LC Column Technology: The State of the Art

In this extended special feature to celebrate the 30th anniversary edition of LCGC Europe, leading figures from the separation science community explore contemporary trends in separation science and identify possible future developments. We asked key opinion leaders in the field to discuss the current state of the art in liquid chromatography column technology.
Compiled byAlasdair Matheson, Editor-in-Chief, LCGC Europe

Gert Desmet: Are there limits to the decrease in size of porous and superficially porous particles?

Tivadar Farkas: Our experience is that there is a clear lower practical limit to particle size imposed by current high performance liquid chromatography (HPLC) instrument limitations. During the development of our 1.3-µm core–shell particles we explored particles down to 1 µm in size. At the time, we were not able to operate columns made with such particles close to their optimum flow rate because of instrumentation limitations in maximum flow and operating pressure.

Professor Wirth’s work with sub-1-µm particles seems to contradict our findings. Still, to our knowledge her results have not been confirmed by an independent report.

Fabrice Gritti: From a technological synthesis viewpoint, 1-µm size-controlled mesoporous silica particles can be routinely prepared and 0.5-µm superficially porous particles can also be prepared from 20-nm to 30-nm monodispersed nonporous core silica nanoparticles. 

However, currently, there are four main limitations to the application of such fine particles for chromatographic separations. First, they are not well suited for the analyses of small molecules using narrow-bore columns (2-mm internal diameter (i.d.) to 3-mm i.d.) because extremely high pressures (3–5 kbar) are required to achieve optimum performance. The current pumps cannot deliver the optimal flow rate of about 1 mL/min at 3–5 kbar. Second, performance losses as a result of extra-column dispersion would be massive with the current instrumentation: A complete integration from sample injection to solvent preheating column and to ultraviolet (UV) or mass spectrometry (MS) analyte detection is still needed. Third, the achievement of reproducibly well‑packed columns (2 mm i.d. to 3 mm i.d.) with sub-1‑µm particles has yet to be demonstrated in the laboratory. Finally, performance loss is expected because of very intense frictional heating.

Despite all of the above, sub-1-µm particles for fully-, superficially-, or nonporous particles can be advantageous but in a capillary format-that is, micro- and nano-LC-for the resolution of large biomolecules, such as proteins and monoclonal antibodies, which require low resistance to mass transfer between liquid and stationary phase and low optimum linear velocities to achieve maximum performance at the usual pressure drops. 

Xiaoli Wang: From a particle synthesis perspective, no. However, the use of even smaller particles, for example, less than 1 µm, is limited by several factors. First, the required pressure would be much higher and the engineering efforts to ensure robust instrument operation dramatically increases. Second, the much higher efficiency requires another order of magnitude decrease in system dispersion. Third, we need column hardware with much smaller dispersion and also new frits for the nanoscale particles. 

Frank Steiner: The big challenge will be mechanical stability under extreme pressures and the required ultrahigh‑pressure liquid chromatography (UHPLC) instrumentation. Decreasing particle size from 1.5 µm to 0.5 µm, and at the same time reducing the column length by a factor of three implies a back pressure increase by a factor of nine to generate equivalent plates in a ninefold method speed-up. This requires operating pressures of several thousand bars. Operating under such conditions will also generate substantial frictional heating with adverse effects on efficiency. All these challenges will make it very difficult to go below a 1-µm particle size.

Mel Euerby: The stationary phase manufacturers’ ability to reproducibly manufacture even smaller porous or superficially porous particles and variations in particle size distributions could have big implications on the resultant pressure and the stability of the packed bed. The second limitation is instrumentation hardware that is capable of reproducibly handling these high pressures. Another consideration is the necessity to further reduce the system volume to prevent excess peak dispersion as a result of the very small peak volumes that these small particles theoretically will generate.

Dave Bell: In present column formats, yes. Smaller particles in the form of fully porous particles or superficially porous particles may, however, be useful in alternative formats, such as open tubular or pillar-array formats. 

Gert Desmet: Given that the speed of analysis is, ultimately, determined by the speed of detection, should future column developments aim to increase speed or increase efficiency and resolution?

