Professor Mary Wirth and graduate students Bingchuan Wei and Benjamin Rogers from Purdue University demonstrate a quantum leap in protein column efficiency. Using colloidal silica particles of submicrometer diameters (470 nm), they obtained plate heights that were as much as 15-fold lower than the theoretical limit for Hagen-Poiseuille flow. The smallest plate height was 15 nm, 500-fold smaller than previous reports for pressure-driven protein chromatography. This remarkable achievement was attributed to "slip flow," which is explained in this installment. The authors also present a practical example using this new concept. 

The plate heights of commercial liquid chromatography (LC) columns have changed little over the last 20 years, following the trend of 

 is the particle diameter (1). Simulations have shown that 

 is the limit for nonporous particles (2). With particle sizes dropping by only a factor of about two during the last 20 years, as such plate heights have improved by the same factor. The reason is not a lack of diligence or creativity; instead, it is because of the fundamental limits of mass transport, with the ultimate limit being the velocity spread of the mobile phase across the spaces between particles. We recently used a nanofluidic phenomenon called 

 (3–8), which greatly reduces this velocity spread, thereby reducing the plate height dramatically. We demonstrated a plate height as low as 0.03 

, which was 15 nm (9). This opens a new horizon for efficiency in ultrahigh-pressure liquid chromatography (UHPLC). 

The concept of slip flow is illustrated in Figure 1. Imagine that the walls are the surfaces of two adjacent particles. The familiar Hagen-Poiseuille flow is depicted with its parabolic velocity profile, in which the velocity goes to zero at the walls. This is the realm of chromatography today. The figure also depicts slip flow, in which the velocity does not go to zero at the walls, thereby giving rise to a narrower distribution of velocities in the column. Slip flow occurs when the molecular interactions between fluid and the surface are weak, allowing the fluid at the interface to have a nonzero velocity. The existence of slip flow is common knowledge in the field of fluid dynamics, where it is typically a small correction factor when there are weak fluid–surface interactions, but it becomes a big factor when the channel reaches the scale of tens of nanometers (3–8). Consequently, slip flow always occurs in reversed-phase LC, but it is not distinguishable from Hagen-Poiseuille flow because the dimensions of chromatographic materials are relatively large. The nanoscale channel dimensions are achieved by using submicrometer particles.  

Figure 1: Illustration of how slip flow gives a narrower velocity distribution compared to Hagen Poiseuille flow. The velocity is far from zero at the wall for slip flow.			

Experimentally, the parabolic flow profile cannot be directly viewed on the nanoscale, but it can be inferred because one can measure a companion to the narrower flow profile: a faster flow rate. Because slip flow gives a nonzero velocity at the wall, the volume flow rate, 

, is enhanced relative to that for Hagen-Poiseuille flow, 

. In equation 1, for a capillary of radius

 is the slip length, which is defined as the ratio of the velocity to shear rate, each at the wall.  

For water in contact with various hydrocarbon surfaces, the slip lengths are all under 100 nm (3,5,7,8); therefore, channel radii less than 100 nm will give noticeable slip flow. Having 

 in the denominator explains that the smaller the channel, the greater the enhancement from slip flow. 

Figure 2 shows experimental results for volume flow rates we have reported for a capillary packed with a 2-cm plug of 470-nm particles (9). These particles formed colloidal crystals inside the capillary (10,11). They were then chemically modified to have a mixed self-assembled monolayer of butyl groups and methyl groups (12). For varying mobile-phase composition, the flow rates are normalized by mobile-phase viscosity after accounting for its small pressure dependence over this range (13). These lines would all coincide for Hagen-Poiseuille flow; instead, they are quite different from one another. Furthermore, the slopes increase in order of immiscibility with alkanes, which is consistent with the idea that weaker interactions between the fluid and the surface give increased slip flow. The data show that the slope for toluene coincides with the line calculated for Hagen-Poiseuille flow (dashed line), whereas the slope for water is five-fold higher. The reason is that slip flow is negligible for toluene because it interacts well with the hydrocarbon surface, whereas slip flow dominates for water because of its very weak interactions with hydrocarbons. The toluene data confirm that the material does not have parasitic flow from cracks, wall gaps, or other defects that would enhance flow in an undesirable way.  

