HPLC 2013: 40th International Symposium on High Performance Liquid Phase Separations and Related Techniques

The prior 39 symposia in the HPLC series span 40 years and global locations in Asia, Europe, and North America, but this was the first in the Southern Hemisphere. The 40th International Symposium on High Performance Liquid Phase Separations and Related Techniques attracted almost 300 chromatographers to the beautiful Grand Chancellor Hotel in Hobart, Tasmania, from November 18 to 21, 2013. The technical program was unusually strong, with exciting reports on developments in column hardware, column packings, LC×LC, Convergence Chromatography (UPCC and SFC [supercritical fluid chromatography])MS detection, and capillary electrophoresis. Reports on enabled applications also covered the globe.

HPLC column technology

The steady evolution of HPLC column technology over 50 years has driven the instruments and enabled new applications. The key features are selectivity, efficiency, and robustness. Despite HPLC’s age (nearly 50), it is clear that LC column technology is far from mature.

Greatly improved column efficiency

Since the early days of HPLC, chromatographers have wondered about the details of flow profile of the mobile phase through columns. At the beginning, models of flow profile were parabolic, with the highest velocity in the center but tailing to a much slower flow at the column wall. Then chromatographers started making measurements. They found evidence of the multipath effect, where the liquid near the wall moved more quickly since the bed structure was less regular. This offered greater permeability than the close-packed structure that characterized the interior of the column. This was confirmed at HPLC 2013 in Hobart, with photomicrographs showing a transition zone of about 10 particles between the irregular jumble adjacent to the column wall and a uniform close-packed structure for the interior.

As particles decreased in diameter to “sub-twos,” frictional (Joule) heating started to become important. Frictional heating is interesting with at least two potential problems for ultrahigh-performance liquid chromatography (UHPLC). In the center of the column, radial dissipation of the heat from frictional heating is hindered by the very low thermal conduction of the column packing. Thus, the heat cannot reach the cool wall and stays in the center, reducing the viscosity of the mobile phase in the column centroid. This accentuates the severity of parabolic shape in fat columns, hence the focus of sub-twos packed in 2.1-mm or narrower columns for UHPLC.

Most chromatographic processes are dominated by enthalpy (delta H). Heating reduces the attraction between the stationary phase and analyte. Thus, in the hotter center, retention is weaker and the analyte moves even faster than in the colder region near the wall. This acts to further increase the severity of the parabolic band profile, which is expected to increase band broadening. This effect should be even stronger with longer columns since more frictional heating would be expected.

Optimizing thermal conductivity

The column chemists at Waters (Milford, MA) responded to the flurry of interest in core-shell particles with introduction of CORTECS™ columns with C18, C18+, and hydrophilic interaction liquid chromatography (HILIC) surface chemistry. The neat feature about CORTECS is that they are also sub-twos, with a particle diameter of 1.6 μm. The solid core is about 1.1 μm. The shell is about 0.24 μm with 90-Å pores. The solid core improves radial conduction of heat to the sidewall, which improves column efficiency, particularly the C term. This enables trebling flow rate with no loss in column performance compared to competitive columns packed with 1.7-μm core-shell particles.

CORTECS technology led to the obvious question: If this works, can the effect be pushed further? Responses were: Diamond has the highest thermal conductivity, with silicon carbide being a close second. Diamond phases are being developed by Diamond Analytics (Orem, UT). The company is marketing column packings involving microdiamonds. What about diamond core shells? Would industrial diamonds with 1 μm diameter be affordable?

Active flow technology

I’m aware of two different approaches to avoiding the flow inhomogeneity in HPLC columns. Prof. Mary Wirth (Purdue University, Lafayette, IN) is using “slip flow” technology with sub-1-μm particles and very narrow (sub-μm) capillaries to achieve astounding column efficiencies. Unfortunately, I did not see any mention of this at HPLC Hobart. What I did see were numerous reports on column fittings that enable active flow technology (AFT), which was developed by a team led by Robert A. Shaliker (University of Western Australia in Sydney) and developed commercially by Thermo Fisher Scientific’s (TFS) HPLC column technology group in Runcorn, U.K.

AFT involves column fittings that partition the flow of mobile phase into the center and perimeter. As described above, the flow along the column perimeter is complex and less uniform. For AFT, column end fittings are constructed with concentric frits feeding related ports that facilitate injection to, and collection from, the column centroid and on to detection. The flow from the bed perimeter flows through the outer frit and outlet fitting to waste.

