Evaluation of ultra performance liquid chromatography Part I Possibilities and limitations
Citation: André de Villiersa, François Lestremaub, Roman Szucsb, Sylvie Gélébartb, Frank Davida, Pat Sandra (2006/09) Evaluation of ultra performance liquid chromatography Part I Possibilities and limitations. Journal of Chromatography A (Volume 1127) (RSS)
DOI (original publisher): 10.1016/j.chroma.2006.05.071
Semantic Scholar (metadata): 10.1016/j.chroma.2006.05.071
Sci-Hub (fulltext): 10.1016/j.chroma.2006.05.071
Internet Archive Scholar (search for fulltext): Evaluation of ultra performance liquid chromatography Part I Possibilities and limitations
Tagged: Analytical Chemistry (RSS), class assignments (RSS)
Summary
Abstract
• Posibilities and limitations of UPLC technique are evaluated via van Deemter and Knox plots on Acquity BEH 1.7 μm columns. In addition, working-range maximum efficiency and analysis time were compared between UPLC and HPLC. Advantages of UPLC conditions in speed of analysis for up to approximately 80,000 theoretical plates were demonstrated with 'Poppe-plots' constructed from experimental data.
1. Introduction
• HPLC is a dominating separation technique routinely used in the industry due to its essential and irreplaceable characteristics, but compared to GC and CE, it lacks high efficiency. This disadvantage is considered to be due to slow diffusion speed of the analytes into the stationary phase which in its turn results from the small diffusion coefficients in the mobile phase.
• Efficiency can be improved by using small particles packings which reduce the diffusion path of the analytes. This is also supported by van Deemter equation: H = A + B/u + Cu, where the most significant impact on the efficiency would be introduced from the C-term (C = f(k)dp²/Dm is the resistance to mass transfer in the mobile phase). Decrease in the particle size and consequent increase in the linear velocity results in significant decrease in the plate height (significant increase in the efficiency).
• Different approaches were taken in improvement of HPLC efficiency. Successful results were achieved with the use of high temperature LC and monolithic columns, but the most promising results were obtained with UPLC. Nevertheless, certain shortcomings such as limitations in the analysis speed and efficiency at the fixed pressure and particle size, possible formation of temperature gradients across the column, and significant demands on columns and instrumentation due to high pressures are still valid even for UPLC.
• Evaluation of the limitations and practical possibilities of UPLC were performed with the aid of van Deemter plots, generated from experimental data on Acquity BEH, 1.7 μm columns. Further, UPLC and HPLC results were compared using kinetic plots constructed by the method of Poppe.
2. Experimental
2.1 Materials
• Sigma-Aldrich HPLC grade acetonitrile and water; Waters Acquity BEH C18, 50 mm or 100 mm x 2.1 mm, 1.7 μm; Waters XBridge BEH C18, 150 mm x 4.6 mm, 3.5 μm and 250 mm x 4.6 mm, 5 μm.
2.2 Instrumentation
• Waters Acquity UPLC (binary pump, DAD detector and custom 1.8 μL sample loop).
• Agilent 1100 HPLC (binary pump, DAD detector and custom 0.127 mm ID PEEK tubing between injector, column and detector).
2.3 Chromatographic method details
2.3.1 Analysis conditions
• Test mix in 30:70 acetonitrile:water – 0.4 mg/L uracil, 2 mg/L caffeine, 40 mg/L pyridine, 2 mg/L aniline, 2 mg/L phenol, 2 mg/L acetophenone and 10 mg/L benzene (prepared fresh daily from individual components stock solutions in acetonitrile). Pre-mixed mobile phase – 30:70 acetonitile:water. Column temperature - 40°C. Wavelength – 210 nm, acquisition rate – 40Hz.
2.3.2 van Deemter curves on Acquity columns
• Injections performed in the full-loop mode with the needle wash (mobile phase) on UHPLC. Varied flow rates were used – from 0.05 mL/min to the flow rate at the maximum pressure of 1034 bar.
2.3.3 van Deemter curves on XBridge columns
• HPLC injections performed were equal to injections on UHPLC (scaled to 1.8 μL of Acquity relative to XBridge dimensions). Varied flow rates were used – from 0.2 mL/min increased in steps of 0.2 mL/min until the maximum pressure of 400 bar was reached.