Figure 2: Plot of flow rate vs. pressure, normalized for viscosity, illustrating that slip flow increases with solvent immiscibility in hydrocarbon. The dashed line is calculated for Hagen-Poiseuille flow; other lines are linear fits to the flow data. L = 2 cm. Adapted from reference 9 with permission.			

In Figure 3, chromatograms for a commercial monoclonal antibody are compared for a silica colloidal crystal (Figure 3a) like the one used for Figure 2, and a 5-cm-long commercial wide-pore UHPLC column (Figure 3b) (9). The silica colloidal crystal can separate the antibody from its aggregates in 1/10 the time and at room temperature. The peak sharpness enables the use of low retention conditions to resolve these analytes at room temperature. Because elevated temperature promotes antibody degradation and aggregation, slip flow allows operation under conditions that minimally alter the protein. Fluorescence imaging was used to sample the sharp peak of Figure 3a with high point density, revealing that the plate height is 70 nm for the monoclonal antibody peak (9). An example of fluorescence imaging is illustrated in Figure 3c for the lowest plate height we have observed, which is bovine serum albumin (BSA). Over-labeling yields two peaks because of one versus two dye labels on the protein. Figure 3c shows an image of the two BSA peaks as they approach the exit of a 2-cm-long capillary. From Gaussian fitting of the peak profiles, shown in Figure 3d, the standard deviation of the earlier eluted peak is 16 μm, which corresponds to a plate height of 15 nm for the separation distance of 17.5 mm, thus approaching the diffusion limit (9). This plate height is a factor of 500 times smaller than the plate height of the protein chymotrypsin, which was measured for 1.5-μm nonporous silica particles in reversed-phase UHPLC (14). The number of plates for this separation distance of 17.5 mm is 1.2 × 10

  plates. The only other example of a plate number exceeding 1 million was a separation of aromatic compounds using silica monolith of 

 = 340 bar, where the separation time was 1480 min (24 h) (15). This is more than 700 times longer than the 2-min separation time for the BSA sample for essentially the same pump pressure. This illustrates that the flattening of the Poiseuille flow profile under slip conditions has a very large effect on the column length needed to obtain a high separation efficiency (9).  

Figure 3: Chromatograms for monoclonal antibody separation obtained using (a) slip flow (isocratic 45:55 acetonitrileâwater, 0.1% trifluoroacetic acid; P = 460 bar; L = 1.45 cm) and (b) UPHLC (Waters Acquity C4, 1.7 Î¼m; gradient: 10â70% acetonitrile over 10 min; flow rate: 1 mL/min). (c) Imaged peaks for BSA, with second peak caused by double labeling (isocratic 25:75 acetonitrileâwater, 0.1% trifluoroacetic acid; P = 330 bar; L = 1.75 cm). (d) Gaussian fit of image data. Adapted from reference 9 with permission.			

Something remarkable happens when separations approach the diffusion limit: Length no longer contributes to resolution. Of course, for extremely short lengths other factors will eventually enter in, such as injected width. But for reasonable lengths, the equation for plate height, 

, in the diffusion limit simplifies to just the 

 term in the van Deemter equation, as given in equation 2:  

In this case, only the diffusion coefficient, D, and the obstruction of diffusion by the limited column porosity, ε, contribute to 

 for a given mobile-phase velocity, ν. The velocity is determined by the pressure, 

, the mobile-phase viscosity, η, and other factors described by the Kozeny-Carman relation of equation 3, for which we added the flow enhancement factor, 

, it is straightforward to show that this maximum number of plates under the diffusion limit, 

, is dictated by pressure and is independent of length.  

This situation is analogous to capillary electrophoresis, where the plate number is proportional to the applied voltage, but not length (16). Ultimately, injected width and heating impose an upper limit to 

 in capillary electrophoresis (16). For UHPLC, these limitations will not be reached as easily: Injected width is minimized in chromatography by stacking at the head of the column, and heating is minimized because pumping brings in fresh fluid, and solids are better heat conductors than liquids. Radial homogeneity of the packing has been a problem in chromatography (17), but this is not the case for colloidal crystals (10,11). A method for rapid homogeneous packing with negligible eddy diffusion is the subject of a pending patent (18). The independence of plate height on separation length explains why it is possible to achieve such high plate numbers for such short separation distances of 

Using equation 4 to calculate the maximum possible plate number under the diffusion limit, for a typical medium sized protein, 

/s, and for the separation parameters of 

 = 470 nm, ε = 0.28, a flow enhancement of five, and 

. It is important to note that when 1.5 × 10

  plates are observed in 1 cm, for example, one cannot claim this corresponds to 1.5 × 10

 plates/m. The calculation also illustrates why the plate heights are so low for proteins: Their diffusion coefficients are an order of magnitude or more lower than those of small molecules. 