The results are impressive: Standard-bore-diameter (4.6-mm) HPLC columns can be eluted five times faster than is customary, thus improving productivity. This is especially useful with MS detection, where peak heights are nearly doubled and signal/noise ratio is improved 10 times.

AFT fittings are usually most effective at the column terminus, but are also beneficial at the inlet. At the inlet, one can explore whether the sample should be injected uniformly or only in the center. With the latter, I’m reminded of the Infinite Diameter Effect discovered by Prof. John Knox of the University of Edinburgh (Scotland) in the late 1970s. Infinite diameter columns showed higher efficiency, provided the column was short enough that radial dispersion of the peak in the column was not sufficient for the sample to reach the wall region.

At HPLC Hobart, one poster compared a narrow-diameter (2.1-mm) AFT column, which had a central collector of 1.07 mm with a 1.0-mm-i.d. column. The AFT column showed a 58% improvement in efficiency compared to a conventional 1.0-mm-i.d. column. The improvement is attributed to avoiding the wall effect, which limits performance in a conventional 1.0-mm-i.d. column. Plus, the plate height vs linear velocity plot is much flatter for the AFT column, which means that there is little penalty for using high flow rate to reduce run time.1

Interestingly, AFT column technology can also be used for reaction flow chromatography (RFC) for postcolumn detection (PCD).2 This uses a postcolumn derivatization reaction to improve chromatographic detection. RFC requires a second HPLC pump delivering the derivatization reagent connected to the fitting at the column effluent. Column effluent and reagent are mixed together to produce the signal. If the reaction is fast, the low dead volume (less than 1 μL) reduces band broadening compared to a conventional fitting T followed by reaction tubes (more than 100 μL). The utility was demonstrated for improved detection of phenols via reaction with 4-aminoantipyrine and potassium ferricyanide. This improved the S/N for phenol detection by about 100× compared to conventional UV absorbance detection.

The AFT development may have much more to offer. Dr. Tony Edge of TFS’s Runcorn group presented a collage of slides showing still greater, unanticipated, and unexplained improvement in column performance. Again, observation leads theory. Stay tuned.

Chromatographic selectivity

Designing selective stationary phases

Dr. Chris Pohl of TFS’s Dionex group has been unusually successful in designing stationary phases for particular applications. Just look at the series of ion chromatography columns that are marketed under the Dionex brand. At the meeting, he introduced four new columns, all employing novel mixed-mode stationary phases, and two employing nanoparticles. Each is designed to solve a particular applications problem.

The first treated existing Dionex ion chromatography columns with glycidol using electrostatic grafting of a sulfonated cation exchange resin. The first glycidol layer covers the polymer surface. Amine treatment gives low-capacity quaternary amine anion exchange behavior. Repeated cycles of epoxide and amine decrease retention of divalent anions (carbonate, sulfate, etc.) while retention of monovalents is unchanged. Thus, one can tune the divalents into a nonbusy section of the chromatogram, improving peaks per minute.

The same process also works on anion exchanges such as AS-19 and AS-15. An interesting advantage is that glycidol chemistry gives a very robust stationary phase since sites for Hofmann elimination reactions involving beta-hydroxyls are avoided.

Mixed-mode stationary phases are designed to provide separations via two or more controllable retention modes. Anion exchange and reversed phase is one example; a cation exchange, anion exchange, and RP is a trifunctional mixed-mode phase. This column packing, called Acclaim™ Trinity™ P1, bonds C10 RPLC ligands and weak anion exchange (WAX) ligands to 3-μm spherical silica (100-Å pore). A weak cation exchange nanoparticle is held electrostatically to the bonded phase. Trinity P1 shows useful selectivity for assay of 10 pharmaceutical counter-ions including tromethamine, mesylate, bromide, and chloride.

Chris’s third introduction was the Acclaim Trinity P2, which starts with covalent bonding of weak cation exchange and HILIC groups followed by electrostatic attachment of nanopolymer strong anion exchange beads. This column provides separation of pharmaceutical anions and cations in the same run. One chromatogram shows separation of K+ and penicillin G (cation and anion) in a single 4-min run. Another showed metformin chloride in 6 min.

Glycans liberated from proteins are a major application in proteomics. Most often, they are separated by HILIC with MS detection. Chromatograms are a complex mixture with no obvious pattern. The Dionex GlycanPac™ AXH-1 is a HILIC/WAX mixed phase that provides classwise separation of mono-, di-, tri-, quat-, and penta-silated glycans in 30 min. Any one of these columns would merit its own lecture.

