Resolution in chromatography is influenced by retention, engineering, and selectivity. Retention relates to how long analytes are in the column; generally, longer retention leads to increased resolution. Engineering relates to the design and operation of the chromatographic system, as expressed by the number of theoretical plates. Generally, more theoretical plates lead to greater resolution. Finally, selectivity is generated by the chemistry of the intermolecular interactions between the stationary phase and the analytes. Larger differences in the strength of these interactions lead to greater selectivity. In this instalment, we will examine the relationship between selectivity and resolution.

Optimum resolution is one goal of any chromatographic separation. Figure 1 shows a simulated chromatogram, using the ProEZGC chromatogram modeler, of two peaks being eluted in a typical gas chromatographic (GC) separation (1). Resolution may be considered a measure of the amount of separation between the two peaks, and it is classically calculated from the chromatogram using equation 1. In short, equation 1 defines resolution as the difference in retention times divided by the average of the peak widths. A resolution calculated as 1.5 or greater indicates baseline separation. The resolution of the two peaks seen in Figure 1 is 1.7, indicating slightly better than baseline resolution, which is seen on the chromatogram.

Resolution can also be thought of resulting from fundamental parameters of the chromatographic system and the separation. These parameters are retention, engineering, and selectivity. For retention, it is intuitive that the longer an analyte pair are on the column, the more time they have to separate, and the better the resolution. For engineering, it makes sense that the dimensions and construction of the instrument and column impact transport efficiency for the analytes and therefore resolution. For selectivity, the difference in intermolecular interaction chemistry between the two analytes and the stationary phase impacts the separation power of the stationary phase and therefore the resolution.

These three ideas are captured in the master resolution equation, seen in equation 2 (2).

The three ideas—retention, engineering, and selectivity—are seen in this equation. Before diving more deeply into the impact of selectivity, we should briefly discuss the other two terms, retention and engineering. Retention is seen in the first term, 

). This term will vary between 0 and 1, depending on the size of the retention factor. Looking at the extremes, at very large retention factor, 

 > 10–20, the term approaches 1 and diminishing return is seen from longer retention time. Likewise at low 

 may have a large impact on resolution. This is why many authors suggest as a first step in method development, to adjust the temperature to give 

 between about 1 and 20. This also provides a limit for run times: if the gas hold-up time is 1 min, this suggests that the total run time should be less than about 20 min; additional run time simply adds time with little impact on resolution.

Engineering is seen in the square root of theoretical plates (

) divided by 4. From our recent discussion of the Golay equation, we see that the number of theoretical plates derives from numerous factors related to the construction of the chromatographic system and the conditions of the separation (3). In contrast with the retention term, longer retention times usually lead to increased 

 through a complex relationship with the retention factor. 

 is more dramatically impacted by column dimensions, carrier gas flow rate, and carrier gas choice, with smaller inside diameter, thinner stationary phase film, and lower molecular weight carrier gas generally favoured. In typical capillary gas chromatographic systems, 

 may vary from about 10,000–100,000, with the term in equation 2 therefore varying from 25–80. Interestingly, while the number of theoretical plates is roughly proportional to column length, here too, there is a diminishing return with longer columns; resolution will only increase by a factor of 1.4 if the column length is doubled, as 

 is doubled. However, the retention time will also be about doubled.

The remaining term in equation 2 relates selectivity and resolution. Not surprisingly, the higher the value for selectivity, the higher the resolution. Since selectivity is calculated as the ratio of adjusted retention times for two adjacent peaks, with the numerator having the larger time, the minimum value is 1. In capillary gas chromatography, selectivity for adjacent, fully resolved peaks can often be as low as 1.02. Before diving into the effect of selectivity on resolution, it is worth taking a quick look at the selectivity term in equation 2. Table 1 shows calculations of the selectivity term, (

 Table 1 shows a roughly linear increase in the selectivity term as the selectivity increases at low selectivity, but curvature and a diminishing return as selectivity increases beyond 1.1. In most situations in capillary gas chromatography, selectivity of adjacent and nearly overlapping peaks will be less than 1.1, which indicates a roughly 10% difference in the two adjusted retention times, or a 1 min difference in adjusted retention time in an overall 10 min separation.

We can now use equation 2 and Table 1 to gain some insight into how selectivity relates to retention (

) and engineering or number of theoretical plates (

) to generate resolution. Table 2 shows calculations of the selectivity required to achieve a resolution of 1.5 with constant 

. This shows the effects of column dimensions and instrument conditions on resolution. Equation 2 was simply rearranged to calculate the selectivity term and then the approximate selectivity was read from Table 1. This is a good point to caution against using equation 2 for rigorous calculations; it is best used for conceptual discussion and thinking about resolution. There are too many underlying variables in the terms used in equation 2 to make it easily useful for rigorous calculations.

 increases, often by using a longer column or one with a smaller inside diameter or stationary phase film thickness, the required selectivity to achieve a resolution of 1.5 decreases, as expected. Remember that an increase or decrease in 

, if achieved using column length, will result in a similar change in retention time. The two extremes in Table 2 are most interesting. At very low 

, as in classical packed columns or very short capillary columns, much higher selectivity is required to achieve separation. This is one reason why there were hundreds of available stationary phases for packed columns when they were in wide use. This is also one reason why we have seen much attention in the literature over the past two decades paid to stationary phase choice in second dimension columns for comprehensive two-dimensional gas chromatography, as these columns are very short, so high selectivity is needed to achieve any separation at all (4).

At the high end, the necessary selectivity continues to decrease with increasing 

, but the return for using a longer or more difficult to use column decreases, with selectivity of approximately 1.02 being about the practical limit for most capillary column configurations. This implies, with an approximate 10 min adjusted retention time, a minimum separation of the peak maxima of about 0.2 min (about 10 s).

Looking now at selectivity and retention, Table 3 shows determinations of the needed selectivity to achieve a resolution of 1.5 with constant 

 = 60,000 (a typical capillary column) and 

, the retention factor, varying from 1–20. As in Table 2, the selectivity term on equation 2 was determined by rearranging equation 2 and the approximate selectivity needed was then read from Table 1. As 

 increases, the necessary selectivity to achieve baseline resolution decreases, but not dramatically. In most cases with this typical column of 60,000 theoretical plates, the necessary selectivity remains in a tight range of about 1.03–1.05. At 

 < 1, which is not advisable as a small variation on 

 can lead to a large shift in resolution, much higher selectivity is needed. At 

 > 20, which is not advisable as the retention times become excessively long and band broadening may make the peaks too short and more difficult to detect, there is not much influence from longer retention on selectivity.