Future Directions in  Slip Flow Separations

Can slip flow be used to advance separations of peptides and small molecules? Reversed-phase LC is widely used for both; therefore, it is worth considering. Smaller particles would be needed for sufficient retention factors. Equation 4 shows that N decreases rapidly with 

. Fortunately, this is offset by an important aspect of slip flow: flow enhancement increases for smaller particles, as per equation 1. This reduces the steep penalty in back pressure as particle diameter decreases. As particle diameters become less than 0.2 μm, the calculated ratio of surface area to mobile-phase volume reaches that of porous silica, offering the capacity that has always been lacking for nonporous particles. With the faster analyte diffusion coefficients, equation 4 shows that 

 would be limited to ~100,000 for an 800-bar pump. These media could thus be quite advantageous in efficiency and even more so for speed. We are beginning to explore this avenue. 

This work was supported by National Institutes of Health (NIH) under grant R01GM101464. 

The senior author has a financial interest in a company that has licensed this technology. 

(1) R.W. Brice, X. Zhang, and L.A. Colón, 

 (2) M.R. Schure, R.S. Maier, D.M. Kroll, and H.T. Davis, 

 (3) C.-H. Choi, K.J.A. Westin, and K.S. Breuer, 

 (4) M. Majumder, N. Chopra, R. Andrews, and B.J. Hinds, 

 (5) X.C. Qin, Q.Z. Yuan, Y.P. Zhao, S.B. Xie, and Z.F. Liu, 

 (6) S. Sinha, M.P. Rossi, D. Mattia, Y. Gogotsi, and H.H. Bau, 

(8) M. Whitby, L. Cagnon, M. Thanou, and N. Quirke, 

(9) B. Wei, B.J. Rogers, and M.J. Wirth, 

(10) D.S. Malkin, B. Wei, A.J. Fogiel, S.L. Staats, and M.J. Wirth, 

(11) B. Wei, D.S. Malkin, and M.J. Wirth, 

(12) R.W.P. Fairbank, Y. Xiang, and M.J. Wirth, 

(13) J.W. Thompson, T.J. Kaiser, and J.W. Jorgenson, 

(15) K. Miyamoto, T. Hara, H. Kobayashi, H. Morisaka, D. Tokuda, K. Horie, K. Koduki, S. Makino, O. Nunez, C. Yang, T. Kawabe, T. Ikegami, H. Takubo, Y. Ishihama, and N. Tanaka, 

 (17) K.D. Patel, A.D. Jerkovich, J.C. Link, and J.W. Jorgenson, 

(18) US patent application 6⅓21,098. Inven-tors: M.J. Wirth, C. Yerneni, S. Ranas-inghe, R.A. Birdsall. 

is a fourth year doctoral graduate student in the analytical chemistry program at Purdue University, where he works in the Wirth group. He earned his B.S. in chemistry at Queen's University in Ontario, Canada.

received his PhD in chemistry from Purdue University in 2011, and he has continued in the Wirth group as a post-doctoral fellow. He earned his B.S. in chemistry at Peking University in Beijing, PRC.

is the W. Brooks Fortune Distinguished Professor in the Department of Chemistry at Purdue University, where she has established a research program combining chemical separations and material science.

Articles

Column Watch

Professor Mary Wirth and graduate students Bingchuan Wei and Benjamin Rogers from Purdue University demonstrate a quantum leap in protein column efficiency. Using colloidal silica particles of submicrometer diameters (470 nm), they obtained plate heights that were as much as 15-fold lower than the theoretical limit for Hagen-Poiseuille flow.