Tivadar Farkas: Speed of analysis in practical terms may not be set by the speed of detection. We see it as a result of a combination of contributors. It is easy to see this point by considering that autosamplers still take 1–2 min to do their “dance”. The current practice of using smaller particles, either fully- or superficially porous, resulted in higher speed (given that optimal flow rates are higher) and higher efficiency, which, in turn, often results in higher resolution. This means that state-of-the-art HPLC columns provide improvements in all these performance criteria, so it seems that they cannot be pursued independently. 

Fabrice Gritti: High-throughput liquid chromatography and analysis speed are limited by the duty cycle times of the sample manager (from sample drawing to injection start), chromatographic separation (column length), detectors (UV–vis sampling rates, MS or MSn scan rates), column equilibration, and the data processing (informatics, software). Therefore, further gains in separation speed implies the combined improvement of all these factors. As the demand for high-throughput analyses remains very strong in the biological and pharmaceutical industries and the primary quality required is not performance but speed, there is still a strong interest in developing faster and better‑integrated high-throughput systems. 

This research development in high-throughput chromatography is not incompatible with imagining and designing more efficient columns in high-resolution liquid chromatography for solving very complex sample mixtures. This involves new silica–polymer–hybrid monoliths prepared in-situ in long capillary columns, well-packed capillary columns with fine particles (fully-, superficially-, or nonporous particles), and micro-fabrication of ordered structures by three-dimensional (3D) printing photopolymerization or stereolithography. 

Comprehensive two-dimensional LC (LC×LC) is also now reaching maturity and will play a growing role in the coming years in high-resolution liquid chromatography. Likely, the development of 1D-LC ultrahigh efficient column technology should then likely attract lesser attention in the next decade because conventional comprehensive 2D-LC is much faster and more efficient than any modern 1D-LC techniques. Also, the concept of 3D-LC (LC×LC×LC) is currently emerging thanks to the development of 3D printing technologies, but it might take over a decade from now before the first successful 3D-LC instruments become a reality. 

Xiaoli Wang: I think this should be more application driven. There is always a need to increase speed, for example, in applications where samples are of reasonable complexity or samples are well characterized, or for the second dimension of 2D-LC. Columns that enable faster speed are always welcome and I believe detection limits can be overcome by smart engineers. On the other hand, I see more demand for higher resolution in bioanalysis applications. The question there is do we use more efficient one-dimensional columns, or do we go for multidimensional LC?

Frank Steiner: Modern optical detectors can operate at data rates of 200–300 Hz, which allows working down to a 10 ms peak base in time. However, detecting such fast transients in an unbiased way will require low time constants of the electronic filters, which increases baseline noise and will negatively affect trace-level detection. Modern MS detection in scan mode, however, is more limited in speed and can hardly cope with sub-second peak widths. Regardless of detector limitations, I see more need for peak capacity and resolution improvements to respond to increasingly complex analytical demands in systems biology research.

Mel Euerby: Most UV detectors are capable of fast detection rates of greater than 100 Hz at the moment, which may be sufficient for the next generation of particle sizes. However, this is not the case for other types of detectors. The pharmaceutical industry seeks a balance between productivity, that is, speed, and the quality of data in terms of optimizing the resolution equation parameters, such as efficiency, selectivity, and retention. It is my opinion that we hopefully will see column developments that will exploit more fully the selectivity term and generate more complementary chromatographic selectivity. 

Dave Bell: The speed of a given analysis is dependent on the efficiency and adequate resolution of the analytes of interest. The continued focus on developing column and instrument efficiency is critical to improving the speed of analysis. Additionally, an improvement in selectivity, that is, surface chemistry, often provides the resolution required to induce fast separations. The right tool for the job is often overlooked when developing methods.

Gert Desmet: Do we need to change the way we install columns in our instruments or is the current approach sufficiently user-friendly and robust? And is there any need for automation? 

