Solid Choices for Chemical, Biochemical Analysis
Cost of ownership has always been a factor in purchasing LC systems, says Bill Foley, senior director, separations product management, at Waters (Milford, MA). But today, as major LC markets become increasingly cost-conscious, lab managers can justify more expensive UHPLC based on value. “More than ever, UHPLC systems are easier to justify from the perspective of cost of ownership.”
Perhaps the most tangible, measurable benefit to UHPLC is reduced solvent usage. Labs pay for solvents coming and going; HPLC/UHPLC-grade solvents are expensive to acquire and costly to dispose of. Supplies of acetonitrile, a preferred LC solvent, are back up after the shortages of several years ago, but prices have not fallen in step with rising inventories of the solvent.
According to Foley, labs that invest in new UPLC® (a Waters trademark, versus the generic UHPLC) systems can save 90 percent or more on solvents. Savings are even greater for Waters’ supercritical carbon dioxide-based ACQUITY UPC2® systems—what Waters has termed “convergence chromatography.”
UPC2 systems are identical to UPLC in hardware design. While supercritical chromatography systems have a reputation for extreme operating conditions, pressures are around the same, or lower, as they are for UHPLC.
Throughput and sensitivity are common selling points for UHPLC. Run times are significantly shorter, resulting in fivefold or higher raw throughput. Yet realworld improvements vary according to several factors. One is instrument location. Users who in the past could access an HPLC on the next bench may now have to walk down the hall to use or check on the availability of the replacement UHPLC.
Another factor relates to whether the UHPLC replaces HPLCs dedicated to one specific method. This would not normally be problematic when both instruments employ the same column and mobile phase. But sometimes they don’t, and users can be picky about how colleagues maintain the instrument.
Waters therefore takes a conservative position on how many HPLCs a UHPLC can replace, so while run times may be shorter by a factor of five or more, the company’s official stance is that one UHPLC can replace two HPLCs and, under ideal circumstances, possibly three.
Not on pressure, particle size alone
During the early days of HPLC, the “P” in the abbreviation stood for “pressure” but eventually gave way to “performance.” Perhaps this historical fact and UHPLC’s use of very high pressures led to the mistaken belief that UHPLC’s superior performance was solely a result of pressure (and small particle size stationary phases).
Not true. As Jason Weisenseel, PhD, chromatography technical leader for aftermarkets at Perkin Elmer (Orlando, FL), notes, a major factor in UHPLC’s effectiveness lies in the super-low system dead volume, which produces tighter bands. “When they’re tighter they’re taller as well, and sensitivity rises,” Weisenseel observes.
Weisenseel relates that during the 1980s and 1990s, as HPLC systems became more robust, users became complacent about sample preparation. But the advent of Waters’ UPLC in 2004 required users to rethink sample prep. “Operators must be more careful of sample and mobile phase cleanliness than they need to be with HPLC.”
Columns are the component most susceptible to damage through particle buildup. “Injecting an untreated, nasty matrix can clog columns irreversibly. Tubing is susceptible as well,” Weisenseel says.
While some UHPLC workflows can withstand “dilute and shoot,” more complex samples such as plasma and serum do better when filtered through 0.2-micron membranes or subjected to solid-phase extraction (SPE).
Ade Kujore, marketing specialist at Cecil Instruments (Cambridge, UK), provides some perspective on the HPLC-UHPLC decision. “UHPLC involves more stringent and prolonged sample preparation and often more frequent maintenance and servicing.” UHPLC systems are more expensive, and while these systems are faster, labs are not always prepared for processing the data UHPLC generates. “Superficially porous particle columns, together with column heaters/chillers, may negate the need for UHPLC in some instances,” he adds.
Sample and mobile phase preparation do not involve high-level engineering, but they do take time and require judgment, training, and skill levels somewhat higher than for HPLC. So while UHPLC does provide higher throughput and sensitivity, sample prep must be factored into any time and cost analysis.
Mass detection has been viewed as a way to avoid some types of sample preparation while augmenting sensitivity. Phillip DeLand, global LC business manager at Bruker (Fremont, CA), notes the historic banter between chromatography and MS, where chromatographers view MS as just another detector and mass spectroscopists see LC as merely a fancy injection device. But clearly the synergy between the two platforms creates analytical value greater than the sum of its parts.
Bruker does not sell UHPLCs as stand-alone instruments but only integrated with its core MS instruments. Its major analytical markets are environmental, food testing, and life sciences proteomics.