Ultra Selective Liquid Chromatography

Especially in LC, selectivity dominates the separation space. A guide to Ultra Selective Liquid Chromatography (USLC®) from Restek (Bellefonte, PA) helps chromatographers make good choices in column selection. One starts with a quick examination of the structure of analytes into hydrophobic, dipolar, acidic, and basic. These properties are matched to stationary phase profiles. Mobile phase selection is next. The company recommends scouting with mixtures of four—aqueous (A) 0.1% formic acid in water, or 0.1% formic acid mixed with 5 mM ammonium formate. The B solvents are neat methanol (protolitic) or acetonitrile (aprotic) in steep gradient elution runs from A to B. Use of formic acid for pH control improves compatibility with mass spectrometers. Restek offers several reversed-phase liquid chromatography (RPLC) columns with indexed selectivity including C8, C18, diphenyl, embedded polar alkyl, and perfluorophenyl.

Column screening

Phenomenex (Torrance, CA) has a significantly different approach, starting with selecting the solid support for RPLC. Choices include traditional porous particles, silica monoliths, core-shell, and organo-silica hybrids; then the surface chemistry is selected. Suffice to say that there are many options. The company offers a screening application (ColumnMatch.com™) that searches its huge database for relevant separations. These are distilled down into predicted column and run conditions.

Monolithic, UHPLC, and HILIC stationary phases

For at least two decades, Merck KGaA (Darmstadt, Germany) has been promoting columns where the stationary phase is a monolithic silica rod. Recently, Merck chemists introduced Chromolith® 2-mm-i.d. columns with the macro pore at 1.5 μm diameter. These provide column efficiencies of 100,000 N/m. The new columns have two main advantages: 1) The pressure drop is much less than the corresponding packed-bed column with similar efficiency; thus one can use higher flow rates to reduce run time. 2) Monolithic columns are less prone to column plugging and are generally more robust than packed-bed columns.

Merck also introduced Purospher® STAR RP-18 UHPLC columns. Conventional Purospher columns utilize 5-μm-diam spherical silica, but the STAR line reduces the packing diameter to 2 μm. This facilitates a tenfold increase in sample throughput while reducing mobile phase per run by 89%.

SeQuant® ZIC®-cHILIC is another interesting stationary phase introduction from Merck Millipore. The surface chemistry is phosphocholine bonded to 3- or 5-μm spherical silica. Choline’s imbedded phosphate group provides a fixed negative charge connected to a terminal quaternary amine with a + charge. ZIC-cHILIC gives rapid separation of polar analytes under HILIC conditions, or useful selectivity for zwitterions under RPLC conditions.

Orthogonal columns

Reversed-phase chromatography, with C18, C8, and even C30 surface chemistry, is remarkably effective in providing chromatographic separations. However, coelutions are still a problem, especially with more than about 10 analytes/sample. This is the case when attempting the assay of nitroaromatic analytes using U.S. EPA Method 8330 A/B. A poster by Udo Huber and colleagues at Agilent (Waldbronn, Germany) utilized two chromatographic runs for the separation.3 The first used a C18 column as specified in Method 8330. The second run was with a phenyl-hexyl column. It appears that the π–π interactions between the nitroaromatics and phenyl group on the stationary phase are successful in resolving ambiguities arising from the C18 phase alone.

SEC columns for protein aggregates

Protein aggregates are a difficult analytical problem in biotherapeutics. Some, such as IgMs, are pentameric in their active form. Agilent has developed Bio SEC (steric exclusion chromatography) columns packed with 5-μm-diam spherical silica with nominal pore size ranging from 100 to 2000 Å in six increments. The latter has an operating range up to about 40 megadaltons for proteins. Run times are about 10 min. Although some aggregates are concentration sensitive and kinetically labile, these columns are an interesting advance. Chemists involved in formulation and stability studies should find these columns particularly useful.

HPLC detection

MS detector

In early October 2013, I received a phone alert regarding a new product introduction from Waters the next morning. Well, I’m glad I tuned in, since Waters introduced the ACQUITY QDa Detector as an extension of its ACQUITY UPLC franchise. The key question is: What if MS detection was suddenly as familiar as optical detection? The QDa is a fresh design from the beginning based on 30 years of MS development experience. The design is protected by 37 patents covering size, utility, and price. It works with Waters’ UPLC, UPC2, and Alliance HPLC systems and is the same size as an optical detector. As straightforward to operate as a common UV detector, the information-rich nature of MS is useful in qualitative and quantitative studies. If the performance and ease of use live up to the advance billing, MS detection may replace optical absorbance as the most commonly used.

