diffusion coefficients caused by significant decreases in density during expansion are apparently balanced by the increase in linear velocity, keeping the ratio between the diffusion coefficient and the linear velocity a constant. Changing modifier concentration to change retention was shown to not be a significant problem. All these issues have been a concern in the past. Diffusion coefficients, and viscosity data needs to be collected at high pressures before the actual limits of SFC can be discovered.

Instrumentation for analytical scale supercritical fluid chromatography.

Analytical scale supercritical fluid chromatography (SFC) is largely a sub-discipline of high performance liquid chromatography (HPLC), in that most of the hardware and software can be used for either technique. The aspects that separate the 2 techniques stem from the use of carbon dioxide (CO2) as the main component of the mobile phase in SFC. The high compressibility and low viscosity of CO2 mean that pumps, and autosamplers designed for HPLC either need to be modified or an alternate means of dealing with compressibility needs to be found. The inclusion of a back pressure regulator and a high pressure flow cell for any UV-Vis detector are also necessary. Details of the various approaches, problems and solutions are described. Characteristics, such as adiabatic vs. isothermal compressibility, thermal gradients, and refractive index issues are dealt with in detail.

On axial temperature gradients due to large pressure drops in dense fluid chromatography.

The effect of energy degradation (Degradation is the creation of net entropy resulting from irreversibility.) accompanying pressure drops across chromatographic columns is examined with regard to explaining axial temperature gradients in both high performance liquid chromatography (HPLC) and supercritical fluid chromatography (SFC). The observed effects of warming and cooling can be explained equally well in the language of thermodynamics or fluid dynamics. The necessary equivalence of these treatments is reviewed here to show the legitimacy of using whichever one supports the simpler determination of features of interest. The determination of temperature profiles in columns by direct application of the laws of thermodynamics is somewhat simpler than applying them indirectly by solving the Navier-Stokes (NS) equations. Both disciplines show that the preferred strategy for minimizing the reduction in peak quality caused by temperature gradients is to operate columns as nearly adiabatically as possible (i.e. as Joule-Thomson expansions). This useful fact, however, is not widely familiar or appreciated in the chromatography community due to some misunderstanding of the meaning of certain terms and expressions used in these disciplines. In fluid dynamics, the terms "resistive heating" or "frictional heating" have been widely used as synonyms for the dissipation function, Φ, in the NS energy equation. These terms have been widely used by chromatographers as well, but often misinterpreted as due to friction between the mobile phase and the column packing, when in fact Φ describes the increase in entropy of the system (dissipation, ∫TdSuniv>0) due to the irreversible decompression of the mobile phase. Two distinctly different contributions to the irreversibility are identified; (1) ΔSext, viscous dissipation of work done by the external surroundings driving the flow (the pump) contributing to its warming, and (2) ΔSint, entropy change accompanying decompression of fluid in the column, contributing either to warming or cooling depending on local density and temperature. The molecular basis for this variation is described. Sample calculations of dissipation and temperature profiles of several model fluids including carbon dioxide-methanol mixtures are presented, based on the NIST REFPROP program including select equations of state and property calculation software.

Minimizing ultraviolet noise due to mis-matches between detector flow cell and post column mobile phase temperatures in supercritical fluid chromatography: effect of flow cell design.

A mis-match between the post-column mobile phase temperature and the UV detector flow cell temperature can cause significant UV noise in supercritical fluid chromatography (SFC). Deviations as little as 5 °C can increase noise as much as 5 times, making the detector unsuited for trace analysis. Two approaches were used to minimize this noise. When a flow cell was in direct thermal contact (metal on metal) with the detector optical bench, the mobile phase temperature was actively controlled to the measured flow cell temperature, by using one of the heat exchangers (HX) in the column compartment. However, with some older, but still widely used flow cell designs, this required repeated, hourly monitoring of the flow cell temperature and repeated manual adjustment of the heat exchanger temperature, due to thermal drift. Flow cell design had a strong influence on susceptibility to this thermally induced noise. Thermally insulating the flow cell from the optical bench made some cells much less susceptible to such thermally induced noise. Five different flow cells, some insulated, some un-insulated, were evaluated. Most had a truncated conical flow path, but one had a cylindrical flow path. Using either approach, the ASTM noise, with a 10mm, 13 µL conical flow cell, could be optimized to ≈0.007 mAU at 2.5 Hz, in SFC, which is very near the 0.006 mAU manufacturer's specification for HPLC. The insulated version of this flow cell required far less optimization, compared to the un-insulated version. At 150 bar, an experimental 3mm, 2 µL flow cell, with only one side insulated, yielded noise slightly too high (≈0.16-0.18 mAU) for trace analysis, at 80 Hz. However, at 200 bar, noise at 80 Hz was <0.06 mAU, which should allow quantification of a 1 mAU tall trace component with a signal to noise ratio (S/N) >10. Even partially un-insulated, this flow cell design was much less susceptible to thermally induced noise. Further insulating this flow cell design failed to improve performance.