In most of the situations seen in Tables 2 and 3, the necessary selectivity to achieve baseline resolution varies from about 1.02–1.05. We can think of this as a “sweet spot” for selectivity in capillary gas chromatography. Estimating a 10 min separation and 

 = 5, this means a minimum spacing between peak maxima of about 10–20 s, which is commonly seen in many separations.

Influence of Temperature on Selectivity and Resolution

Figure 2 shows a series of simulated chromatograms of a separation of benzene and ethanol on a polyethylene glycol (PEG) stationary phase using Pro-EZGC. Temperature ranges from 70–110 °C with isothermal conditions, and the chromatograms are annotated with the resolution. As discussed in last month’s column (5), selectivity derives from differences in the standard free energy of the mobile phase-stationary phase transition, which is in turn determined by the intermolecular forces that exist between the analyte and the stationary phase. As the temperature increases, the forces of attraction between the analyte and stationary phase decrease as the vapour pressure increases and more of the analyte molecules enter the vapour phase. This is seen in the reduced retention times. Since selectivity is based on differences in these forces, it follows that the strength of each analyte’s forces would decrease at a different rate as the temperature is increased. This effect is clearly seen in Figure 2, as benzene elutes first at the lower temperature, then they are coeluted, and then ethanol is eluted first at the higher temperature. All of these separations were simulated with constant column dimensions and stationary phase.

In equation 2, temperature has effects on all three terms. In Figure 2, in addition to the effects on selectivity, we see the clear effect on retention, lowering retention time and retention factor, and lowering the retention term in equation 2. The effect of temperature on the engineering term and 

, the number of theoretical plates, is complex, but, in short, higher temperature generally leads to smaller 

, through reduced retention factor, which is an important term in the classical van Deemter equation.

Influence of Stationary Phase Choice on Selectivity and Resolution

Stationary phase choice is the major driver of selectivity and a change in stationary phase can often have dramatic and confusing effects on analyte retention and peak spacing. Figure 3 shows the dramatic difference in the separation of benzene and ethanol on two stationary phases: polydimethyl siloxane (PDMS), which is generally considered nonpolar, and polyethylene glycol (PEG), which is considered polar. The separations were simulated at 70 °C, and all of the other conditions and column dimensions were kept identical so the chromatograms are directly comparable. More detailed discussion about the differences in retention for these two compounds on these two stationary phases can be found in last month’s column (5). On the PEG stationary phase, both compounds are more strongly retained than on PDMS, yet the selectivity and, consequently, the resolution is much lower. The longer retention on PEG also leads to broader peak width, showing the effect of engineering. The somewhat higher value of 

 has little influence on the selectivity.

Selectivity, a measure of the ability of a stationary phase to perform a given separation, mainly derives from the intermolecular interactions that occur between the stationary phase and the two analytes that make up a closely eluting peak pair and the separation temperature. Resolution, the goal of any separation method, is a function of retention, selectivity, and engineering of the system. Selectivity influences resolution, as seen in the master resolution equation, which can be used conceptually to analyze the impact of retention, engineering, and selectivity on resolution. Finally, we should note that this is conceptual thinking based only on isothermal separations. The situation with temperature programming is far more complex and will be the subject of a future column.

. https://ez.restek.com/proezgc (accessed 2023-02-01).

. https://www.chromacademy.com/channels/gc-training-courses/principles/gc-chromatographic-parameters/ (accessed 2023-02-01).

Snow, N. H. Is Golay’s Famous Equation for HETP Still Relevant in Capillary Gas Chromatography? Part 2: Assumptions and Consequences. 

 (2), 57–61. DOI: 10.56530/lcgc.na.mm7085y7

Stefanuto, P-. H.; Focant, J-. F. Columns and Column Configurations. In 

Basic Multidimensional Gas Chromatography

Snow, N. H. Using Thermodynamics and Simulations to Understand Selectivity and Retention in Gas Chromatography. 

 (2), 58–62. DOI: 10.56530/lcgc.na.bf1581b9

 is the founding endowed professor in the Department of Chemistry and Biochemistry at Seton Hall University, USA, and an adjunct professor of medical science. During his 30 years as a chromatographer, he has published more than 70 refereed articles and book chapters and has given more than 200 presentations and short courses. He is interested in the fundamentals and applications of separation science, especially gas chromatography, sampling, and sample preparation for chemical analysis. Direct correspondence to: amatheson@mjhli

Articles

GC Connections

What is the relationship between selectivity and resolution?

How Chemistry Determines Separations: The Influence of Selectivity on Resolution

Szabolcs Fekete, winner of the HTC 2022 Innovation Award, reveals the benefits of ultra-short columns to investigate COVID-19.

Q. Your work on the innovative potential of ultra-short columns has been praised highly by key opinion leaders in the chromatography community. Briefly, what was innovative about the development of ultra-short columns and how did this challenge conventional wisdom?

It was already known, and has been studied in the past that, as a result of the on/off elution mechanism observed with large solutes, such as proteins and nucleic acids, in liquid chromatography (LC), the column length has less impact on controlling their retention compared to small solutes (1,2). More precisely, under commonly used linear gradient conditions, only a short inlet segment of a column (0.1–1 cm) effectively retains the large molecules (3). One might refer to this “useful” inlet column section as the “effective column length”. For example, when running a fast linear gradient for an intact monoclonal antibody (mAb) in reversed‑phase LC, an effective length of shorter than 0.5 cm is expected. After the solute passes this inlet section, its local retention factor will be very low, allowing for an even shorter column length with faster gradients.

Therefore, we challenged the current practice of leveraging 10–25-cm long columns for reversed-phase LC with a novel 0.2–2-cm long column format. As we predicted, equivalent separations could be performed on these ultra‑short columns, allowing for a 0.5–1 min analysis time instead of the routine 10–30 min. The main challenge was to redesign the column hardware to minimize the “extra-bed” volume within the column and to estimate the effect of system band dispersion when using these very small volume columns on commercial ultrahigh-pressure liquid chromatography (UHPLC) systems. We found that 15 × 2.1 mm and 20 × 2.1 mm dimensions provided the best compromise between analysis time and efficiency loss caused by extra-bed and extracolumn dispersion (4). The redesign has allowed for very efficient gradient separations of mAbs and antibody-drug conjugate (ADC) samples in the 0.5–1 min interval.