Ultrahigh-Efficiency Protein Separations with Submicrometer Silica Using Slip Flow

In this study a novel, simple, rapid, and effective method was successfully developed for the determination of pesticide residues in red wine samples. Sample preparation involved extraction of pesticide residues into acetonitrile by QuEChERS (quick, easy, cheap, effective, rugged, and safe) and cleanup with a rapid push-through mini-cartridge filter instead of dispersive solid-phase extraction (dSPE). The red wine extract was cleaned up by passing it through a mini-cartridge containing primary secondary amine sorbent that retains organic acids, sugars, and polyphenolic pigments. The cleaned extract was analyzed by liquid chromatography–tandem mass spectrometry (LC–MS-MS). The cleanup procedure was simple and required less than 1 min per sample. Satisfactory recoveries ranging from 81.6% to 112.2% with relative standard deviations less than 10.8% were achieved. The linear dynamic range was 2–400 ng/mL with a correlation coefficient greater than 0.9940. The limit of detection and limit of quantification were in the range of 0.01–0.40 and 0.05–1.33 ng/mL, respectively. Six commercially available red wine samples were tested in this study, three of which were found to be positive for the presence of pesticides.

According to the International Organization of Vine and Wine, in 2010 about 26 billion liters of wine were produced worldwide, of which about 24 billion liters were consumed (1). Wine, especially red wine, is a rich source of polyphenols such as resveratrol, catechin, and epicatechin. These polyphenolic compounds are antioxidants that protect cells from oxidative damage caused by free radicals. Research on antioxidants found in red wine has shown that they may inhibit the development of certain cancers such as prostate cancer (2). In addition, consumption of red wines has been believed to have heart-healthy benefits (2). The application of pesticides such as fungicides and insecticides to improve grape yields is a common practice in vineyards. However, the applied pesticides may permeate through plant tissues and remain in the harvested grapes and subsequent processed products, such as grape juice and wine. Because pesticide residues are a potential source of toxic substances that are harmful to human beings, it is important to test for the levels of pesticide residues in grapes, juice, and wine. Although the European Union (EU) has set maximum residue levels (MRLs) for pesticide residues in wine grapes of 0.01–10 mg/kg (3,4), it has not yet established MRLs for wine. A study of 40 bottles of wine bought within the EU revealed that 34 of the 40 bottles contained at least one pesticide. The average number of pesticides per bottle was more than four, while the highest number of pesticides found in a single bottle was 10 (5).  

The analysis of pesticide residues in red wine is challenging because of the complexity of the matrix, which contains alcohol, organic acids, sugars, phenols, and pigments (such as anthocyanins). Traditional red wine sample preparation methods include liquid–liquid extraction (LLE) with different organic solvents (6,7) and solid-phase extraction (SPE) with reversed-phase C18 and polymeric sorbents (8–10). However, LLE is labor intensive, consumes large amounts of organic solvents, and sometimes forms emulsions, making it difficult to separate the organic and aqueous phases. In contrast, SPE uses less solvent without emulsion formation, but demands more effort for method development. Other methods such as solid-phase microextraction (SPME) (11,12), hollow-fiber liquid-phase microextraction (13), and stir-bar sorptive extraction (SBSE) (14) use little or no organic solvent but are less reproducible. Typical instrumental detections systems include gas chromatography (GC), GC coupled to mass spectrometry (GC–MS), and liquid chromatography coupled to tandem mass spectrometry (LC–MS-MS) (6–14). 

QuEChERS (quick, easy, cheap, effective, rugged, and safe) is a promising sample preparation method that was first reported in 2003 by Anastassiades, Lehotay, and colleagues for the determination of pesticide residues in vegetables and fruits (15). Since then QuEChERS has been widely used for the analysis of pesticides and other compounds of concern in various food, oil, and beverage matrices (16–18). The QuEChERS procedure involves extraction of pesticides from a sample with high water content into acetonitrile with the addition of salts to separate the phases and partition the pesticides into the organic layer. This is followed by dispersive solid-phase extraction (dSPE) to clean up various matrix coextractives and is achieved by mixing an aliquot of sample extract with sorbents prepacked in a centrifuge tube.  