2.3.4 Data handling
• Peak width at ½ height were used for theoretical plates calculations. Van Deempter non-linear curves were constructed using Origin 6.1.
2.3.5 Multiple column experiments
• Series of connections of 2.1 mm Acquity BEH C18 columns was facilitated by very short stainless steel tubing. The test mixed was analysed at the maximum flow rate that corresponded to maximum pressure - just below the pressure limit. Polaratherm Series 9000 column heater with force air circulation was used to maintain column temperature at 40°C. In order to avoid void volumes, the built-in solvent pre-heater and cooler were by-passed. The data was acquired at 20Hz rate.
2.4 Construction of kinetic plots
• The following equation was used to calculate the maximum pressure in Pa to construct logo/N versus logN kinetic plots:
ΔP = фηN (A (uo⁴´³)/(Dm¹´³dp²´³)) +B(Dm/dp²) + C(uo²/Dm)
Experimental data was entered into the Knox equation to calculate A, B and C.
h = Aν¹´³ + B/ν + Cν
3. Results and discussion
3.1 Evaluation of Acquity 1.7 μm columns
• Xbridge and Acquity BEH columns were chosen because they’re packed with almost the same stationary phase to keep retention time differences to a minimum.
• Experimental data obtained for acetophenone (k = 3.6 – 3.9) on Acquity BEH, 1.7 μm and on XBridge 3.5 μm and 5 μm with 0.5mL/min flow rate and a system pressure of 675 bar was used to construct van Deemter plots to compare performance of the columns.
• Analysis of van Deemter plots demonstrated that the highest optimal velocity was observed for the smallest particle size, and that at higher velocities the appearance of the plot was relatively flat. This confirmed theoretical concepts that the achievement of desired characteristics for mass transfer would be facilitated by the increase of the optimal velocity when the particle size is reduced, and that increase in the linear velocity above the optimum value would result in insignificant reduction in efficiency.
• Comparisons with UPLC results in the literature reveal that similar to 100 x 2.1mm, 1.7 μm column’s minimum plate height values were obtained for 150 x 1 mm, 1.5 μm column, and even lesser values for non-porous particles < 1.5 μm.
• Compared to 5 μm particles, 1.7 μm particles offer higher efficiency and higher optimal linear velocity. In order to keep the same efficiency on a shorter column, analysis time has to be increased. For faster analysis, when operated at higher than optimal linear velocities, insignificant lose in efficiency would be expected.
• Columns can also be compared using Knox plots with reduced plate height versus reduced linear velocity. This comparison should not depend on particle size.
• However, reduction in particle size resulted in greater values for minimum reduced plate height. Disagreement of reduced plate height values for 3.5 μm and 5μm particles was attributed to variation in packing efficiency due to troublesome process of manufacturing/packing of smaller particles.
• Further evaluation of possible causes contributing to increase in reduced plate heights for 1.7 μm particles suggested that there are at minimum two contributing factors: 1) inadequacy in packing efficiency for columns with IDs below 4.6 mm, and particle sizes less then 5 μm, and 2) the formation of radial temperature due to gradient frictional heating at high pressures.
• Investigation of the effect of weak radial temperature gradient by measuring efficiencies for the non-isolated and isolated in polystyrene foam column revealed the loss in efficiency for the non-isolated column operated at high (close to the maximum) system pressure, and no change in efficiency for the columns operated at optimum flow rate.
• It makes sense that the optimal flow rates, and as the consequence reduced pressures, result in lower amount of frictional heat. However, changes in analyte diffusion coefficients and the mobile phase viscosity resulted from frictional heating and affected by high pressures are more than likely to have a significant impact on the reduced plate height.
• As a result, all contributors discussed above (and extra-column volume) are reflected in the values of the coefficients A, B, and C of Knox equation. It was observed that different compounds with different retention times (weakly retained vs. highly retained) exhibit, so far unexplainable, significantly different behavior in terms of optimal reduced linear velocities and reduced plate height.
• In order to operate 1.7 μm particle columns at high optimal flow rates, high pressures are necessary, along with reduced system volumes, low-volume flow cell and a data collection rate of no more then 40 Hz.