The ultra‑short columns can be used with other retentive chromatographic modes, including ion-exchange chromatography (IEC), hydrophilic interaction liquid chromatography (HILIC), and hydrophobic interaction chromatography (HIC) (5,6).

Q. You recently used these columns to speed up the analysis of COVID-19 antibody therapeutics (6). What was the rationale behind this research?

The rationale behind this research was the COVID-19 pandemic, which necessitated the emergency use authorization (EUA) of several new therapeutics and vaccines. With an EUA submission, characterization data on a new drug candidate and drug development are often performed simultaneously. Hence, there is little time for misguided method development or traditional turnaround times. We have taken this situation as a chance to propose new means of speeding up the chromatographic characterization of protein therapeutics. To achieve high‑throughput reversed‑phase LC, IEC, HILIC, and size-exclusion chromatography (SEC) separations for two COVID-19‑related mAbs (casirivimab and imdevimab), automated systematic retention modelling was combined with the ultra-short columns (6). The optimized chromatographic methods allowed for the comprehensive characterization of a COVID-19 mAb cocktail with a sub-1-h turnaround time. Retention models were also developed, facilitating the virtual transfer of methods without the need for wet experiments, reducing the time needed for development and optimization. This approach presents laboratories with an easily deployable method supported on ultra-short and conventional column formats.

Q. Can you describe the individual applications you investigated and their benefits for the analyst?

 We have developed a variety of applications on ultra-short columns, with one of the most interesting being the demonstration of a short bed column (15 × 2.1 mm) packed with an optimized stationary phase for LC–high-resolution mass spectrometry (HRMS) quantitation of intact proteins (7). This new column allowed rapid separations, facilitated attomole levels of detection, and could be successfully applied to detect and characterize mAbs in biofluids when paired with a new aptamer-based immunocapture approach. Another study reported the use of an ultra-short column for very high temperature, mAb-complex separations (8). In reversed-phase LC, mAb complexes may require extremely harsh, superheated conditions to dissociate and elute as interpretable profiles. On-column degradation is a consequence of superheated conditions, and the application of ultra-short columns and fast gradients serve as a potential solution to reduce degradation by-products and analysis time.

We also recently applied a very unusual approach to decrease the intrinsic gradient steepness of a reversed-phase LC method to separate a complex therapeutic protein sample in 6–8 min. Based on scouting experiments, we found that a small gradient slope (flat gradient) favoured separation. Normally flat gradients are performed by either increasing the gradient time and/or flow rate, or by decreasing the absolute change of mobile phase composition (narrower ΔB% range), leading to a longer analysis time. Instead, we had the idea to decrease the intrinsic gradient steepness using an ultra-short column, which has a low column volume (please note that the intrinsic gradient steepness is proportional with the column volume; however, this fact is not often considered). Figure 1 illustrates the impact of gradient steepness on the separation of this complex protein sample. The chromatograms in Figure 1(a) and 1(b) used a 10-cm long column, while Figure 1(c) corresponds to a 2-cm long column. Using ultra‑short columns, the analysis time was kept at 7 min—countering the commonly accepted notion that longer columns are required for better resolution.

Q. What were the main analytical challenges you had to overcome?

We were extremely confident about the analytical workflow developed for protein separations using ultra-short columns, multi-isocratic elution mode, and on-column fractioning, and were convinced that any sample could be separated. However, we indeed faced an issue with one project that resulted in a set of closely eluting intact proteins. We attempted a number of conditions, which included one of our PhD students evaluating various multi‑isocratic programmes. One day she programmed a negative step gradient that finally solved the issue and a baseline separation could be attained, leading to subsequent work for us to understand how large molecules elute under these conditions (9). To be honest,

I have never seen the use of multiple isocratic steps together with the addition of a steep negative gradient segment. When inserting a negative segment by decreasing the eluent strength, a slightly more retained peak can be parked somewhere along the column while a slightly less retained compound

is eluting. This negative segmented multi-isocratic elution mode can potentially be applied to improve the separation of notoriously heterogeneous biopharmaceutical samples to separate critical peak pairs (for example, intact mAb variants, Fc-fusion proteins, bispecific‑mAb, antibody mixtures, or ADC species). Moreover, an equidistant peak distribution can be achieved if desired.

Another developing challenge is the nonspecific adsorption and undesired secondary interactions between column hardware and biomolecules that continue to plague laboratories, presenting numerous analytical challenges. Novel hydrophilically modified hybrid surfaces continue to expand and facilitate the deployment of more robust and sensitive analyses (10). This approach advances separations with aqueous mobile phases in both SEC and ion exchange for proteins and nucleic acids.

 Recently, we illustrated the potential of so-called “dual stationary phase gradients” (11), corroborating previous reports that a stationary phase gradient can produce selectivity in a separation as effectively as a mobile phase gradient programme—especially under mobile phase gradient elution conditions. In reality, no column is based on one type of interaction mechanism, with many designed to elicit multiple solute to stationary phase interactions. These mixed‑mode columns inspired us to give more careful consideration to the idea of dual stationary phase gradients. With the theory applied in our recent work, modulation of two unique interaction mechanisms across a chromatographic column has the potential to provide previously unseen selectivity. With the increasing prevalence of mixed‑mode columns, we believe there will be ample opportunity to explore these new concepts in experimental work.

We have also presented at HPLC 2022 and ISC 2022 on so‑called “pressure‑enhanced liquid chromatography” (PE-LC), which leverages pressure changes to alter selectivity (12). An apparatus for the facile manipulation of column pressure was assembled through a two-pump system and postcolumn flow restriction, allowing us to program various constant pressure changes and even pressure gradients. With our proposed setup, it was possible to combine eluent and pressure gradients in the same analytical run, improving the separation of insulin from one of its forced-degradation impurities. We believe that this novel PE‑LC approach opens up new possibilities in liquid chromatography, especially for samples containing solutes of moderate to large sizes. I continue to work on understanding peak formation under these pressure gradients and how PE-LC may benefit from large peak compression effects under negative pressure gradients.

We also have numerous ideas and research projects primarily on the chromatographic separation of new modalities such as cell and gene therapies, focusing on new size- and charged-based separations and trying to understand their underlying phenomena.