The aim of this study is to develop a method using QuEChERS extraction, but an easier and faster cleanup method compared to dSPE to cleanup red wine coextractives. This novel sample cleanup method is based on a filter-and-clean concept: The red wine extract is pushed through a mini-cartridge containing anhydrous magnesium sulfate and primary secondary amine (PSA) sorbent, residual water is adsorbed onto the anhydrous magnesium sulfate, and red wine coextractives are retained by the PSA sorbent. The purified extract is collected into an autosampler vial and injected into an LC–MS-MS system for analysis without the need for further filtration with a syringe filter. This cleanup procedure is simple and takes less than 1 min per sample. Red wine extracts were assessed for cleanliness based on visual appearance and full-scan chromatograms after cleanup with four traditional dSPE approaches containing different amounts of PSA sorbent and the rapid mini-cartridge filtration approach. The rapid mini-cartridge approach produced a slightly cleaner extract than the dSPE approach containing the same amount of magnesium sulfate and PSA sorbent. However, the cleanup procedure with push-through mini-cartridge filtration was found to be much faster than dSPE. Eight pesticides belonging to insecticide, fungicide, and parasiticide classes were selected for analysis in this study. Polarities of the eight selected pesticides were very different, with the logarithms of the octanol water partition coefficient (LogP) ranging from -0.779 to 5.004. The classes, structures, LogP, and p

values are listed in Figure 1. Among the eight pesticides analyzed in this study, cyprodinil was most often detected on grapes, with chlorpyrifos, diazinone, and methamidophos also frequently found on grapes (19). The recoveries of planar pesticides included in this study (carbendazim, thiabendazole, pyrimethanil, and cyprodinil) are often adversely affected by graphitized carbon black (GCB), a sorbent that is widely used in dSPE to clean up pigmented samples. In this study, PSA sorbent was used instead of GCB for cleanup of red wine samples and the recoveries of these planar pesticides are reported.  

Figure 1: Classes, structures, LogP, and pKa values of the eight pesticides selected in this study.  			

Finally, six commercially available red wine samples were analyzed using this simple, rapid, and effective sample preparation method. Carbendazim was detected in three red wine samples, although the detected concentrations (parts per billion) are much lower than the European or Japanese regulated levels (parts per million) in grapes (20,21). 

HPLC-grade acetonitrile and LC–MS-grade methanol were purchased from Spectrum. Methamidophos, carbendazim, pyrimethanil, diazinone, and chlorpyrifos (all 100 ppm) were purchased from Chem Service. Thiabendazole (1000 ppm), cyprodinil (100 ppm), and pyrazophos were purchased from Ultra Scientific. Triphenyl phosphate (TPP, 5000 ppm) was purchased from Cerilliant and was used as the internal standard (IS) in this study.  

A 100 ppm thiabendazole solution was made by mixing 200 µL of the 1000 ppm stock solution with 1.8 mL acetonitrile. A 2 ppm pesticide working standard solution was made by adding 80 µL each of the eight 100 ppm standards with 3.36 mL of acetonitrile. A 5 ppm IS solution was made by diluting 10 µL of the 5000 ppm TPP stock solution with 10 mL of acetonitrile. 

The six red wine samples tested in this study were provided by coworkers. Portions (10 mL) of the red wine samples were added into 50-mL polypropylene centrifuge tubes (UCT). To prepare fortified samples, red wine samples were spiked with appropriate amounts of the 2 ppm pesticide working standard solution, vortexed for 30 s and allowed to equilibrate for 15 min. A 10-mL volume of acetonitrile was added to each sample and then shaken for 1 min. Salts (4000 mg of anhydrous magnesium sulfate and 2000 mg of sodium chloride) packed in a Mylar pouch (UCT) were added, and the samples were shaken vigorously for 1 min and then centrifuged at 5000 rpm for 5 min. The upper layer red wine extract was then ready for cleanup. 

