Automated Chromatographic Solid-Phase Extraction Using an Autosampler

Solid-phase extraction (SPE) is a preferred tool for isolating target analytes from complex matrices because the availability of a diverse range of chromatographic sorbents enables targeted approaches based on the specific chemistry of the analytes and matrices. Also, SPE offers the ability to enrich or preconcentrate analytes in samples. Enrichment is valuable because it allows one to match the analyte concentrations to the approach used to measure them. Given these unique capabilities, SPE is considered the “gold standard” in analytical sample preparation.

SPE using a cartridge on a vacuum (or pneumatic pressure) manifold requires proper training, as care must be taken to maintain appropriate liquid flow in real time. As a result, the number of samples that can be processed is limited to 30 per day per technician. If enrichment is needed, a sample dry-down step is required, which reduces throughput.

A parallel approach with tubes, pipet tips, cartridges or 96-well plates increases the number of samples, but at a cost. Parallel liquid flow is highly variable from tube to tube or well to well. Increased variability in parallel flow results in additional variability in analyte recovery, requiring the use of internal standards and acceptance of lower quality, specifically, higher LLOQs (lower limits of quantification) and inconsistent assay results. Also, this approach only addresses sample/solvent flow over the SPE sorbent—not pipetting sample/solvents to each well. Thus, going parallel only modestly increases sample throughput, but also the probability of human error due to the greater number of pipetting steps.

To increase SPE throughput, a parallel approach can be coupled with robots to perform the pipetting (robotic SPE, then LC/MS/MS or GC/MS/MS measurement); however, results are lower in quality (for variable parallel flow reasons given above) in this two-workflow process. Another drawback is the additional cost for robots ($200–500k), programming/integration and technicians with robotic experience. From a results point of view, this seems to mirror the “speed, cost, quality triangle,” where gains in any one of the three take away from the other two. Breaking through the “speed, cost, quality triangle” requires fundamental change in the SPE device to enable both automation and higher quality using only existing analytical instrumentation instead of high cost robots.

Introduction of precision-controlled micro-SPE using the autosampler

Central to automated, rapid, and higher-quality SPE is a patented, single-use micro cartridge (ITSP, Hartwell, Ga.)^1-5 (Figure 1) containing user-defined, packed chromatographic media. The crimped-on septum and needle guide (upper 80% of cartridge) enable automation by facilitating accurate cartridge transport on a syringe needle. Automated SPE begins by using the autosampler syringe for cartridge conditioning, sample loading and sample washing over a waste receptacle. The syringe is used to perform elution over a clean vial or well, and the used cartridge is discarded in a different waste receptacle. Automation is complete after the syringe mixes the freshly eluted sample and then injects it into the LC/MS/MS or GC/MS/MS that will be used to measure the sample (see the CTC/PAL autosampler in Figure 2).

Figure 1 – Internal view of an ITSP SPE single-use cartridge and how it interfaces with the PAL autosampler syringe and needle support. This facilitates all SPE cartridge transport and delivery of sample and solvents.

Figure 2 – PAL RTC autosampler (CTC Analytics, Lake Elmo, Minn.) equipped for online SPE-LC/MS/MS. Hoses connect the SPE cartridge tray and syringe wash station to a laboratory solvent waste container. Used cartridges are typically discarded by the PAL into a box under the wash station and LC valve.

This simple, automated, single-workflow process integrates SPE directly into the LC/MS/MS or GC/MS/MS software as an on-line workflow, and does not require additional skills beyond those needed to operate the LC/MS/MS or GC/MS/MS. It is simply a different method selection in the MS software already in use with any LC/MS/MS or GC/MS/MS equipped with a PAL autosampler.

Figure 3 – Execution timeline for SPE in parallel with LC/MS/MS for the analysis of 71 drugs in urine. In this process, overnight measurement of two 96-well plates of samples per LC/MS/MS is routine and results are ready for review when the lab opens in the morning.

In the single-workflow procedure, SPE and LC/MS/MS (or GC/MS/MS) are performed serially, but in parallel with each other (see Figure 3). As shown, cycle time is frequently a function of the LC/MS/MS (or GC/MS/MS) measurement and not the SPE. With this form of parallelism, the only cost in time is SPE of the first sample. By combining the processes it is considerably faster than the two-workflow process and requires nothing more than a good initial choice in autosampler in order to achieve complete automation.

The top of the ITSP micro-SPE cartridge enables precise automation, while the bottom of the cartridge performs chromatographic SPE, cleaning and enriching the sample to produce high-quality results and maintaining the LC/MS/MS or GC/MS/MS to ensure long-term, robust operation. As shown in Figure 1, the syringe needle is placed at the frit containing the packed chromatographic sorbent to transfer sample, without significant dispersion, directly to the packed sorbent. Flow of the sample and solvents is positive liquid pressure, syringe pump-driven (flow adjustable at a resolution of ±10 nL/sec) and is precise and accurate. The extra-column volume of the cartridge (Figure 1, below sorbent) is just 16 μL.

Elution volumes of 50–100 μL facilitate control of the analyte concentrations delivered for measurement. As a result, sample volume can be adjusted to match the capability (sensitivity) of the measurement instrumentation and the necessary cutoffs (LLOQs, S/N ≥20) for proper interpretation of test results. This is achieved without the customary dry-down step used with vacuum or pneumatically driven flow forms of SPE, which tend to require ≥5× larger elution volumes.