Characterization of a 2.6 µm Kinetex porous shell hydrophilic interaction liquid chromatography column in supercritical fluid chromatography with a comparison to 3 µm totally porous silica.

The first systematic study of the performance of a porous shell, hydrophylic interaction liquid chromatography (HILIC) column in supercritical fluid chromatography (SFC) is presented. Observed efficiency on 2.6-µm porous shell particles exceeded all reports using UHPLC on 100-mm long columns packed with <2-µm totally porous particles. A Kinetex 4.6×150 mm, 2.6 µm HILIC column significantly outperformed a 3 µm Luna totally porous silica of the same length and diameter. A 17 component, low molecular weight test mix, consisting of a range of small drug-like molecules was separated isocratically on each column, with similar selectivity, but the porous shell column required ½ the time (≈2 min vs. 4 min), with almost 50% higher efficiency. Even little retained compounds (k<0.5) exhibited more than 30,000 plates under some conditions. Reduced plate heights were higher than previously reported on porous shell particles in both HILIC and rHPLC, with the lowest value of 1.62. Significant fronting was sometimes observed. The cause of the fronting was not determined. The least symmetrical peaks showed the highest apparent efficiency. Pressure drop at optimum velocity (2.5 ml/min) and low modifier concentrations was <60 bar, and only exceeded 250 bar at near double optimum flow and 65% modifier. Peak widths were mostly just over 0.01 min (20 Hz) wide. There was a loss of efficiency when the injection volume was increased. The chromatograph was shown to have extremely low extra-column dispersion, on the order of 5-10 µL(2), which is also the lowest reported in an SFC, in spite of using standard components. This is likely due to turbulent flow in the tubing and fittings.

Minimizing UV noise in supercritical fluid chromatography. I. Improving back pressure regulator pressure noise.

Pressure fluctuations and resulting refractive index changes, induced by the back pressure regulator (BPR) can be a significant source of UV detector noise in supercritical fluid chromatography (SFC). The refractive index (RI) of pure carbon dioxide (CO(2)) changes ≈0.2%/bar at the most commonly used conditions in supercritical fluid chromatography (SFC) (40 °C and 100 bar), compared to 0.0045%/bar for water (CO(2) IS 44× worse). Changes in RI cause changes in the focal length of the detector cell which results in changes in UV intensity entering the detector. The change in RI (ΔRI/bar) of CO(2) decreases 8-fold at 200 bar, compared to 100 bar. A new back pressure regulator (BPR) design representing an order of magnitude improvement in the state of the art is shown to produce peak to peak pressure noise (PN(p-p)) as low as 0.1 bar, at 200 bar, and 20Hz, compared to older equipment that attempted to maintain PN(p-p)<1bar, at <5Hz. With this lower PN(p-p), changes in baseline UV offsets could be measured as a function of very small changes in pressure. A pressure change of ±1 bar at 100 bar, common with some older BPR's, produced a UV baseline offset >0.5 mAU. A pressure change of ±0.5 bar representing the previous state-of-the-art, resulted in a UV offset of 0.3m AU. Baseline noise <0.05 is required to validate methods for trace analysis. The new BPR, with a PN(p-p) of 0.1 bar, demonstrated UV peak to peak noise (N(p-p))<0.02 mAU with a >0.03 min (10Hz) electronic filter under some conditions. This new low noise level makes it possible to validate SFC methods for the first time.