3.2 Evaluation of possibilities and limitations of UPLC using kinetic plots
• The disadvantage of van Deemter and Knox plots is that they don’t take into account backpressure limitations (degree of column permeability) that would allow more accurate comparison between columns.
• The best method to graphically visualize the compromise between speed, efficiency and column back pressure is a Poppe plot of plate times, =H/uo, or to/N, versus plate number, N.
3.2.1 Effect of increased pressure capabilities
• From Poppe kinetic plots it can be determined at what conditions the fastest analysis can be achieved for the required efficiency on a column with the specified particle size. However, there is a limit to the efficiency that can be obtained for at a specific maximum pressure and temperature.
• Further evaluation of kinetic plots reveals that fast analyses requiring low efficiency can be achieved at pressures below 1000 bar. Increase in the maximum pressure will result in higher efficiency, and higher efficiency in its turn always requires longer analysis time. From the data present it is evident that operation of short columns with 1.7 μm particles under no more than 400 bar pressures will result in fast, but low efficiency analysis.
• Theoretically, pressure increase would produce the same effect on columns packed with 3.5 μm and 5 μm columns.
3.2.2 Comparison of UPLC with conventional HPLC
• Comparison of kinetic plots for columns with 3.5 μm and 5 μm particles used at 400 bar and 1.7 μm particles columns at 1000 bar aids evaluation of dependency of efficiency and analysis speed on particle size and operating pressure.
• Smaller particle size columns provide faster analyses, but depending on the efficiency level required, analysis on the bigger particles size columns maybe faster than on smaller particle size column due to pressure limitations. So that an analysis would be performed faster on 1.7 μm particle size column at 1000 bar for up to 68,000 plates, or on 5 μm particle size column at 400 bar for beyond 68,000 plates than on 3.5 μm particle size column at 400 bar.
• Thus, UPLC offers major advantage in the speed of analysis than HPLC.
• Smaller particle size columns offer reduced column heating requirements and reduced consumption of solvent.
• In addition, UPLC offers high resolution (e.g. for complex sample mixtures) since longer columns can be used with smaller particles at high pressures. For ~ 70,000 – 80,000 theoretical plates, UPLC offers faster analysis than HPLC, but higher efficiency will require larger particle size.
• The range of acceptable analysis time can be estimated from the maximum efficiency points on the kinetic plots where 1.7 μm particles offer shorter analysis time than 3.5 μm and 5 μm particles. If operated within this range, shorter analysis time will always be obtained on UPLC.
3.2.3 Experimental confirmation of kinetic plots
• Supporting experiments were conducted on UPLC to confirm observations made with Poppe plots.
• Slight deviation of experimental data, obtained for columns (with the lengths of 50 and 100 mm) thermostated by Acuity column heater, from theoretical was noted. This was attributed to the differences between the maximum pressure used for theoretical calculations and the actual pressure obtained during experiments.
• Deviations observed for the columns (with the lengths of 200 mm and longer) thermostated by Polaratherm were explained by the differences in the instrument set-up such as an increase in system volume due to additional connecting tubing between columns, increase in mobile phase viscosity since it is no longer constant at high pressures and a radial temperature gradient due to favorable heat transfer.
• These experimental data deviations from theoretically calculated values are not significant, and therefore previous theoretical conclusions are justified.
• There were no discrepancies between the data obtained on single columns and on tandem columns.
• At high pressures, long columns with small particles provide high efficiency with adequate analysis time, which is not possible with HPLC instrumentation.
• Analysis time can be further reduced by operating columns at higher temperatures at fixed maximum pressures. Expected temperature effect would be directed towards lowering to/N numbers in the kinetic plot, which would result in approximately the same analysis time for all particle sizes.
4. Conclusions
• UPLC was recognized as a superior technique compared to HPLC due to its high efficiency and short analyses times. Even such flaws as radial temperature gradient, imperfections in small particle packing efficiency and extra column band broadening do not compromise its excellent performance.
• Furthermore, kinetic plots constructed from experimental data for UPLC versus HPLC comparison demonstrated that a combination of higher pressure and smaller particle size is beneficial in speeding up the analyses without compromising the efficiency.
References
• 31 references were used in the paper
Summary of the article provided by A.Caltabiano for Chemical Information Retrieval Class at Drexel University - Assignment #1 on www.getcheminfo.wikispaces.com/assignments