Two cleanup methods, traditional dSPE and rapid push-through mini-cartridge filtration, were compared for cleanup efficiency of the red wine extract. Four 2-mL dSPE tubes containing 110 mg MgSO

 and 180 mg PSA (D) were tested to compare the cleanup efficiency against the rapid push-through mini-cartridge containing 110 mg MgSO

 and 180 mg PSA (UCT, ECPURMPSMC). For dSPE cleanup, 1 mL of the red wine extract was transferred into the 2-mL dSPE tube, shaken for 30 s, then centrifuged at 10,000 rpm for 5 min. A 0.5-mL volume of the cleaned extract was transferred into a 2-mL autosampler vial and 10 µL of the 5 ppm TPP (IS) solution was added. For rapid push-through mini-cartridge cleanup, 1 mL of the red wine extract was loaded using a nonsterile latex-free syringe with a Luer-lock tip (VWR), the loaded syringe was attached to the mini-cartridge and the extract was pushed through in a slow, drop-wise fashion. The first 0.5-mL portion of the cleaned extract was collected in a 2-mL autosampler vial and 10 µL of the 5 ppm TPP (IS) solution was added. For both cleanup methods, about half the 1-mL portion of the red wine extract was adsorbed onto the sorbents in the dSPE tube or mini-cartridge.  

An Agilent 6890 GC system coupled with a model 5975C single-quadrupole mass-selective detector (MSD, Agilent) was used in this study for the acquisition of full-scan chromatograms of extracts that were prepared using the different cleanup methods. The GC system was equipped with a 30 m × 0.25 mm, 0.25-µm 

 Rtx-5MS capillary column integrated with a 10-m guard column (Restek). A splitless liner with dimensions of 4 mm × 6.5 mm × 78.5 mm (i.d. × o.d. × 

) packed with deactivated glass wool (UCT) was used to introduce the extract onto the GC column. Splitless injections (1 µL) at 250 °C were made with a 50-mL/min split vent at 1 min. Ultrahigh-purity helium at a constant flow rate of 1.2 mL/min was used as the carrier gas. The oven temperature was initially held at 40 °C for 1 min; ramped at 10 °C/min to 300 °C and held for 3 min. The total run time was 30 min with data acquisition beginning at 4 min. The detector interface, ion source, and quadrupole temperatures were set at 280 °C, 250 °C, and 150 °C, respectively. Chromatograms of the red wine extracts were obtained in full-scan mode with a scanning range of 35–700 amu. 

An Accela 1250 LC system coupled to a TSQ Vantage triple-quadrupole MS system was supplied by Thermo Fisher Scientific. A PAL autosampler (CTC Analytics) was equipped for automated sample injections. Xcalibur (version 2.1) software (Thermo Fisher Scientific) was used for data acquisition and processing. The separation of the eight target pesticides was performed on a 100 mm × 2.1 mm, 3-µm 

 Sepax HP-C18 column with a 20 mm × 2.1 mm, 3-µm 

 Restek C18 guard column. The column temperature was maintained at room temperature (~20 °C). The injection volume was 10 µL at 15 °C. Mobile-phase A was 0.1% formic acid in Milli-Q water (EMD Millipore), and mobile-phase B was 0.1% formic acid in methanol. A flow rate of 200 µL/min was used. The gradient program was as follows: 5% B for 1 min, 5–50% B over 2 min, 50–95% B over 5 min, 95% B for 6 min, 95–5% B in 0.2 min, and 5% B for 2 min. 

Tandem MS was operated with heated electrospray ionization (HESI) in positive mode, and the conditions were as follows: spray voltage: 3000 V; sheath gas: nitrogen at 40 psi; auxiliary gas: nitrogen at 10 psi; ion transfer capillary temperature: 350 °C; collision gas: argon at 1.5 mTorr; Q1 peak width: 0.2 Da FWHM (full width half maximum); Q3 peak width: 0.7 Da FWHM. Optimization of the MS-MS transitions (collision energies and S-Lens RF values) was performed individually for each pesticide by infusing 1-µg/mL standard in acetonitrile at 10 µL/min with 50:50 (v/v) mobile phases A and B at a flow rate of 200 µL/min. The two most intense and characteristic precursor–product ion transitions were chosen for selected reaction monitoring (SRM). Acquisition was divided into three segments (0–5 min, 5.01–11 min, and 11.01–16 min) based on the retention times of the target analytes. The retention times, precursor and product ions, collision energies, S-Lens RF values, and dwell times are listed in Table I. 