“In those areas, we see a common theme in reducing detection limits within a variety of matrices, where interferences exist or with low-concentration analytes,” DeLand says. The most significant boost to sensitivity occurs in the interface between the LC and the MS, most commonly with the gentle electrospray ionization technique. Bruker has developed an ion source, Captive Spray, which gets more of the analyte into the MS’s ion optics. This builds on previous efforts to increase ionization efficiency through addition of a solvent that enhances ionization by adjusting the charged state of the analyte, which in turn allows more efficient charge transfer during the electrospray operation.
HPLC automation is mostly limited to pre-analysis, which falls within the realm of general lab operations. Yet the speed and throughput of modern LC systems have shifted the workflow bottlenecks from the chromatography run to sample preparation, to the point where labs contemplating high-throughput operation need to study automation possibilities.
One consideration involves vial or microplate handling just prior to auto injection. Conventional plate stackers can store dozens of plates at the proper temperature, feed them to the injection component, and return them to storage. Waters’ Sample Organizer product, for example, stores and delivers up to 20 plates.
Very high throughput labs, or those that live and die by consistency of results, must consider a more comprehensive automation of sample preparation that includes dispensing or liquid handling, temperature control, filtration, dilution, solid phase extraction, and sample container shaking or washing. One could even imagine workflows where microplate reading might antecede injection into an LC.
“An ideal system would clean up the sample throughout the workflow without user intervention,” Foley says. “This frees scientists for other tasks and eliminates human error.”
Regarding automation, Kujore suggests that purchasers look into systems that allow automation on the lab’s terms—when and to the degree that is needed and not as an expensive add-on at the time of purchase.
Reliability of various automation components is a primary concern, according to Kujore. “The last thing you want after leaving for a long weekend is worrying if your robotics are doing what they’re supposed to do.” Users should also factor in acquisition and maintenance costs. “If you find yourself with a negative payback—such as routinely devoting extra time and resources to use, set up, and maintain and automate—you’ll wonder if you actually need it.”
UHPLC systems have significantly improved separations and peak capacities relative to HPLC, but the need exists for even higher performance. Improvements may be possible by applying even higher pressures to even smaller stationary phase particles, but the resulting physical demands on instrumentation—back pressures—increase exponentially as particle sizes shrink.
One elegant work-around, says Dr. Michael Frank, Agilent’s senior director, global marketing, liquid phase separation business (Waldbronn, Germany), is two-dimensional LC (2DLC). The technique involves subjecting each peak in the first dimension to separation in a second, orthogonal dimension, for instance, reverse phase followed by ion exchange. Thus, compounds that coelute by virtue of affinity to C18 will separate based on charge.
Any complex sample or samples with difficult-to-separate components are candidates. Life science separations, which often involve complex samples and/or low-concentration analytes, come immediately to mind, as do clinical and bioprocessing samples. Even small molecule drug assays containing structural isomers, closely related metabolites, or coeluting impurities are often difficult to separate through one-dimensional LC.
2DLC is gaining traction in nontraditional HPLC industries as well. Today, the technique quantifies small differences in beer samples and helps the Chinese herbal medicine industry standardize their formulations. It can also characterize samples from hydrocarbon and polymer processing, where components differ only slightly. “Some of these separations are achievable only through 2DLC,” Frank says. “Having two different columns or two different mobile phases operating during one run tremendously increases your chances of separating compounds of interest.”
2DLC reduces somewhat the need for high-end MS detection, provided peaks are well characterized in both dimensions, for example, samples containing five compounds that separate easily. A single-quad MS detector may suffice for providing confirmatory mass assignment when analyzing a pharmaceutical active ingredient that includes ten to 15 impurities.
For protein digests, however, MS increases the likelihood of obtaining the most information from each injection, Frank says. “2DLC theoretically multiplies the peak capacities of the two individual separations. MS provides additional peak capacity—not in the retention time dimension but in a separated mass dimension. MS/MS adds two massselective peak capacities, and with ion mobility you can increase peak capacity to the power of five, all in one run.”
The advantages of 2DLC come at a cost. As with the migration from HPLC to UHPLC, 2DLC adds complexity to the chromatographic workflow. “There’s no such thing as a free lunch,” Frank comments. With UHPLC the prices are somewhat higher maintenance and more extensive sample preparation. For 2DLC the principal downside is system complexity—the need for an additional pump and column. On the plus side, because 2DLC operates at normal system pressures, the system is under reduced stress relative to UHPLC.
Paradoxically, run times under 2DLC are not disadvantages when all factors are considered. Runs on complex samples may be lengthy under HPLC and even under UHPLC. According to Frank, 2DLC greatly reduces method development times. “Overall, if you have to develop a completely new method [under HPLC/UHPLC] until all components in a sample are identified, 2DLC would decrease overall work times.”