Tivadar Farkas: Column installation is becoming a challenging task that may compromise chromatographic performance. Luckily, well-designed adaptors and connectors make the task less flawed. Automation could be helpful, but is perhaps “overkill”. Assuming that a typical practitioner installs between zero to four columns a day, more effective connectors seem to be the way to go rather than typically expensive automation.

Fabrice Gritti: Instrumentation is still lacking robustness and simplicity of use today. Remarkably, the overall instrument architecture has not significantly evolved over the last 30 years of research and development of new modern systems. Pumps, injectors, column ovens, and detectors (optical or mass) are still interconnected in complex and often nonoptimized ways. Automation and reduced troubleshooting are essentially what nonexpert users are looking for in their daily applications. There is a definite need to improve size or system architecture, automation, usability, and the troubleshooting procedures of current instruments.

Xiaoli Wang: Yes! There are connections commercially available that are robust and reasonably user‑friendly. However, it has come to a point where further miniaturization, for example particle size or column dimension, is limited by the dispersion created by the numerous connections in the system. Ideally, automated connection that requires little to no manual operation from the users and ensures consistent operation at ultrahigh pressures would be really good for today’s users. Even if full automation is not possible, new connections that reduce the number of steps, are easy to replace, and are compatible with a wide range of columns and instruments would make a chromatographer’s life easier.

Frank Steiner: I would say we are constantly improving this already. There has been a big leap in robustness and user‑friendliness over the years from steel ferrule-based plumbing to the modern fingertight UHPLC fittings that are currently available. There is also powerful column change automation available with low dead volume multiposition valves. We may dream of super-simple click systems for plumbing, even to enable fast robotic column changes. Sample throughput is still limited by many other factors in real laboratory workflows, so in my opinion we do not need to switch columns faster and more frequently just yet.

Mel Euerby: The transition from conventional HPLC to UHPLC highlighted the necessity to properly install the column into the UHPLC system to obtain the claimed efficiencies. This was partially an educational issue. Operators were incorrectly positioning the tubing and end fittings into the column and, consequently, introduced extra dead volume, which fortunately is not critical for the peak volumes associated with conventional HPLC column formats. It can, however, significantly reduce the efficiency obtained in UHPLC as a result of the small peak volumes.

Dave Bell: One can certainly foresee alternative column formats that could lead to more efficient and dependable installations. This is especially true for chip-based devices. The column format has been largely left unaltered for many decades, especially for narrow-bore to standard‑bore formats. Modern 3D printing may help to catalyze the development of novel connections and column formats that could lead to higher efficiency, and more robust, friendlier installation. 

Gert Desmet: Do you think sufficient attention has been paid to the column hardware, given the improvements in particle quality?

Tivadar Farkas: Column hardware has been the focus of HPLC column developers and manufacturers. It has been dissected, reevaluated, and reengineered many times. Perhaps not all changes (some of them internal and miniature) are evident to chromatographers. Still, more remains to be done in terms of improving hardware performance while keeping costs manageable. 

Fabrice Gritti: As far as column efficiency is concerned, the stainless steel open tube in which particles are packed is inevitably limiting column performance because it is directly responsible for the heterogeneity of the packed bed density across the column diameter following optimized slurry packing procedures. It is known as the wall effect. Columns packed with either fully (minimum reduced plate height h_min = 2) or superficially (h_min =‑1.4) porous particles are not delivering the maximum expected performance given their bulk random packing: The performance of the corresponding infinite diameter column is h_min = 0.7–0.9. The solution to that problem has been proposed with the development of curtain flow and segmented flow chromatography, which both eliminate the nefarious contribution of the wall region to the overall band broadening process. However, this solution is limited to small molecules and the new column hardware is anything but an easy-to-use tool. Additionally, as far as high-throughput liquid chromatography with short columns is concerned, little attention has been given to the problem of efficiency as a result of the border effect related to the presence of porous metallic frits. Insufficient attention has been given to the redesign of the whole column hardware in order to approach the existing performance of the bulk random packings in both high-resolution liquid chromatography and high-throughput chromatography.