I feel extremely lucky for the continued support of my colleagues during this work and want to thank Davy Guillarme, Jean-Luc Veuthey, and Amarande Murisier (University of Geneva), and Matthew Lauber, Jennifer Nguyen, and Mike Fogwill (Waters Corporation) for all of their help and contributions.

Snyder, L. R.; Stadalius, M. A.; Quarry, M. A. Gradient Elution in Reversed-Phase HPLC. 

, 1412A−1430A. DOI: 10.1021/ac00264a716

Stadalius, M. A.; Quarry, M. A.; Snyder, L. R. Optimization Model for the Gradient Elution Separation of Peptide Mixtures by Reversed-Phase High-Performance Liquid Chromatography: Application to Method Development and the Choice of Column Configuration.

, 93−113. DOI: 10.1016/S0021-9673(01)81640-X

Fekete, S.; Bobály, B.; Nguyen, J. M.; et al. Use of Ultrashort Columns for Therapeutic Protein Separations. Part 1: Theoretical Considerations and Proof of Concept. 

, 1277–1284. DOI: 10.1021/acs.analchem.0c04082

Fekete, S.; Murisier, A.; Nguyen, J. M.; et al. Use of Ultra-short Columns for Therapeutic Protein Separations, Part 2: Designing the Optimal Column Dimension for Reversed-Phase Liquid Chromatography. 

, 1285–1293. DOI: 10.1021/acs.analchem.0c01720

Navarro-Huerta, J. A.; Murisier, A.; Nguyen, J. M.; et al. Ultra-Short Ion-Exchange Columns for Fast Charge Variants Analysis of Therapeutic Proteins. 

, 462568. DOI: 10.1016/j.chroma.2021.462568

Duivelshof, B.; Zoldhegyi, A.; Guillarme, D.; Lauber, M.; Fekete, S. Expediting the Chromatographic Analysis of COVID-19 Antibody Therapeutics with Ultra-Short Columns, Retention Modeling and Automated Method Development.

, 221, 115039. DOI: 10.1016/j.jpba.2022.115039

Nguyen, J. M.; Liu, X.; DeLoffi, M.; et al. Aptamer-Based Immunoaffinity LC–MS Using an Ultra-Short Column for Rapid Attomole Level Quantitation of Intact mAbs.

, 122694. DOI: 10.1016/j.jchromb.2021.122694

Bobály, B.; Keresztfalvi, A.; Graber, T.; Schwarz, M. A. Superheated Reversed Phase Chromatography with Ultrashort Columns for the Analysis of Therapeutic Proteins.

, 114462. DOI: 10.1016/j.jpba.2021.114162

Fekete, S.; Murisier, A.; Nguyen, J. M.; Lauber, M. A.; Guillarme, D. Negative Gradient Slope Methods to Improve the Separation of Closely Eluting Proteins. 

, 461743. DOI: 10.1016/j.chroma.2020.461743

Fekete, S.; DeLano, M.; Harrison, A. B.; et al. Size Exclusion and Ion Exchange Chromatographic Hardware Modified with a Hydrophilic Hybrid Surface. 

94, 3360–3367. DOI: 10.1021/acs.analchem.1c05466

Fekete, S.; Lauber, M. A. Studying the Possibilities of Dual Stationary Phase Gradients to Explore Alternative Selectivities in Liquid Chromatography.

, 463492. DOI: 10.1016/j.chroma.2022.463492

Fekete, S.; Fogwill, M.; Lauber, M. A. Pressure-Enhanced Liquid Chromatography, A Proof of Concept: Tuning Selectivity with Pressure Changes and Gradients. 

, 7877–7884. DOI: 10.1021/acs.analchem.2c00464

 worked in the pharmaceutical industry in analytical R&D for 10 years before moving to the University of Geneva in Switzerland and working as a postdoc and later as scientific collaborator for a decade. In April 2021, he joined Waters Corporation and now works as a consulting scientist. He has contributed to ~170 peer-reviewed journal articles, co-authored 10+ book chapters, and co-edited three handbooks. His current interests include: separations of new chemical modalities (including oligonucleotides, RNA therapeutics, proteins, antibody-drug conjugates, and gene therapy products), fundamentals of chromatography, column technology, new method development approaches, and modelling. He is well known for his various collaborations with both academic and industrial partners.

He has an h-index of 47, with over 6700 citations (2023, January, Google Scholar).

Featured Interview

Chromatography and  COVID-19

The primary analytical challenge is to selectively extract the target analytes using a suitable sample preparation technique and introduce them into the downstream analytical instrument.

The critical step in the chemical analysis is sample preparation. Sorptive sample preparation techniques are among the new generation of microextraction approaches, and are compliant with green analytical chemistry principles. A recent intercontinental collaboration between two academic research laboratories—the Aristotle University of Thessaloniki, Greece, and the Florida International University, USA—has yielded a significant number of analytical/bioanalytical methods using fabric phase sorptive extraction (FPSE), magnet integrated fabric phase sorptive extraction (MI-FPSE), and capsule phase microextraction (CPME) for the isolation of various analytes from different complex sample matrices. A brief description of these techniques with regards to principle, synthesis, applications, and advantages and disadvantages along with paradigms is presented.

Analytical chemists are routinely tasked with a wide range of challenges, ranging from statistically meaningful sampling all the way to the final step of reporting the analytical data. To execute a successful analytical workflow, all relevant factors must be thoughtfully considered so that reliable analytical data are produced as the final outcome of the process. As an integral component of the workflow, sensitive, precise, accurate, and robust analytical/bioanalytical methods are required to deal with a wide variety of complex sample matrices where target analytes may possess broadly dispersed physicochemical properties and be present at trace levels. In addition, many types of interfering compounds may also be present in the sample matrices, making the isolation of analytes of interest a challenging task.

Apparently, the most critical step in the analytical/bioanalytical workflow is sample preparation. No analytical/bioanalytical chemist can argue that this is also the most tedious, time-consuming, and error-prone step in the entire workflow, and may lead to erroneous results and wrong conclusions. At the same time, this is also the most challenging and uncertain step for separation scientists (1,2).

