5
capable of 1000 bar or more. In UHPLC, the very large pressure drops required result in an increase in the mobile phase temperature [1-3], due to the expansion of the mobile phase. Small ID columns, such as 2.1 or even 1 mm, are thought necessary to minimise the length of radial thermal gradients, caused by the decompression of the mobile phase. Radial gradients potentially cause serious loss of effi ciency. Smaller ID columns require smaller optimum fl ow rates.
Supercritical fl uid chromatography (SFC) is inherently 3 to 5 times faster than HPLC on the same sized columns, due to lower viscosity [4] and higher diffusivity of/in the mobile phase. Pressure drops are 1/3rd to 1/5th those encountered in HPLC. Thus, pressure drops are usually no more than a few hundred bar. Since the pressure drops in SFC are much lower, high pump pressure capability is much less important, compared to UHPLC.
Modifi ed CO2 mixtures can cool, warm
or maintain the same temperature when expanded [5], depending on modifi er concentration. The combination of low pressure drops and modest levels of warming or cooling seldom, if ever, distort peaks or cause losses of effi ciency. Thus, the potential for forming signifi cant radial thermal gradients in SFC is much lower than in UHPLC. One would think that the increasing use of sub-2 µm particles in UHPLC, with the associated large pressure drops, and the resulting large thermal gradients should stimulate the development of ultra-high performance supercritical fl uid chromatography (UHPSFC), since in SFC those issues are greatly reduced. However, there have been no attempts by the vendors of such equipment toward developing true UHPSFC.
Minimising thermal gradients has been the main justifi cation for the use of 2.1 mm columns in UHPLC, but since such gradients are much less likely in SFC, the need for such columns has less justifi cation. Columns 3 mm ID, with the same sub-2 µm packings, have 4X the dispersion compared to 2.1 mm columns, allowing 4X higher extra column dispersion in the system. It appears that 3 mm ID columns are, at present, easier to pack, and more effi cient, compared to 2.1 mm columns. Further, a 2.1 mm ID column can tolerate only ≈ 1/2 the injection volume of a 3 mm ID column. Never the less, the future development of SFC instrumentation ought to aim for the use of sub-2 µm particles in 2.1 µm ID columns with minimal extra-column dispersion (hmin
≈ 2), even
though there is little current compelling justifi cation for such development.
It appears to be a common perception that columns are inherently well packed, supported by the fact that column technology has largely outstripped instrumentation. It is clear that the SFC instrument confi gurations shipped from the manufacturers are inadequate for use with sub-2 µm particles. However, many of the modifi cations used to improve HPLC instruments to UHPLC performance, such as shorter, smaller ID tubing, and smaller detector cell volumes have resulted in confusing and anti-intuitive results when used in SFC. A major problem is that it is unclear, a priori, how well columns are actually packed, and whether any poor effi ciency observed is due to the column, the instrumentation, or both.
There have been a modest number of attempts to characterise the improvements needed to achieve the goal of hmin
≈ 2 with
sub-2 µm particles. This report summarises those attempts.
Guillarme
In 2012, Guillarme [6] compared a Waters UHPLC to a Waters UPC2 which he called a UHPSFC. Both chromatographs were used as plumbed by the manufacturer. The UHPLC was found to have an extra-column variance of ≈ 3 µL2
min-1
for 3 mm ID columns and 1 mL-min-1 in 2.1 mm columns. Knowing column length allows calculation of retention times and peak widths. At k = 2, h = 2, the dispersion of a 3x100 mm column is ≈ 73 µL2
, whereas a
2.1x100 mm column has a dispersion of only ≈ 17.5 µL2
. A rough ‘rule of thumb’ says that extra-column dispersion should be < 1/5th column dispersion. This suggests that the extra-column dispersion should be < 3.5 µL2 when using a 2.1x100 mm column with 1.8 µm particles.
Guillarme stated that although 2.1 mm columns tend to be favoured in UHPLC, they could not be used in SFC due to the high extra-column dispersion of the commercial systems. He proceeded to compare the performance of the 2 instruments using 2.1 mm ID reversed phase columns in UHPLC, but 3 mm ID normal phase columns in UHPSFC, all with 1.7 µm packings. The SFC was found to be 4 times faster, but had a minimum reduced plate height (hmin while the UHPLC had hmin
) > 2.8, ≈ 2.2. The 3x100
mm column should have 4X the column dispersion, allowing 4X higher extra- column dispersion, compared to the 2.1 mm column, but still exhibited substantially worse effi ciency. Guillarme made no attempt at improving the performance of the SFC, through changing the tubing or UV fl ow cell, but he clearly outlined the major differences between the UHPLC and the SFC.
. The tubing between the
injection valve and the column was 250 mm long, 130 µm ID, including a passive heater. The tube between the column and detector was 150 mm long, 100 µm ID. The injector had a 5 µL loop. The detector fl ow cell was 0.5 µl.
On the other hand, the SFC had a nearly 30X higher variance of ≈ 85 µL2
, with 600 mm
of 175 µm tubing before the column and 600 µL of 175 µm tubing after the column. The injector had a 10 µL loop. The fl ow cell was 8 µL. This very high extra-column dispersion in the SFC negates any pretence that the results could be called ‘ultra-high performance’. However, most other current commercial SFC’s have similar, very high extra-column dispersion [7] as plumbed and shipped by their manufacturers, and are similar to conventional HPLC’s and SFC’s in the 1990’s.
It is informative to calculate the theoretical dispersion of various column sizes in order to estimate the extra-column dispersion allowed. Assuming a particle size of 1.8 µm, the optimum fl ow rate on various ID columns can be estimated from experience at ≈ 2 mL-
Broeckhoven
On the other hand, Broeckhoven stated [8] that 2.1 µm ID columns must be used in SFC in order to achieve high velocities with commercial instruments, due to the high optimum fl ow rates of the columns and limited maximum fl ow rates of the pumps. However, he also agreed that the extra-column dispersion was excessive, as shipped by the manufacturer. Much of the extra-column dispersion occurs in the connector tubing and to a lesser extent in the detector fl ow cell. The instrument used was an Agilent 1260. Attempting to minimise such extra-column dispersion, he varied the composition of the sample solvent, the injection volume, and the size and length of the pre- and post-column tubing, and detector volume.
Tubing ID in front of the column was decreased from 175 µm to 125 µm, and the shorter (125 mm), built-in heat exchanger (HX) was used (with 175 µm internal tubing). The standard 13 µL detector fl ow cell was replaced with a 1.7 µL cell, with a 310 mm long, 125 µm inlet tube, and eventually with a 0.6 µL fl ow cell.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68