Xiaoli Wang: Probably not, but this is normally a highly sensitive area where efforts and results are not shared openly in the community. Column hardware quality can have a huge impact on column performance. Improvement in column hardware should certainly be made and I think this will continue to be driven by column vendors. For example, some column dimensions, such as narrow bore, that we consider difficult to use today might become as good as traditional dimensions with improvements in column hardware.

Frank Steiner: I would actually challenge this a bit. While plumbing systems have seen true innovation and progress, I am still missing this with stationary phase retainer technology. Removing the adverse effects of frits and sieves on column efficiencies with well packed, highly efficient particles could in my opinion not keep pace with the improvement of the particles. When it comes to column internal diameters below 2 mm, the overall quality of the column hardware, the internal surface smoothness, and the accuracy of diameters also have room for improvement. I acknowledge the difficulties with all that, but think it should get more attention.

Mel Euerby: Frits in all their types and shapes are a major concern because their surface area is very large and contaminants can be easily transferred onto the highly pure stationary phase in the column. We have previously had major problems with metal contamination of columns from frits when exposed to high organic mobile phase conditions.

Dave Bell: No. Developments in column hardware and sample–mobile phase transfer may be one of the next major breakthroughs in liquid chromatography technology. Modern 3D printing may help to evolve the development of novel connections and column formats that could lead to higher efficiency and improved installation.

Gert Desmet: What new stationary-phase chemistries are needed?

Tivadar Farkas: Octadecyl (C18)! It is amazing how much effort went into the development of an endless number of C18 reversed phase columns over several decades of HPLC history. While many of them are only incrementally different, we are still pursuing the dream column that could do it all. While C18 columns fall short of this dream, perhaps they are getting closer to it? 

It is hard to predict what speciality surface chemistry is going to be uniquely suitable for a particularly difficult application. At the same time, I am sceptical about the prospects of developing a novel surface chemistry that could come close to the success of the octadecyl surface ligand.

Fabrice Gritti: Improved hydrophilic interaction liquid chromatography (HILIC) chemistry phases are yet to be developed for higher retentivity and selectivity of polar and highly polar compounds, such as sugars and polyalcohols. Hybrid organic–inorganic superficially porous particles should be improved to increase their lifetime under harsh eluent compositions, such as basic eluent pHs. Finally, the surface chemistry of particles can sometimes be very active towards large biomolecules. This makes their identification difficult with standard polymers in size‑exclusion chromatography (SEC) and lowers the level of method sensitivity in adsorption chromatography because of severe peak tailing. Developing a more inert surface for packing materials is the key towards successful polymer–biomolecules separation and identification. Yet, each class of analyte has its own reactivity and the ideal particle is likely an unreachable target. Instead, efforts are often directed to the selection of the proper eluent composition. 

Xiaoli Wang: New phase chemistries that can address special applications are always needed, but the application size is becoming smaller. In general, I believe more MS-friendly phases are needed, for example, phases with better stability with lower MS bleed. We also need phases that require less salt–buffer to obtain good peak shape and have fast reequilibration. Specifically, I see more need for new phases in HILIC and supercritical fluid chromatography (SFC), where no C18-equivalent phase exists. I also see a need for new phases with less nonspecific binding for biomolecules.

Frank Steiner: I cannot easily define what the chemistries should be, but I would like to see better phases for highly generic methods. Reversed-phase phases with some intrinsic secondary interactions in a long range gradient is not enough for this, in particular not when it comes to highly polar analytes. I do not think that HILIC has truly fulfilled all expectations. Intelligent and highly efficient multimode phases that call orthogonal retention mechanisms in sequential gradients may be a promising approach.

I also think we need a truly efficient alternative to SEC for several very popular characterization workflows of biotherapeutics, which all suffer from the slow diffusion of macromolecules in too-narrow pores. 

Mel Euerby: This is the million dollar question and the answer will be different for different chromatographers, especially those working in niche areas. From a pharmaceutical perspective, I would expect to see more high pH-stable phases with differing chromatographic selectivity to the ubiquitously used C18 phases that are in the high pH market at the moment. The transition from small drug molecules to biomolecules will result in new ranges of phases specifically designed for them, possibly on a nonsilica backbone. In addition, there is a growing trend for solution-specific columns to meet particular needs. Above all, new phases will have to be shown to be reproducible from batch to batch.