Fortunately, a plethora of classical and modern sample preparation techniques have been developed during the last couple of decades to achieve the following objectives:

1. To reduce the sample size, simplify the sample matrix and quantitative release of the target analytes from the sample matrix;

2. To make the sample compatible with the analytical instrument used;

3. To remove interferences and cleaning up the sample;

4. To preconcentrate the target analytes;

5. To obtain the highest amount of analytical information from the sample,
thus maximizing the advantages of sensitive and sophisticated instrumentation;

6. To meet the demands of modern high-throughput analytical/bioanalytical laboratories by optimizing relevant parameters including the analysis time, cost, sensitivity, and compliance with strict regulatory guidelines;

7. To comply with the principles of sustainable development and green analytical chemistry demands with regards to samples, reagents, method, waste, operator, and instrument (3,4).

As such, the primary challenge in sample preparation is to selectively and efficiently extract the target analytes from the sample matrix and elute in a suitable solvent compatible with the chromatographic system. As of now, different solvents, sorbent materials, extraction mechanisms, and design configurations are available.

Sorbent-based sorptive extraction techniques have enjoyed tremendous growth during the past decade and continue to be more popular compared to solvent-based microextraction techniques. It should be mentioned that sorbent-based extraction techniques are considered to be able to cover most of the analytical/bioanalytical sample preparation needs, if not all (4).

To this end, a recent bilateral academic research collaboration between two laboratories, one at the Aristotle University of Thessaloniki in Greece, and the other at Florida International University in the USA, have teamed up to advance the benefit of novel sol-gel-based material systems and unique delivery devices, which have yielded a significant number of novel analytical methods. The research collaboration was primarily focused on novel sorbent-based sample preparation techniques including fabric phase sorptive extraction (FPSE), magnet integrated fabric phase sorptive extraction (MI-FPSE), and capsule phase microextraction (CPME) for the isolation of various analytes from complex sample matrices such as environmental water, food, and biological fluids. Under a memorandum of understanding, Florida International University was responsible for developing new, environmentally friendly sample preparation technologies, while Aristotle University of Thessaloniki assumed the responsibility of developing new applications for these innovative sample preparation technologies. A brief description of these techniques with regards to principle, synthesis of sorbents, fabrication of the devices, and advantages and disadvantages along with paradigms are illustrated in this article.

Sorbent-Based Sorptive Extraction Techniques

The seminal invention of solid-phase microextraction (SPME) by Janusz Pawliszyn and Catherine L. Arthur in 1990 (5) and subsequent invention of stir-bar sorptive extraction (SBSE) by Erik Baltussen, Pat Sandra, Frank David, and Carel Cramers in 1999 (6) have initiated a new era of sorbent-based sample preparation technologies. These techniques gained attention as alternatives to solid-phase extraction (SPE) and liquid–liquid extraction (LLE) due to their solventless/solvent minimized operation and many other positive attributes. Although SPME, SBSE, and their different modifications including thin-film microextraction (TFME) (7) undoubtedly deserve credit for beginning a new direction in analytical/bioanalytical sample preparation known as “sorbent-based microextraction”, and for highlighting the importance of sample preparation in the overall analytical/bioanalytical workflow, they also suffer from several shortcomings. These include a limited number of commercially available sorbents, a narrow range of pH stability (pH 2–10), instability and swelling of the coating in organic solvents, poor extraction performance for highly polar analytes, and fragility of the coated fibres; in the case of SPME, limited commercially available sorbent chemistries and recommended only for nonpolar analytes (Log K

 >4), and for SBSE, the necessity of special instruments for thermal desorption.

Most of these shortcomings originate from the viscous nature of the pristine polymers (polydimethylsiloxane, polyethylene glycol) and the polymer immobilization technology on the substrate surface that is primarily based on a free-radical cross-linking process. The high viscosity of the polymers used as the sorbents in SPME and SBSE slows down the mass transfer of the analytes from the sample matrix to the extracting sorbent, resulting in prolonged sample preparation time. In addition, SPME and SBSE do not provide sample preparation solutions for acidic, basic, or zwitterionic compounds. Both techniques primarily serve for gas chromatographic (GC) separation, although there are a few SPME fibres for liquid chromatographic (LC) analysis via solvent desorption, with a potential risk of sorbent coating loss. It is fair to say that SPME and SBSE do not provide the complete sample preparation solution that separation scientists need to serve their analytical challenges (5–8).

As such, many academic research groups have continued to address the shortcomings of SPME and SBSE by developing new sorbent chemistries that possess higher extraction efficiency, faster extraction kinetics, selectivity, pH stability, porosity, and thermal stability. Unfortunately, these new sorbents are not commercially available.

A critical review on the desired attributes of a robust sample preparation technique reveals that such a technique should deliver the following qualities:

(a) Capable of handling all sorts of original samples (biological, environmental, pharmaceutical, food, forensic) without any matrix clean-up to minimize irreversible loss of analytes during the sample preparation;

(b) Capable of simultaneously preconcentrating neutral, acidic, and basic, as well as trace metals;

(c) Able to prepare the sample in a way that can be analyzed by GC, LC, or capillary electrophoresis (CE) to maximize the analytical information;

(d) Easily field-deployable to reduce the cost of transportation, minimize the change in the sample matrix during transportation and storage, and minimize the potential loss of analytes due to the adsorption by the collection container;

(e) Able to eliminate or minimize the usage of organic solvent;

(f) Should be able to support the extensive demands of high-throughput laboratories;

(g) Compatible with green analytical chemistry principles, and meeting as many of these as possible with regards to solvent consumption, energy consumption, waste produced, human health, and environmental impact;

(h) Able to retain the extracted analytes for a prolonged period so that instrument breakdown/natural calamity/power failure or other unintended incidents do not incur any economic/scientific loss to the concerned laboratory;

(i) Economical and able to favour multiple use to reduce the overall cost;

(j) Able to select the best sorbent material for a specific application using a mathematical model-based approach without requiring a subjective decision;

(k) Able to simplify the overall sample preparation workflow;

(l) Able to offer a holistic solution so that an analyst doesn’t need to use different techniques for different type of samples or analytes;

(m) Able to maximize the number of available sorbents so that the analyst has enough freedom in sorbent selection (9).

The players in the “game” of sample preparation include: analytical performance, time and safety issues, sample matrix, sample type, interferences and undesired components, analytes of interest with regards to their physicochemical characteristics and concentration levels, and the analytical instrumentation with regards to the operating costs and availability in the marketplace.

Conventional sample preparation techniques such as SPE and LLE may exhibit better extraction efficiency or clean-up performance, but modern microextraction techniques comply better with the demands for green analytical chemistry principles. A new principle, known as 

 has also been proposed recently that aims to reconcile the principles of green analytical chemistry and functionality (10).

