Investigating the performance attributes of an ultra-high pressure liquid chromatography system

Methods

Material and solutions: Injection linearity and precision were investigated using a Supelco HPLC gradient system diagnostic mix (P/N 4-8271) containing uracil (dead volume retention marker), phenol, methyl paraben, ethyl paraben, propyl paraben, butyl paraben and heptyl paraben. For carryover determination, caffeine (Millipore Sigma P/N C0750), chlorhexidine (Millipore Sigma Pharmaceutical Secondary Standard; Certified Reference Material P/N: PHR1421) and amitriptyline (Cerilliant P/N: A-923 1.0 mg/mL in methanol, ampule of 1 mL, certified reference material) were used. The following solutions were prepared: caffeine, 4 mg/mL and 10 µg/mL in methanol/water (20%/80%; v/v); chlorhexidine, 500 ng/µL and 1.25 ng/µL in water + 0.1% formic acid; amitriptyline, 285 ng/mL and 1 ng/mL in methanol/water (20%/80%; v/v). For flow rate precision determination, a total of 18 drugs and 11 deuterated internal standards were obtained from Cerilliant and combined in one solution.

Chromatography: LC separations were performed using the SCIEX ExionLC 2.0 system. For injection linearity and precision determination a 2.6 μm Phenomenex Kinetex C18 column (2.1 x 50 mm, P/N: 00B-4462-AN) was chosen, and a simple gradient of water and acetonitrile, both containing 0.1% formic acid, was used. The analytical run including equilibration was 10 minutes to ensure maximum reproducibility. The syringe speed was set to low and the speed factor to 0.1. For caffeine and chlorhexidine carryover determinations, a flow restrictor (15 m yellow PEEK tubing, 0.18 mm ID) with 2 low dead volume unions was used to ensure measurements are made with a back pressure of about 140 bar. For amitriptyline carryover and flow-rate precision determinations (using 18 compound drug mixture), a 2.6 μm Phenomenex Kinetex Phenyl-Hexyl column (4.6 x 50 mm, P/N: 00B-4495-EO) was chosen. Mobile phase A was ammonium formate in water. Mobile phase B was formic acid in methanol. The LC flow rate was 1 mL/min and the LC run-time was 6.5 minutes. The SCIEX ExionLC 2.0 system autosampler was used in the standard configuration consisting of a 250 µL syringe, 100 µL sample loop, 250 µL buffer tubing and 15 µL needle tubing. For all injections, the µL pick-up plus mode was selected in order to minimize the injection cycle time whilst optimizing sample consumption.

Mass spectrometry and diode array detector conditions: Ultraviolet-visible detection was performed using an ExionLC 2.0 system with integrated ExionLC 2.0 diode array detector HS, equipped with a 10 mm, 10 μL, 300 bar flow-cell. For the Supelco HPLC gradient system diagnostic mix, the detector was operated at 254 nm and, for the caffeine experiments, at 272 nm. All diode array detector data collection was collected with a data rate of 10 Hz.

Mass spectrometry used a SCIEX Triple Quad 5500 system. The ionization source was operated using electrospray ionization (ESI) in positive mode. Three MRM transitions were monitored for chlorhexidine and two for amitriptyline.

Data acquisition was performed using Analyst software 1.7.1 with Components for the ExionLC 2.0 system. It is worth noting that The ExionLC 2.0 system is also fully supported for instruments in which data acquisition is performed using SCIEX OS software.

Data processing: Data processing of both mass spectrometry and diode array detector acquired data was performed using SCIEX OS software 2.0.1 in which calibration curves, precision and accuracy statistics were generated.

Injection linearity and precision

First a mixture of compounds was analyzed using the integrated ExionLC 2.0 diode array detector HS for detection. Very good separation of the components was achieved (Figure 2). 

Replicate injections (n=5) were performed across a broad injection volume range of 0.5 and 25 µL using the µL pick-up plus injection mode. The area counts were measured and plotted against corresponding injection volumes for each component in the Sulpelco HPLC gradient system diagnostic mix.

Figure 3 shows that the linear coefficient of determination (r²) is > 0.999.

The injection reproducibility was also computed from this same data, for all the various injection volumes. Very low variance was observed, even for the very low injection volumes. Figure 4 shows that the %CVs for all injection volumes are below 1%.

When optimizing sample consumption, the µL pickup plus injection mode is the preferred option as no excess sample is utilized. The injection sequence has been optimized so that the time to injection is not compromised over other injection modes which can be made in less than 17 seconds.     

Injection carryover determination

For determining carryover, 3 different compounds were selected that are prone to adsorption: caffeine, chlorohexidine and amitriptyline. For each compound, carryover was determined using the following sequence of 6 injections: Blank 1, Low concentrated standard, High concentration (highly concentrated standard, exceeding the linear range of the detector), Blank 2, Blank 3, Blank 4. The following equation was used to calculate carryover.

As shown in Figure 5, carryover for caffeine was 0.0006% with a simple wash cycle in the UV experiment. 

The more challenging experiment is using the chlorhexidine compound, which is notoriously ‘sticky’. Here, carryover was measured using the more sensitive MS experiment. Multiple wash solutions and extended needle washes were required to achieve desired results in this case, but the flexibility of the autosampler enabled construction of an optimized wash method. The total wash time is dependent on the total volume used during each wash step. Varying wash solvent and transport solvent conditions were investigated to minimize total wash time while providing minimal carryover. Figure 6 shows the chlorhexidine carryover was calculated to be 0.0005% using the optimized wash volumes.

For additional carryover testing, 1 ng/mL and 285 ng/mL amitriptyline solutions were prepared in 10% acetonitrile, the latter providing approximately 20 million area counts for the most intense MRM transition in the MS experiment. Figure 7 shows the amitriptyline % carryover from the first blank following the high concentrated standard was determined to be 0.01% using a minimal volume wash solvent composed of 20% acetonitrile, 20% methanol, 40% water and 20% isopropanol by volume with 0.1% formic acid.

Flow rate precision

Figure 1 shows the retention time stability with overlaid chromatograms from 100 consecutive injections. As shown in Figure 8 (top), the retention time precision of each of the analytes across a range of retention times for these injections is less than 0.3% RSD. For most compounds tested, the maximum retention time difference over 100 injections was <1 second (Figure 8, bottom). 

Conclusions

The ExionLC 2.0 system is a flexible and robust UHPLC system that is suitable for today’s challenging LC-MS workflows.

• High autosampler precision and accuracy are essential for quantitative experiments. As demonstrated here, this system can achieve <1% CV across the full range of injection volumes tested (0.5 – 50 µL) with excellent linearity (linear correlation coefficient (r) regression analysis was > 0.999 for all analytes).
• Low carryover is also critical in many LC-MS applications. The flexible autosampler wash options (with multiple wash solutions and extended needle wash capabilities) were shown to reduce carryover, even for a very challenging analyte, down to 0.0005%.
• Retention time precision can also be critical when running multi-analyte panels and time scheduled MRM assays. High flow rate precision of the ExionLC 2.0 system provided RSD of <0.3% for the analytes tested here.