Dave Bell: I often hear the argument that more stationary phase chemistry options are not needed and that there are too many options already. The plethora of column choices is said to confuse the end user. Alternative chemistries, however, provide different balances of interactions that can make or break a given separation. Since no stationary phase can provide the selectivity required for every separation need, I do believe there is a need for more selection. What is not needed are additional surface modifications that provide the same or very similar selectivity compared to existing phases.

Gert Desmet:Is there any commercial future for monolithic columns, produced using either the current sol-gel or polymerization routes or by new technologies, such as micromachining or 3D printing? Why or why not? What will their impact be?

Tivadar Farkas: Monolithic and micromachined columns are the dream of most of us preoccupied with the effective packing of sorbents into column hardware. Unfortunately, the present technological challenges and the financial aspects involved with both monolithic and micromachined columns are major obstacles. Once overcome, such columns could provide much improved permeability and resilience to fouling, and, ultimately, easier operation and improved lifetime.

Fabrice Gritti: Silica monolithic columns prepared by the classical sol-gel process are reaching their limit in terms of domain size and pressure tolerance. It is also difficult to prepare uniform monolithic structures across wide column formats from 3.0-mm- i.d. to 4.6‑mm i.d. Overall, they can hardly compete with 1.7-µm- to 2.5-µm-particulate columns, which are increasingly efficient, reproducible, and more robust over time at pressures as high as 1 kbar. Polymer monolithic columns have a large margin of improvement for the analysis of small molecules: The design of the polymer skeleton needs to be revisited to produce a rigid mesoporous structure. 

Three-dimensional printing of polymers by extrusion, photopolymerization, and stereolithography will open new opportunities in separation science provided that these technologies will be able to produce highly ordered mesoporous structures over large enough build ranges-say, a few cm-along with small enough feature sizes of around 1 µm. Generating suitable mesopore structures with proper surface chemistries in 3D printing materials may also remain a practical limitation.

Xiaoli Wang: I don’t believe that monoliths produced using the current processes are competitive compared to packed column technology. The first generation of monoliths was less efficient than sub-2-µm totally porous particles combined with UHPLC and, in my opinion, the same is true for the second generation of monoliths compared to superficially porous particles. Also, because each monolith is a separate manufacturing process, column-to-column reproducibility is always a concern, which is perhaps a perception hard to change, especially for the quality assurance (QA) and quality control (QC) laboratories in the pharmaceutical industry. Micromachining and 3D printing are encouraging in terms of eliminating the radial heterogeneity in current monoliths. It remains to be seen how mesopores can be created and controlled. This would be the key for them to be competitive with particle columns.

Frank Steiner: There is clear potential in these phase architectures with respect to optimizing the efficiency‑to‑permeability ratio. However, manufacturing reproducibility is still impeding their success, mainly because of (at the moment) conventional approaches for surface modification that follow after the micromachining or 3D printing fabrication of the support structures. If this can be overcome, it should have a great commercial future for column manufacturing in the nano- and capillary range and facilitate substantial increase of column lengths. However, if LC does not go nano or capillary in the future, such column types will remain a niche.

Mel Euerby: To date monolithic columns in a standard column format have not really taken off. However, the new micromachined and 3D printed column formats possess the potential to have an impact on all fields of nano-LC where connections, that is, tubing and fittings, are a major cause of concern because they introduce significant amounts of dead volume. It must be stressed that to realize their potential, these columns must be shown once again to be reproducible from column to column and that an array of stationary phase chemistries must also be developed.

Gert Desmet: Are current column designs, in terms of dimensions, particle sizes, and housing, also the best possible designs to be used as second-dimension columns in 2D-LC ?