Sorptive extraction techniques are widely used prior to chromatographic analyses in a range of applications. FPSE, MI-FPSE, and CPME are notable among the novel approaches, with minimum steps needed to achieve satisfactory clean-up and analyte preconcentration, without any prior (filtration/centrifugation/protein precipitation) or post (solvent evaporation
and sample reconstitution) sample modification needed. These novel sample preparation techniques were introduced recently by the research group in Florida International University, as described in the patents cited in the references herein (11–14).

Fabric phase sorptive extraction is a new generation sample preparation technique that was invented to address most of the sample preparation challenges that other sorptive microextraction techniques do not. Unlike other microextraction techniques, FPSE combines the extraction principles of SPME (equilibrium extraction) and SPE (exhaustive extraction) by its unique design. The simultaneous exploitation of the two major extraction principles into a single device results in faster mass transfer, short sample preparation time, and exhaustive or near exhaustive extraction of the target analytes even when the extraction is performed in direct immersion extraction mode (as practised in SPME). FPSE also exploits the surface chemistry of the fabric substrate (hydrophilic/hydrophobic), linker chemistry (polarity enhancer or reducer), and organic/organic-inorganic polymer chemistry that collectively determine the overall selectivity and extraction efficiency of the FPSE membrane. By changing any of the three building blocks of the FPSE membrane, one may fine-tune the selectivity and extraction efficiency of the membrane. Unlike classical microextraction techniques, FPSE has adopted sol-gel coating technology for the sorbent coating. A sol-gel coating chemically binds the organic/organic-inorganic polymer to the fabric substrate via a sol-gel linker that results in thermal, chemical, and solvent stability to the sorbent coating, with pH stability ~1–13. Sol-gel-derived sorbents are inherently porous and possess sponge-like porous morphology. As a result, the sample matrix can easily permeate through the sorbent as well as the pores of the membrane and the analytes are extracted onto the sol-gel sorbent coated on the FPSE membrane via numerous analyte-sorbent intermolecular/ionic interactions. Due to the strong intermolecular/ionic interactions between the analytes and the sorbent, analytes remain retained on the FPSE membrane for a prolonged time (more than a year), even when the FPSE membranes are stored in a regular laboratory environment. The sponge-like porous morphology of the sol-gel sorbent requires a very small volume of organic solvent for quantitative analyte back-extraction. As a result, no solvent evaporation and sample preconcentration is needed in a FPSE workflow, minimizing the risk of analyte loss and improving the quality of the analytical data (15).

Fabric phase sorptive extraction has also expanded the available sorbent chemistries by developing phases that are either exclusively used as SPME and SBSE coating or as an SPE sorbent. As of now, FPSE offers ~30 unique sorbent phases including polydimethyl siloxane (PDMS), polyethylene glycol (PEG), polyacrylate (PA), C

, cation exchanger, anion exchanger, mixed-mode, zwitterionic, and multimode zwitterionic sorbents. The selection of an SPME fibre and SBSE phase primarily depends on the experience of the analyst. To eliminate the error involved in the subjective judgement of the analyst during sorbent selection for a given application, FPSE has developed a mathematical model-based approach using the logK

 (octanol-water coefficient) values of the analytes. Every FPSE sorbent chemistry has its own absolute recovery calculation equation. The analyst can easily calculate the predicted percentage absolute recovery value of a specific analyte on a particular FPSE sorbent using this model, and can decide which sorbent to try for the analysis (9,16). For example, absolute recovery in sol-gel poly(ethyl glycol) 300 coated cellulose FPSE membrane can be calculated using the equation: Recovery % = 14.758805 +16.309632*Log K

. Figure 1 demonstrates the quadratic mathematical model developed for a sol‑gel poly(ethylene glycol)-coated cellulose FPSE membrane. The model was created using 10 compounds at a concentration of 10 µg/mL. Extraction conditions are stated in Figure 1. The 10 test compounds with varied physicochemical properties and functional compositions included: furfural alcohol (FA), piperonal (PIP), phenol (PHE), 4-nitrotoluene (4NT), 9-anthracene methanol (9AM), naphthalene (NAP), tetramethyl benzene (TMB), tricosan (TCL), and diethylstilbestrol (DESB). As can be seen, the absolute recovery of a test compound depends on the logK

 as well as the functional composition of the analyte. As such, absolute recovery of phenol was lower than piperonal in sol-gel poly(ethyl glycol) 300 sorbent, despite the fact that piperonal is more polar than phenol.

Unlike other sorbent-based extraction and microextraction techniques, FPSE simultaneously exploits the surface chemistry of at least three building blocks: (a) the fabric substrate; (b) the sol-gel linker; and (c) the organic/inorganic-organic polymer. The overall polarity/selectivity/extraction efficiency of an FPSE membrane depends on all three building blocks, although the individual impact may not be equal. The FPSE membranes behave differently when the substrate is changed from 100% cotton cellulose to 100% polyester, keeping the other two building blocks unchanged. A schematic representation of an FPSE membrane and its different building blocks is presented in Figure 2.

Figure 3 displays the typical steps involved in an FPSE process.

Magnet Integrated Fabric Phase Sorptive Extraction (MI-FPSE)

Magnet integrated fabric phase sorptive extraction is a new type of FPSE that was created (a) to improve the extraction reproducibility; (b) to expedite the analyte mass transfer; (c) to eliminate the necessity of an external magnet; and (d) to use multiple sorbent chemistry in-tandem to maximize the extractive power when a wide range of analytes (neutral, acidic, basic, polar, nonpolar) are simultaneously present in the sample matrix and are all analytes of interest. MI-FPSE offers all the advantages of classical FPSE in addition to several others. Due to the precise control of the spinning rate (rpm), MI-FPSE demonstrates higher run-to-run extraction reproducibility. In addition, the sample matrix permeates through the FPSE membrane in MI-FPSE at a faster rate compared to classical FPSE, with the extraction equilibrium a significantly shorter time. Figure 4 presents MI-FPSE devices and the extraction process (17).

MI-FPSE improves extraction reproducibility by providing controllable spinning on a magnetic stirrer. During the spinning, it collides with the aqueous sample matrix and forces the sample to permeate through the membrane, resulting in faster mass transfer and shorter extraction equilibrium times. Although FPSE allows sandwiching of several FPSE membranes together, MI‑FPSE by design requires a minimum of two membranes in the device. As a result, MI-FPSE allows easy adaptation of different sorbent chemistries to maximize the extraction efficiency.