Improved ion chromatography suppressor

Suppressors in ion chromatography are used to remove the conductivity of the background eluent and thus improve analyte detection. In the early days, suppressors were packed beds, which with time were replaced by membrane designs that facilitated continuous operation. The next step was to include electronic regeneration, which further simplified operation. Membranes were fast, but not very robust. Thermo Fisher Scientific (Sunnyvale, CA) announced the Dionex ERS 500 electrolytic suppressor designed around a planar array of beads, which improves temperature and pressure tolerance and minimizes peak dispersion. Specifically, the thin packed bed is compatible with higher pressure generated by Dionex 4-μm packed IC columns as well as MS, ICP, and ICP-MS detectors. The 500 replaces the popular Dionex SRS™ 300 Self Regenerating Suppressor.

Convergence chromatography

SFC method development

On the analytical scale, a poster by Paula Hong and colleagues at Waters described method development with ACQUITY UPC2  with columns packed with sub-2-μm particles.4 With the current state-of-the-art, the first step is to make two scouting runs with UPC2 BEH and UPC2 BEH 2-EP, where EP is ethylpyridine. If peak shapes are poor, it is recommended to try pH control additives. If retention and/or selectivity are not sufficient, additional runs with UPC2 HSS C18 and CSH FP (controlled surface hybrid fully porous) are recommended. Once the column has been chosen based on best selectivity, experimental variables such as column pressure, temperature, and gradient profile are optimized.

Process SFC

Three years ago, Waters boldly predicted that many traditional HPLC separations would be replaced by Convergence Chromatography. Poster 146 by Martin Enmark and colleagues of Karlstad University, Sweden, described measuring adsorption isotherms for industrial-scale SFC.5 The primary motivator is to take advantage of the improved throughput and reduced environmental footprint. The authors pointed out that SFC with CO2 is more complex, because it requires more sophisticated techniques for isotherm measurement than simple frontal analysis. They reported that elution characteristic points, retention time, and inverse methods gave useful prediction of elution profiles.

LC×LC

LC×LC is also referred to as comprehensive LC and 2DLC. The goal is to improve peak capacity without simply using longer columns and corresponding even lengthier run times.

In 2-D separations, the ideal case is to have orthogonal separation modes operate first on one dimension and then the second. The second dimension spreads the results of the first dimension out, providing a peak capacity of the product of the two separations. So, if the first dimension has a peak capacity of 30 peaks, and the second has 10, the theoretical product is 300.

LC×LC is following in the footsteps of GC×GC and 2-D electrophoresis, both of which are quite successful in improving the resolution of complex samples, with peak capacities extending into the thousands. But LC×LC is not nearly as successful. Several authors presented interesting separations, but the second dimension provided an increase in actual peak capacity of 30% of theoretical. I discussed this with several delegates.

The key difference is that in GC×GC, one thermally traps the effluent from the first column in a tight band, which provides a concentrated sample band for release into the second dimension. This trapping is not generally practical in LC. Later eluting bands are smeared out, and thus the concentration of analytes in the second dimension is usually below the detection limit of the detector.

Interesting applications

Lipoproteins interact with glycoaminoglycans in several atherosclerotic diseases. Normally the reactions are ionic, as with the binding of anionic sulfated glucosamine with cationic apolipoproteins. Since both are ionic, high-performance capillary electrophoresis (HPCE) should be a suitable assay technology choice. A team in the laboratory of Prof. Marja-Liisa Reikkola (University of Helsinki, Finland) used partial filling-affinity capillary electrophoresis (PF-ACE) to characterize the interactions between the ligand and analyte. First, the capillary is partially filled with ligand and probed with increasing volume of analyte aliquots. Second, the series is repeated with a constant volume but increasing concentration of analyte. The protocol was used to measure the affinity constants (Ka) for lipoproteins (HDL and apoA-1) and phospholipid transfer protein. The measured log Ka’s are 7.1 and 6.6, respectively. Since the run time is only 5 min, a Ka measurement, which requires about six consecutive runs, can be completed in about an hour.

Herbal medicines

Australia’s Department of Health has an active program monitoring herbal medicines for potency, contamination, and adulteration, including substitution. UHPLC  with diode array detection is most commonly used along with thin layer chromatography (TLC). However, Poster 23 described using LC/MS for fingerprinting of herbals or when improved detection sensitivity is required.6 Applications snippets included assay of aristolochic acids, which are banned in Australia due to their inducing of renal failure; misidentification of colchicine in Gingko biloba; and assay of alkaloids in therapeutic products. Products with greater than 0.2 mg alkaloids/dose require prescriptions.