Fabrice Gritti: The second dimension in comprehensive 2D-LC requires short, relatively wide-with respect to the internal diameter of the first dimension column-efficient, and extremely durable columns capable of sustaining many successive pressure transients and fast runs. This suggests developing columns with minimum border effects (frit design) as well as new packing procedures ensuring durability because the chemical and pressure conditions are rapidly changing. Additionally, the hardware design of the second-dimension column in comprehensive 2D-LC could be directly integrated to the switch valve without the need for connecting tubes.

Xiaoli Wang: This depends on the mode of 2D-LC. In heart-cutting or multiple heart-cutting 2D-LC, normal columns can be used in the second dimension. But on-line comprehensive 2D-LC requires extreme speed. In this case, shorter columns packed with small particles with lower dispersion would be highly desirable for the second dimension. Most importantly, such columns need to be highly stable and robust!

Frank Steiner: It depends on how we define 2D-LC. Speed in the second dimension is important and appropriate handling of fraction sizes and fraction solvents from the the first dimension is required. For multi-heartcut workflows, the current column designs are fit for purpose in my opinion, though they typically consume a lot of mobile phase for the second dimension. In comprehensive 2D-LC, however, where a high number of low-volume fractions need to be efficiently transferred, we may go for intelligent microfluidic devices to manage both the fraction modulation and the second dimension separation with much better efficiencies than current switching valves, tubings, fittings, and columns can support.

Dave Bell: I think this is an area that requires significant research. Developments for hardware used in all dimensions and, in addition, interfaces between the dimensions, could benefit from targeted research in this area. There have already been several reports of 3D printing used to develop interfaces.

Gert Desmet: Apart from the size of the packing, what differences do you see for preparative packings compared with analytical packings?

Tivadar Farkas: It all depends on the scale of the purification. Low-surface-area core–shell particles are successfully used today in fast generic purification schemes. Large-scale purification still favours the traditional particle morphology. 

Xiaoli Wang: Preparative LC is a different ball game from analytical LC. It is very economically driven. Packings need to be cheap, available in the order of kilogrammes and tons, and must have very high sample capacity. Ideally, similar packings in smaller particle sizes should be available for both analytical and preparative columns to ease method transfer. In bioprocessing, polymeric particles are preferred because of their bioinertness, but often they have limited mechanical strength. Therefore, more rigid polymer particles with similar separation performance would be highly desirable.

Mel Euerby: From any method translation standpoint, the properties of the small- and large-particle materials must be the same; otherwise the same selectivities may not be produced. The packing efficiency of small versus large column formats can also be an issue.

Gert Desmet: How has mass spectrometry detection affected columns and how might it further affect them in the future? 

Xiaoli Wang: As I mentioned previously, we need columns with less bleed and higher stability because a mass spectrometer is an extremely sensitive detector. Columns that require less salt or buffer to achieve good peak shape would be highly desirable. In terms of speed, faster MS is needed because current highly efficient columns are generating very narrow peaks that are too fast for MS. I believe the situation will continue to improve with the development of new MS technologies.

Fabrice Gritti: Mass detection requires eluent composition that does not severely hamper the ionization yield of the analytes before they enter into the spectrometer. Indirectly, MS is then pushing the development of new column chemistries that will maintain the highest selectivities and resolution levels for MS-friendly buffers, salts, ionic strength, and organic solvent. Despite the undisputable advantage of MS detectors over optical ones for peak identification, MS-compatible conditions may still limit the selectivity and performance levels of LC columns. The best LC–MS performance is usually found by selecting not only the right column chemistry, but also the eluent composition and, sometimes, the post-column make-up flow for better ionization yield.

Mel Euerby: Historically, mass spectrophotometric detection has driven the internal diameter of columns down because of the need to introduce a lower flow rate into the mass spectrometer. Alternatively, this could be achieved by splitting the flow, but this resulted in peak dispersion and hence smaller internal diameter columns are favoured. However, newer MS detectors do not have such impositions so this driver is decreasing, but most people are still moving towards 2-mm i.d. columns on UHPLC systems and 3-mm i.d. columns on conventional HPLC systems with the aim of reducing solvent consumption. The introduction of chip columns directly coupled to MS may tempt users to move to this form of capillary or nano‑LC.