Although FPSE and MI-FPSE have provided a holistic solution to the analytical challenges often encountered in modern analytical and bioanalytical laboratories, when the sample contains high percentages of interfering components such as plant debris, soil particles, proteins, fat molecules, and other cellular debris, the sample generally requires sample pretreatment processes such as filtration or centrifugation. Sample pretreatment is not only laborious and time-consuming but it may also inflict loss of analytes during the process. Direct contact of the interfering elements with the sorbent may shorten its active service life. The interferents may irreversibly adsorb onto the surface of the extraction sorbent coating and severely impede analyte-sorbent interaction, resulting in poor extraction recovery. To eliminate the need for sample pretreatment from the sample preparation workflow, the CPME device was developed as a stand-alone sample preparation device. The apparatus works as a filtration and extraction device, with built-in stirring capability. Figure 5 presents the CPME device, the scanning electron microscope (SEM) image of the membrane, and the image of a magnet used in the device fabrication (18).

CPME devices utilize two porous polypropylene tubes with internal diameter (i.d.) 1.8 mm, outside diameter (o.d.) 2.7 mm, and wall thickness 0.45 mm conjoined together with a magnet encapsulated in one tube and the sorbent microparticles in the other. The sol-gel-derived sorbents are distributed within the pores of the tube as well as in the lumen of the tube. The sorbent is created by an 

 sol-gel synthesis while keeping the porous polypropylene tubes submerged into the sol solution. The sol solution enters the porous tubular membrane and fills up the lumen. The sol solution subsequently turns into a solid gel, forming scattered particles inside the porous walls and a monolithic bed inside the lumen of the empty tube. To expand the contact surface area of the sol-gel sorbent, the CPME device is subjected to ultra-sonication, during which the monolithic bed inside the lumen disintegrates into microparticles and expands the available contact surface area greatly. As such, CPME devices possess equivalent sorbent loading and comparable contact surface area when compared to FPSE membranes. When a CPME device is exposed to the sample for analyte extraction, the aqueous sample easily permeates through the porous walls (0.2 μm pore size) and the analytes interact with the sorbent particles trapped inside the porous walls, as well as with the particles inside the lumen, and are extracted rapidly. The porous walls of the device act like a filter. As such, the device can be inserted directly in a sample containing high percentages of matrix interferents. The magnet encapsulated within the device spins the device on a magnetic stirrer. After the extraction, the CPME device is transferred into a back-extraction tube and back-extraction is performed using 200–500 μL organic solvent. Similar to FPSE and MI-FPSE, an CPME device can be used many times without sacrificing its extracting power (18).

The workflow for CPME is presented in Figure 6.

One can compare a CPME device with hollow fibre membrane-protected SPME; however, they are built on very different concepts. Hollow fibre membrane-protected SPME utilizes a porous polypropylene capillary to protect the SPME coating from fouling. It does not have a built-in stirring mechanism. A CPME device is a stand-alone sample preparation device with built-in filtration and mechanical stirring capability. The sol-gel monolith sorbent technology allows the sol-gel sorbent to be created as a monolithic bed within the lumen, as well as inside the porous walls of the polypropylene capillary tube. The monolithic bed of the sol-gel sorbent is disintegrated into microparticles. In this regard, a CPME device mimics a protected micro-SPE device.

FPSE, MI-FPSE, and CPME Method Development

Similar to classical sorbent-based microextraction techniques, FPSE, MI‑FPSE, and CPME need to optimize several process parameters to maximize the extraction efficiency of the target analytes from a given sample matrix. These process parameters include: (a) selection of the sorbent chemistry; (b) optimization of the sample volume; (c) optimization of the extraction time; (d) optimization of the desorption solvent; (e) optimization of the desorption time; and (f) optimization of the stirring speed. The volume of the desorption solvent can be optimized, but instead of optimizing the solvent volume, the solvent volume can be set at the level that is needed to submerge the membrane or the device. If the analytes are polar, evaluating the impact of adding salt to the sample is also recommended. The addition of salt decreases the solubility of polar compounds in water and forces the analytes to be extracted into the microextraction membrane or device. Instead of this one factor at a time (OFAT) approach, statistical design of experiments (DOE) can also be used to screen the relevant process parameters and optimize the most applicable parameters using response surface methodology (RSM) (9).

Applications of FPSE, MI‑FPSE, and CPME

The successful and highly productive research collaboration was begun in 2014, and 22 analytical/bioanalytical methods have now been published, covering a wide range of applications. A list of these methods are presented in Table 1 (19–39).

All sample preparation materials were prepared at the International Forensic Research Institute, Department of Chemistry and Biochemistry of Florida International University, and the protocols in each application were developed and evaluated in the Laboratory of Analytical Chemistry at the School of Chemistry of Aristotle University of Thessaloniki.

It should be mentioned that FPSE, MI-FPSE, and CPME offer many biocompatible sorbents. As a result, no protein precipitation step was needed when dealing with any biological samples presented in the list.

The new generation sorbent-based sorptive extraction techniques based on FPSE, MI-FPSE, and CPME seem to be very promising for the extraction and preconcentration of various target analytes from a wide range of sample matrices. The unique geometrical shape and characteristics of the FPSE membrane result in the rapid extraction of analytes due to the fast extraction equilibrium. Usage of a small volume of solvent, as well as elimination of sample pretreatment, solvent evaporation, and sample reconstitution steps from the sample preparation workflow have made these sample preparation techniques environmentally friendly and cost-effective. With regards to bioanalysis, the FPSE membrane can be directly introduced into the whole blood, intact urine, or unfiltered saliva samples without deproteinization or other sample pretreatment processes. A minimum number of sample preparation steps reduces potential sources of errors. The direct introduction to the high performance liquid chromatography (HPLC) instrument after extraction makes this method suitable for routine analysis either for clinical studies or forensic purposes. The FPSE membrane can be reused several times depending on the application.

MI-FPSE efficiently integrates the stirring and extraction mechanism into a single sample preparation device. This configuration permits the development of simple, direct, reproducible, and accurate processes. Fast extraction equilibrium can be achieved, resulting in shorter extraction times. The extraction device can be reused several times with no carryover effect and no discernible loss in the extraction efficiency.