High-performance capillary electrophoresis

Historically, the HPLC community nurtured HPCE, since the early instruments share many design features, particularly detectors. This association continues today, with HPLC Hobart running several sessions on electrodriven separations. eDAQ (Colorado Springs, CO) featured its contactless conductivity detector electronics built into the column module for the Agilent 7100. The conductivity cell is located just ahead of the view port used by the optical detectors. Close coupling minimizes the time offset when multiple detectors are used.

Interesting applications of HPCE

Capillary electrophoresis is ideally suited for laboratory settings, but applications in process control are rare. Poster 294, from a team at the VIT Technical Research Center in Espoo, Finland, described online CE for monitoring carboxylic acids in an aerobic bioreactor.7 The key development was designing a flow-through sample vial that interfaced with a Beckman P/ACE™ MDQ with PDA detection (Beckman Coulter, Brea, CA). A cross-flow filter (0.45 μm porosity) was placed ahead of the sample vial. The system is used to monitor production of food-grade carboxylic acids, proteins, and starch, plus degradation of lignocellulose for biofuels. Close monitoring of the production process facilitates harvesting of higher-purity product.

Summary

At the close of the meeting, I thought about the future impact of the proceedings. UPCC including SFC appears to be developing per the vision set forth by Waters three years ago. AFT column technology is a promising new paradigm with near-universal impact. Second-generation core-shell columns (CORTECS) involving smart design will provide huge advances in speed if the instruments can keep up. Rational design for column selectivity should facilitate development of applications-specific analyzers. Waters made another audacious prediction that the QDa would lead to MS replacing UV absorbance as the most common detector for LC in 10 years. If this prediction came from another source, I’d be more skeptical, but Waters has a track record of making bold steps and has the market position to make the predictions come true. I see encouraging signs from early adopters.

On the other side, unless trapping technology for LC×LC is developed, it will be another example of the adage “what is important in GC is not useful in LC, and vice versa.”

Credits

Profs. Paul Haddad and Emily Hilder of the University of Tasmania deserve special thanks for organizing the technical program with the help of the Scientific Committee and Organizing Committee. Thanks also to Ms. Alexis Maill for organizing the logistical and social program. All in all, an excellent show! The next symposium in the HPLC series is scheduled for May 11–15, 2014, in New Orleans, LA. See you there.

References

  1. Soliven, A.; Foley, D. et al. Improving the Performance of Narrow-Bore HPLC Columns Using Active Flow Technology; Poster 195; HPLC 2013, Nov 18–21, Hobart, Tasmania, Australia.
  2. Soliven, A.; Selim, M. et al. Selective Detection Using Reaction Flow Chromatography; Poster 288; HPLC 2013, Nov 18–21, Hobart, Tasmania, Australia.
  3. Gratzfeld-Huesgen, A.; Huber, U. Optimizing the Separation of 20 Nitro-Aromatics Using Consecutively a Phenyl-Hexyl Column with π-π Interaction and a C-18 Column on the Agilent 1290 Infinity Quaternary Method Development Solution; Poster 152; HPLC 2013, Nov 18–21, Hobart, Tasmania, Australia.
  4. Hong, P.; Durieux, I. et al. Achiral Chromatographic Method Development Strategies with Carbon Dioxide Mobile Phases; Waters, Milford, MA; http://www.waters.com/waters/library.htm?locale=en_US&lid=134743667.
  5. Fornstedt, T.; Enmark, M. et al. Adsorption Isotherm Determination in Supercritical Fluid Chromatography—Pitfalls and Possibilities; Poster 15; HPLC 2013, Nov 18–21, Hobart, Tasmania, Australia.
  6. Fagan, P.; Weerasuria, M. et al. Herbal Medicines—Identification and Analysis; Poster 238; HPLC 2013, Nov 18–21, Hobart, Tasmania, Australia.
  7. Parvathy V.; Lee, R. Stability Indicating RP-HPLC Method for Cyclosporine Eye Drops; Poster 294; HPLC 2013, Nov 18–21, Hobart, Tasmania, Australia.

Robert L. Stevenson, Ph.D., is Editor, American Laboratory/Labcompare; e-mail: rlsteven@yahoo.com .

Comments