CHROMATOGRAPHY WITH OTHER EXISTING TECHNOLOGIES FOR FAST AND HIGH-RESOLUTION LIQUID CHROMATOGRAPHY
5. METHOD TRANSFER FROM HIGH-PRESSURE LIQUID CHROMATOGRAPHY TO ULTRAHIGH-PRESSURE LIQUIDCHROMATOGRAPHY TO ULTRAHIGH-PRESSURE LIQUID
CHROMATOGRAPHY
In various fields of application (i.e., pharmaceutical, environment, food, etc.), it is essential to be able to transfer existing methods (performed in conventional HPLC conditions) to faster separa- tions involving the use of columns packed with sub-2mm particles. As most providers now offer equivalent stationary phases packed with 5, 3, and sub-2mm particles, a geometrical transfer can be performed if the stationary phase chemistry remains identical between the original and final sets of conditions. For this purpose, some rules have to be strictly applied in both isocratic and gradient modes.
5.1 THE RULES FOR ISOCRATIC MODE d THEORY AND APPLICATIONS
For an isocratic transfer between conventional HPLC and UHPLC, two important parameters have to be adjusted, namely the injection volume and the mobile phase flow rate (Guillarme et al., 2007a,b,c).
To avoid a detrimental extra-column band broadening and maintain equivalent sensitivity, it is necessary to adapt the injection volume in agreement with the change of column dimensions. In LC, the injected volume should represent only 1%e5% of the column volume. The latter should be adjusted proportionally to the column internal diameter (dc) and length (L). Therefore, the injection volume is independent of the particle size and only proportional to the column volume. The new injected volume
Vinj2
can be determined simply by maintaining the ratio of column dead volume and injected volume constant between regular HPLC and UHPLC.
Vinj2¼Vinj1$d2c2 d2c1$L2
L1
(1.7) In this equation, subscripts 1 and 2 are related to HPLC and UHPLC column dimensions, respectively. For example, from a conventional 1504.6 mm, 5mm column to a UHPLC 502.1 mm, 1.7mm column, the injected volume should be decreased by 14-fold. To maximize sensitivity, it is possible to inject larger volume but in this case, the sample should be dissolved in a solvent of weaker eluent strength than the initial mobile phase composition (sample focusing).
Regarding mobile phase flow rate, this parameter should be adapted to remain close to maintaining a constant reduced mobile phase linear velocity (v). In HPLC, it is well known that the mobile phase linear velocity (u) is directly proportional to the square of column diameter and also depends on the particle size (dp) of the support. It is however, completely independent of the column length. For a successful method transfer, it is mandatory to maintain the product udp constant, to take into
account simultaneous changes in column diameter and particle size of the support. Therefore, for a geometrical transfer, the UHPLC flow rate (F2) can be calculated with the following equation:
F2¼F1$d2c2 d2c1$dp1
dp2
(1.8) As an example, from a regular 1504.6 mm, 5mm column to a UHPLC 502.1 mm, 1.7mm column, the mobile phase flow rate should be decreased by 1.6-fold.
The expected analysis time of the transferred methodðtana2Þis directly proportional to the change in column dead time and can be estimated according to:
tana2¼tana1$dp2
dp1
$L2
L1
(1.9) The expected backpressure (DP2) can be calculated from Darcy’s law, which shows thatDPis inversely proportional tod3p (at the optimal linear velocity) and is strictly related to the column length:
DP2¼DP1$L2
L1$d3p1
d3p2 (1.10)
Finally, the expected solvent consumption of the transferred method (V2) can be calculated by taking into account the change in internal diameter and column length.
V2¼V1$dc22 dc21$L2
L1
(1.11) Therefore, from a regular 1504.6 mm, 5mm column to a UHPLC 502.1 mm, 1.7mm column, the analysis time is reduced by ninefold. For the abovementioned transfer, the efficiency would be identical, while the backpressure should be ninefold higher and the solvent consumption reduced by 14-fold. This shows the obvious benefits of the UHPLC strategy.
It is possible to find in the literature a large number of applications showing the possibility to transfer isocratic HPLC methods to columns packed with sub-2mm particles, providing that the chemistry of the two analytical supports remains identical. One example is reported inFig. 1.9 and presents a method transfer from a conventional 1504.6 mm, 5mm column to a UHPLC 502.1 mm, 1.9mm column (Russo et al., 2008). Both columns provide an equivalent efficiency of around 10,000 plates (similarL/dpratio). As shown in the chromatograms, efficiency, selectivity, and resolution remain equivalent for the separation of seven common anxiolytic agents. After adjustment of mobile phase flow rate, the analysis time is decreased by a factor of 7 (22 vs. 3.2 min), as expected from theory for a transfer from 5 to 1.9mm particles.
5.2 THE RULES FOR GRADIENT MODE d THEORY AND APPLICATIONS
The rules for gradient method transfer between conventional HPLC and UHPLC are not only more complex than isocratic ones but also based on the basic principles of chromatography. First, the injection volume and mobile phase flow rate should be adapted in a similar way as the isocratic mode (seeEqs. 1.7 and 1.8) (Guillarme et al., 2008).