Frank Steiner: Mass spectrometry can deal much better than optical detectors with decreasing peak volumes, where the required reduction of flow cell volumes comes at the expense of signal-to-noise (S/N) ratios, even with the current light-pipe technology that brought substantial improvements. So it drives the miniaturization of column diameters and will continue to do this in the future. While MS can see column bleed at high sensitivity, this is not a problem with optical detectors. MS will continuously drive the development of low‑bleed stationary phases, which is still a challenge with many polar stationary phase types and respective elution modes, such as HILIC.

Dave Bell: Mass spectrometry has had an impact on all aspects of LC. MS has been both an enabling technology, that is, allowing for less dependence on chromatographic resolution as a result of the inherent selectivity (in some cases) as well as a “needy” technique where improvements in column technology are necessary for some separations. MS will certainly continue to play a role in column hardware and stationary phase design. Technologies such as ion mobility spectrometry (IMS) may, indeed, negate the need for chromatographic separations for some applications in the near future.

Gert Desmet: Do you see any new particle technology on the horizon?

Tivadar Farkas: While not real “particles”, micropillar arrays may emerge as fascinating structures in our field. Their qualities have been proposed by Fred Regnier and amply demonstrated by Gert Desmet’s group. Fingers crossed!

Fabrice Gritti: From the breakthrough of commercial sub‑2‑µm particles in 2004, not much has been seen. Smaller sub‑micrometre particles are not expected on the horizon soon because the constraints on the UHPLC instrument and columns are very severe. They might emerge and support niche applications in bioseparations by nano- or micro-LC. In terms of intraparticle structure, nothing is expected for the separation of small molecules because liquid–solid mass transfer resistance is no longer a limiting factor for these highly diffusing analytes. For all that, new particle technology could definitely emerge with either new ordered mesoporous architecture, for example, radially ordered pores in core–shell particles, or wider mesopores to speed up mass transfer of large molecules in analytical columns packed with core–shell particles. This could be beneficial in both high-throughput and high‑resolution applications. 

Xiaoli Wang: I believe in ordered structures. This can mean the use of 3D printing or micromaching to make a perfectly ordered macrostructure with a homogeneous bed. This can also mean creating ordered mesoporous structures, for example, using a method such as pseudomorphic transformation. In addition, driven by the rapidly growing demands of biopharmaceutical analysis, new particles better suited for biomolecule separations will be developed in the near future, including particles with larger pore sizes and less nonspecific binding.

Mel Euerby: Another million dollar question. Silica is still king, but with the ever-increasing move towards biomolecules, I expect that its days will now be numbered for these types of drugs. I would predict a resurgence of polymer-based stationary phase research by all the major players and expect a steady influx of these products onto the market in the near future.

Dave Bell: The past few decades have offered new particle sizes (sub-2-µm) and architecture (superficially porous particles). One area that has received limited attention is the control and optimization of pore size distribution and shape. This is an area that requires advances in both particle development as well as techniques used to visualize pores in a direct manner. Another area of great interest is in alternative surface materials to silica.

Tivadar Farkas is a senior managing scientist at Phenomenex, Inc., in Torrance, California, USA.

Xiaoli Wang is an R&D section manager in the Liquid Phase Separations Division at Agilent Technologies, in Waldbronn,
Germany.

Fabrice Gritti is a principal research scientist at Waters Corporation, in Milford, Massachusetts, USA.

Frank Steiner is a senior manager, application development and scientific advisor in the Chromatography and Mass Spectrometry Division at Thermo Fisher Scientific, in Germering, Germany.

Mel Euerby is a principal in the Liquid Chromatography Centre of Excellence, Shimadzu UK, and a visiting professor at the University of Strathclyde, Glasgow, UK, and the Open University, Milton Keynes, UK.

Dave Bell is a manager in Separations Technology and Workflow Development at MilliporeSigma, in Bellefonte, Pennsylvania, USA.

Gert Desmet is a full professor in biochemical and chemical engineering at the Vrije Universiteit Brussel, in Brussels, Belgium.