This concept is best illustrated with common, in-use assays. For broad-panel drug measurement in urine samples using a mid-range LC/MS/MS, 225 μL of sample is loaded onto the SPE cartridge and is then eluted with 75 μL. This 3× analyte enrichment achieves the necessary 1-ng/mL cutoffs for low-dose drugs and produces LC peaks for all drugs—sufficiently intense for automatic integration across the concentration ranges observed in real samples. For measurement of the same broad-panel drug assay in oral fluid samples, 1 mL of sample is loaded onto the SPE cartridge and is then eluted with 75 μL. This 13× enrichment provides the 0.2-ng/mL cutoffs for low-dose drugs. For measurement of priority pollutants in drinking water, 10 mL of water sample is loaded onto the SPE cartridge, followed by 50 μL elution to achieve a 200× analyte enrichment and LLOQs at or near single-digit parts-per-trillion levels. In all of these examples, the analytes are ready to measure immediately after SPE without performing dry-down of the eluent.

The ITSP micro-SPE cartridge provides accurate syringe pump flow control over the packed sorbent bed, allowing SPE separations to be performed at their van Deemter optimum velocity⁶ (see Figure 4). This finding was unanticipated, because with single-use SPE devices it is expected that lower flow⁷ yields higher recovery. The data clearly shows that accurate flow control and optimization are equally important for SPE and LC. With method optimization experiments performed in the same ways as LC, SPE recoveries are systematically high and precise (not common with other forms of SPE, particularly for ion exchange and chelation mechanisms). The accuracy and variance in the final test results depend most on the LC/MS/MS or GC/MS/MS measurement (results CV = 3–5%) rather than the recovery of SPE sample preparation performed by the CTC PAL autosampler (Figure 2).

Figure 4 – van Deemter curves for reversed-phase (C₁₈ endcapped) SPE using two different particle diameters (10 and 50 μm). The image also shows the source of variable recovery with vacuum or pneumatically driven flow, because as flow drifts away from the optimum, recovery suffers. Δ%Recovery squared (relative variance, σ², in the amount of sample recovered) is the dimension plotted on the y-axis; the numbers indicated on the y-axis are absolute %recovery (not adjusted based on IS) as an aid to the reader. This is analogous to the %RSD approach to determining plate height described by Neue,⁶ but is used here only for flow optimization.

Additional capabilities

Automated method development/optimization experiments can be carried out for multiple sorbents and solvents as a series of 5–6 run lists (Design of Experiments [DOE]), where each list provides optima fed into the next list. The sum of these lists (including measurement/elimination of breakthrough of all condition, load and wash steps; flow optimization; and a sample loading study) can be measured in three days to result in a highly optimized SPE method. ITSP SPE achieves precise chromatographic separations, preconcentration of sample (without dry-down), robust operation and total automation.

PAL-based SPE does not limit one to a single dimension of SPE. For lipid profiling, uncharged oils can be isolated from fatty acids and phospholipids using anion-exchange SPE. The phospholipids can then be isolated from the fatty acids using chelation in a second SPE step (easily achieved with a PAL autosampler). Finally, each of the three isolated samples (oils, fatty acids and phospholipids) can be directed to the LC/MS/MS or GC/MS/MS measurement approach best suited for that lipid class. Similarly, for proteomics, reversed-phase SPE can be used for desalting, followed by a second chelation SPE to isolate peptides from phosphopeptides for separate LC/MS/MS measurement of the large numbers of peptides. In these examples of 2-D SPE, the PAL is often operated as a standalone sample preparation device; multiple instrument types and/or methods can process the different compound classes for the same samples in parallel. This is also useful in applications in which the analytical measurement time significantly exceeds the SPE time. For example, in the GC/MS/MS measurement of large pesticide panels in food, one standalone PAL autosampler can use SPE to clean up enough QuEChERS extracts for continuous, around-the-clock measurement by five GC/MS/MS systems.⁸

Conclusion

A novel micro-SPE device offers highly effective automation using only existing analytical instrumentation while simultaneously improving SPE performance.

References

1. Gamble, K. U.S. Patent 6969615; 11/1/2005.
2. Gamble, K. and Martin, W. EU Patent 1174701; 09/11/2007.
3. Gamble, K. CDN Patent 2316648; 7/4/2004.
4. Gamble, K. and Martin, W. U.S. Patent 7001774; 2/21/2006.
5. Gamble, K., Fitzgerald, R. U.S. Patent 7798021, 9/21/2010.
6. Neue, U.D. HPLC Columns: Theory, Technology, and Practice; Wiley VCH: Weinheim, 1997, p 13.
7. Jordan, L. Automating a solid phase extraction method. LC·GC 1993, 11, 634–8.
8. Lehotay, S.J.; Han, L. et al. Automated mini-column solid-phase extraction cleanup for high-throughput analysis of chemical contaminants in foods by low-pressure gas chromatography—tandem mass spectrometry. Chromatographia  2016, 79, 1113–30.

Mark Hayward, Rick Youngblood and Kim Gamble are with ITSP Solutions Inc. (ITSP), 10 South Carolina St., Hartwell, Ga. 30643, U.S.A.; tel.: 706-395-8300; e-mail: [email protected]; www.itspsolutions.com. Jonathan Ho and Tom Moran are with Shimadzu Scientific Instruments, Somerset, N.J., U.S.A. Matthew T. Hardison is with Assurance Scientific Laboratories, Bessemer, Ala., U.S.A.