CPME offers simplicity in sample handling and rapid extraction kinetics, and it can be considered as an environmentally friendly extraction technique compared to conventional sample preparation approaches. Moreover, the extraction step can be parallelized (for example, on a multi‑position stirring plate) for the simultaneous preparation of many samples. The microextraction capsules are reusable at least 10 times. High enrichment factors result in satisfactory limits of detection (LOD) and limits of quantification (LOQ) values that may also be improved using more sensitive instrumentation (LC–tandem mass spectrometry [MS/MS] or GC–MS/MS instruments).

Finally, an important benefit of CPME is that the method does not require sample filtration prior to the sample preparation. As such, CPME can be employed for the direct sample preparation of environmental samples or biological samples that contain solid particles.

Malcangi, S.; Romagnoli, M.; Beccaria, M.; et al. Modern Sample Preparation Approaches for Small Metabolite Elucidation to Support Biomedical Research. 

, e100017. DOI: 10.1016/j.sampre.2022.100017

Clark, K. D.; Zhang, C.; Anderson, J. L. Sample Preparation for Bioanalytical and Pharmaceutical Analysis. 

 (23), 11262–11270. DOI: 10.1021/acs.analchem.6b02935

Gałuszka, A.; Migaszewski, Z.; Namieśnik, J. The 12 Principles of Green Analytical Chemistry and the SIGNIFICANCE Mnemonic of Green Analytical Practices. 

, 78–84. DOI: 10.1016/j.trac.2013.04.010

Kabir, A.; Locatelli, M.; Ulusoy, H. I. Recent Trends in Microextraction Techniques Employed in Analytical and Bioanalytical Sample Preparation. 

Arthur, C. L.; Pawliszyn, J. Solid Phase Microextraction with Thermal Desorption Using Fused Silica Optical Fibers. 

 (19), 2145–8. DOI: 10.1021/ac00218a019

Baltussen, E.; Sandra, P.; David, F.; Cramers, C. Stir Bar Sorptive Extraction (SBSE), A Novel Extraction Technique for Aqueous Samples: Theory and Principles. 

 (10), 737–747. DOI: 10.1002/(SICI)1520-667X(1999)11:10<737::AID-MCS7>3.0.CO;2-4

Bruheim, I.; Liu, X. C.; Pawliszyn, J. Thin-Film Microextraction. 

Kabir, A.; Samanidou, V. Fabric Phase Sorptive Extraction: A Paradigm Shift Approach in Analytical and Bioanalytical Sample Preparation. 

Nowak, P. M.; Wietecha-Posłuszny, R.; Pawliszyn, J. White Analytical Chemistry: An Approach to Reconcile the Principles of Green Analytical Chemistry and Functionality.

, 116223. DOI: 10.1016/j.trac.2021.116223

Kabir, A.; Furton, K. G. Fabric Phase Sorptive Extractors. US 9283544, 2014.

Kabir, A.; Furton, K. G. Fabric Phase Sorptive Extractors. US 9557252, 2016.

Kabir, A.; Furton, K. G. Field Sampling Kit for Chemical Recovery, Storage, and Profiling, Method of Making and Using the Kit, and Dynamic Fabric Phase Sorptive Extraction (DFPSE) Medium. US 9527059, 2016.

Kabir, A.; Furton, K. G. Microextraction Capsules and Method of Making. US 10352833, 2019.

Kabir, A.; Furton, K. G.; Malik, A. Innovations in Sol-Gel Microextraction Phases for Solvent-Free Sample Preparation in Analytical Chemistry. 

, 197–218. DOI: 10.1016/j.trac.2012.11.014

Zilfidou, E.; Kabir, A.; Furton, K. G.; Samanidou, V. Fabric Phase Sorptive Extraction: Current State of the Art and Future Perspectives. 

 (3), 40. DOI: 10.3390/separations5030040

Alampanos, V.; Kabir, A.; Furton, K.; Samanidou, V. Magnet Integrated Fabric Phase Sorptive Extraction of Selected Endocrine Disrupting Chemicals from Human Urine Followed by High-Performance Liquid Chromatography - Photodiode Array Analysis. 

, 462459. DOI: 10.1016/j.chroma.2021.462459

Samanidou, V.; Georgiadis, D. -E.; Kabir, A.; Furton, K. G. Capsule Phase Microextraction: The Total and Ultimate Sample Preparation Approach. 

Samanidou, V.; Kabir, A.; Galanopoulos, L.; Furton, K. G. Fast Extraction of Amphenicols Residues from Raw Milk Using Novel Fabric Phase Sorptive Extraction Followed by High-Performance Liquid Chromatography-Diode Array Detection. 

, 41–50. DOI: 10.1016/j.aca.2014.11.036

Samanidou, V.; Kaltzi, I.; Kabir, A.; Furton, K. G. Simplifying Sample Preparation Using Fabric Phase Sorptive Extraction Technique for the Determination of Benzodiazepines in Blood Serum by High-Performance Liquid Chromatography. 

Karageorgou, E.; Manousi, N.; Samanidou, V.; Kabir, A.; Furton, K. G. Fabric Phase Sorptive Extraction for the Fast Isolation of Sulfonamides Residues from Raw Milk Followed by High Performance Liquid Chromatography with Ultraviolet Detection. 

, 428–436. DOI: 10.1016/j.foodchem.2015.09.060

Anthemidis, A.; Kazantzi, V.; Samanidou, V.; Kabir, A.; Furton, K. G. An Automated Flow Injection System for Metal Determination by Flame Atomic Absorption Spectrometry Involving On-Line Fabric Disk Sorptive Extraction Technique. 

, 64–70. DOI: 0.1016/j.talanta.2016.05.012

A recent intercontinental collaboration between two academic research laboratories—the Aristotle University of Thessaloniki, Greece, and the Florida International University, USA—has yielded a significant number of analytical/bioanalytical methods using fabric phase sorptive extraction (FPSE), magnet integrated fabric phase sorptive extraction (MI-FPSE), and capsule phase microextraction (CPME) for the isolation of various analytes from different complex sample matrices. 

March 2023 | LCGC Europe

How Chemistry Determines Separations: The Influence of Selectivity on Resolution

Chromatography and COVID-19

Novel Sorptive Sample Preparation Techniques for Separation Science

A Rapid, Automated Extraction Platform to Assess Drug Product Potency by Online Liquid Chromatography (LC)