Care and maintenance
Of all the components in an HPLC system, pumps require the most care and maintenance. Kujore advises against allowing gases into the pump, which makes online or off-line degassing of all mobile phases a must. “When a pump has been unused for eight hours or more, before switching it on, check that no air bubbles are visible within the mobile phase tubing. If there are, first purge each pump according to the manufacturer’s instructions.” In addition, when changing the contents of the mobile phase container, operators should prime all pumps before switching them on.
Another way to protect the system is by filtering all mobile phases and components before introducing them to the system. Users should consider an in-line, 0.45-micron mobile phase filter and a smaller-pore filter for samples.
Acetonitrile, which remains decidedly unkind to solvent budgets, is equally problematic for check valve seals, as it tends to cause sticking. Kujore therefore recommends priming with water or methanol after using an acetonitrile mobile phase. Operators should also note other pump-unfriendly solvents, per the user manual, and mobile phase components that are incompatible with normal or reverse phase columns.
Remember that pump back pressure increases as guard columns become fouled. One key to longer analytical column life is a regular change of the guard column. “Frequent pump back pressures of greater than 33 MPa at flow rates of less than 1.4 ml/minute in columns shorter than 25 cm are cause for concern,” Kujore advises.
When buffer salts are present in mobile phases, operators should not allow them to precipitate out of solution, which will cause pump parts to stick. One technique during downtime is to circulate buffer through the pumps at a low flow rate, such as 0.2 ml/minute. “Never let the pump lie unused for more than, say, twenty minutes, if the mobile phases contain buffer salts,” Kujore cautions.
Finally, after using the buffer, flush the system with 15 percent methanol in water for two hours and back-flush pump pistons, either manually or automatically, after using buffers with very high salt content.
Is SFC special?
Despite SFC’s reputation as an exotic form of LC, care and maintenance are amazingly similar to what users might expect of standard LC. The back-pressure regulator incorporated into SFC systems is the only major difference. Additional maintenance relative to HPLC involves checking these back-pressure seals and seats.
“Aside from that additional component, SFC systems are virtually identical to HPLC, in terms of both maintenance and component longevity,” says D.J. Tognarelli, chromatography product specialist at JASCO (Easton, MD).
If anything, column lifetime tends to be longer in SFC because supercritical CO2-based mobile phases, even with the addition of cosolvents, tend to be gentler than typical HPLC solvents. Also absent are problems related to running HPLC with salt buffers that increase the likelihood of clogging or salt precipitation.
Pressures employed in HPLC and SFC are virtually identical, Tognarelli adds. “In our SFC systems, the cosolvent pump is an HPLC pump, so the pressures are not meaningfully different.” The only pressure difference is in the flow cell, which in HPLC operates at low pressure. In SFC the flow cell is maintained at high pressure. “But this does not affect maintenance.”
UHPLC pressures, by contrast, are significantly higher than for SFC or HPLC and the source of some maintenance issues. JASCO SFC systems operate at about 500 bar maximum pressure, whereas UHPLC may go as high as 1,300 bar, which subjects seals and valves to at least twice the stress of SFC or HPLC.
Potential purchasers of LC systems need to do their homework. Those on the fence regarding HPLC or UHPLC should refer to the opening section of this article.
It is not unusual for labs to send one or more group members to one of the larger instrumentation or automation trade shows specifically for purposes of “kicking the tires.”
Generically speaking, purchase decisions should be made under the assumption that the instrument will be in service for around eight years, possibly longer.
That is why Kujore’s list of factors to consider are heavily based on cost of ownership: • Cost and ability to implement accessories and components.
- Ease of routine maintenance, such as the cleaning of pump check valves, mobile phase switchovers, column changes, flow cell changes, lamp changes, and the ability to physically access system components.
- Software ease of use and functionality: Will there be a long, steep learning curve? Can your laboratory afford for staff to take time off from their normal duties to attend training sessions?
- Reliability, continuity, efficiency, and competence of support, whether from the manufacturer or a third-party support organization.
- System reliability, longevity, and the availability of spare parts and optional accessories.
- Cost of authorized repairs and service.
- The cost of special consumables: Can you use ordinary consumables, or are you restricted to those from one specific manufacturer? What happens if that manufacturer for any reason ceases to supply the necessary consumables or will supply them only under a severe increase in pricing and/or terms and conditions?
- Flexibility and choice of analytical columns.
- Detector specifications such as low drift, noise, and stray light.
- Pump specifications such as low pump pulsation and speedy and accurate gradient mixing.
- Autosampler specifications such as carryover, injection precision, numbers of injections, and availability of accessories.
- Column heater/chiller specifications such as compartment size, temperature ranges, speed of temperature changes, and the accuracy and stability of required temperatures.