In linear or multilinear gradient elution, the gradient profile can be decomposed as the combination of various isocratic and gradient segments. The rules for efficient gradient transfer originally established by Snyder and Dolan (1998)should be strictly followed. For both parts, it is important to scale the gradient volume in proportion to the number of column volumes to yield identical elution patterns, whereas the initial and final compositions should be constant. In fact, the number of column volumes percolated during the gradient in the regular HPLC system should be equivalent to that of the UHPLC setup.
For any isocratic step within the gradient (i.e., initial isocratic step, isocratic step during a multilinear gradient and also reequilibrating time), the ratio between the isocratic step time (tiso) and the column dead time should be maintained equivalent between conventional HPLC and UHPLC conditions. Therefore, the UHPLC isocratic stepðtiso2Þcan be determined using:
tiso2¼tiso1$dp2
dp1
$L2
L1
(1.12) As an example, from a regular 1504.6 mm, 5mm column to a UHPLC 502.1 mm, 1.7mm column, the isocratic steps which occurred during the gradient process should be reduced by ninefold.
Hypersil Gold 4.6 x 150 mm, 5 µm
F = 1000 µL/min Vinj= 20 µl
ORIGINAL METHOD
0 2 4 6 8 10 12 14 16 18 20 22
Minutes 0
5 10 15 20 25
mV
1 2
3 4
5 6
7
Hypersil Gold 2.1 x 50 mm, 1.9 µm
F = 550 µL/min Vinj= 1.4 µL
TRANSFERRED METHOD
22 min
1
2 3
4 5 6 7
0.00
Minutes 0.00
0.02 0.04 0.06 0.08
0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20
0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20
AU 3.2 min
(A)
(B)
FIGURE 1.9
Isocratic method transfer from regular high-pressure liquid chromatography to ultrahigh-pressure liquid chro- matography. Separation of a benzodiazepines mixture in isocratic mode with a mobile phase containing ACNewater (31:69, v/v) with 0.1% formic acid, T¼30C andl¼254 nm. (A) Hypersil GOLD 1504.6 mm, 5mm,F¼1000mL/min,Vinj¼20mL. (B) Hypersil GOLD 502.1 mm, 1.9mm,F¼550mL/min,
Vinj¼1.4mL.
For slope segments, it is mandatory to keep the initial and final gradient composition (%B) constant. The new gradient timeðtgrad2Þcan be expressed as:
tgrad2¼ð%Bfinal1%Binitial1Þ slope2
(1.13) The gradient slope (slope2) should be calculated to maintain the product of gradient slope and column dead time constant. The new gradient slope (slope2) can be expressed as:
slope2¼slope1$d2c1 d2c2$L1
L2$F2
F1
(1.14) As an example, from a regular 1504.6 mm, 5mm column to a UHPLC 502.1 mm, 1.9mm column, the gradient slope during the gradient process should be increased by ninefold.
When transferring a gradient method from regular HPLC to UHPLC, some changes in selectivity could occur during the gradient run because of differences in dwell volume between the original and the UHPLC configuration. The system dwell volume (Vd) refers to the volume between the mixing point of solvents and the head of the analytical column (Dolan, 2006). After starting the gradient, it will take time until the selected percentage of solvent reaches the column. Because the gradient dwell volume may differ from one system to another, this extra isocratic step would be different and could result in retention time variations affecting resolution for early eluting peaks when transferring a method. To overcome this problem, the ratio of system dwell time (td) and column dead time (t0) must be held constant while changing column dimensions, particle size or mobile phase flow rate.
As the column dead time is reduced by around ninefold between a regular 1504.6 mm, 5mm column and a 502.1 mm, 1.7mm column, the system dwell time should be reduced by the same factor. As mentioned previously, it is then mandatory to work with a UHPLC system possessing a low dwell volume (no more than a few 100mL) to limit its influence.
Again, many applications of the above discussed approach can be found in the literature. One example has been selected, and it is presented inFig. 1.10. The original separation of 12 pharma- ceutical compounds was achieved using a 1504.6 mm, 5mm, C18 column and subsequently transferred to UHPLC with a 502.1 mm, 1.7mm column possessing strictly similar chemistry (Guillarme et al., 2008). The original separation was performed in w27 min and was efficiently transferred to UHPLC, giving a separation that was complete in less than 3 min (reduction by a factor of 9, as expected from theory). In addition, both separations were equivalent in terms of sensitivity, peak capacities, and resolution, mainly because of an adequate reduction of system dwell volume (from 1 mL to 100mL for HPLC and UHPLC, respectively). Finally, the mobile phase flow rate can be increased up to the maximal operating pressure of instrumentation to further reduce analysis time. As illustrated inFig. 1.10C, at 1000mL/min the analysis time was cut to only 1.6 min but with a slight loss (about 10%) of chromatographic performance.