Table I: Retention times, SRM transitions, and dwell times for target analytes and internal standard (IS) 			

Red wine extracts that underwent cleanup with the rapid mini-cartridge filtration, and with four traditional dSPE tubes containing 110 mg anhydrous MgSO

 and different amounts of PSA were compared as outlined in the "Cleanup" section. Red wine extracts without any cleanup, with dSPE cleanup A, B, C, and D, and with rapid mini-cartridge filtration are illustrated in Figure 2. The red color in the extracts decreases as the amounts of the PSA sorbent increases in the dSPE tubes. The samples analyzed with dSPE (D) and the rapid mini-cartridge containing the same amounts of magnesium sulfate and PSA yielded a similar colorless appearance. Large amounts of PSA (180 mg) contributed to the efficient removal of various matrix coextractives such as organic acids, sugars, phenols, and pigments in the red wine. Figure 3 shows the full-scan chromatograms of four extracts that underwent cleanup with traditional dSPE (Figures 3a–3d) and one with mini-cartridge filtration (Figure 3e). The chromatogram of rapid mini-cartridge filtration was slightly cleaner than that with dSPE (D). In addition, the rapid mini-cartridge filtration approach based on the filter-and-clean concept was simpler and faster than the dSPE approach, and was therefore selected for the cleanup of the red wine samples in this study. 

Figure 2: Photograph of red wine extracts without any cleanup, cleaned up with dSPE A, B, C, and D, and cleaned up with rapid mini-cartridge filtration. 			

Matrix Matched Calibration, LOD, and LOQ 

Calibration curves were obtained by analyzing matrix matched standards, which were prepared by spiking appropriate amounts of the 2 ppm pesticide working solution into blank red wine extracts after cleanup with rapid mini-cartridge filtration. Six matrix matched calibration standards at concentrations of 2, 10, 40, 100, 200, and 400 ng/mL were analyzed. The linear dynamic ranges, regression equations, and correlation coefficients (

Table II: Matrix matched calibration, LODs, and LOQs			

The limit of detection (LOD) and limit of quantification (LOQ) are the concentrations that give signal-to-noise ratios (S/N) of 3 and 10, respectively. In this study they were estimated according to the S/N values of the lowest matrix matched calibration level of 2 ng/mL. The calculated LOD ranged from 0.01 to 0.40 ng/mL and the LOQ ranged from 0.05 to 1.33 ng/mL (see Table II). The minimum reporting limit (MRL) in this study was set at the lowest calibration level of 2 ng/mL. 

Figure 3: Full-scan chromatograms of red wine extracts cleaned up with (a) dSPE A, (b) dSPE B, (c) dSPE C, (d) dSPE D, and (e) rapid mini-cartridge filtration. 			

A chromatogram of red wine sample 1 fortified with 10 ng/mL of the target pesticides is shown in Figure 4. All the peaks, except for methamidophos, were sharp and offered reliable quantification. The satisfactory separation of the eight pesticides is also evident in the chromatogram. This allowed the data acquisition to be divided into three segments which ensured the optimal performance of the MS system, including dwell time (scanning speed) for each analyte. 

Figure 4: Chromatogram of red wine sample 1 fortified with pesticides (10 ng/mL).			

Red wine samples fortified with 10, 50, and 100 ng/mL of the target pesticides were extracted with QuEChERS and cleaned up using the rapid mini-cartridge filtration procedure. The recovery and relative standard deviation (RSD) data are listed in Table III. Recoveries of 81.6–112.2%, with an overall recovery of 97.0%, were achieved with this simple, rapid, and easy-to-use procedure. RSDs of four replicates for each of the three spiking levels were less than 10.8%, which indicated that this method is suitable for the determination of pesticide residues in red wine samples. 

Table III: Accuracy and precision data			

Six red wine samples were tested using the newly developed and validated method. The results of the red wine samples tested are listed in Table IV. Several of the pesticides were detected at low concentrations less than the method MRL. Carbendazim was the only pesticide detected above the MRL. The detected carbendazim concentrations were at 10.2, 8.7 and 2.3 ng/mL in samples 4, 5, and 6, respectively. However, the detected concentrations are much lower than the European (0.5 mg/kg) or Japanese (3 mg/kg) regulated levels in grapes. 

Table IV: Pesticide residues detected in six red wine samples. The minimum reporting limit (MRL) of the method is 2 ng/mL. 			

A simple, fast, novel, and effective cleanup method for red wine samples was successfully developed. Pesticide residues in red wine samples were extracted using the nonbuffered QuEChERS procedure. Cleanup was carried out by passing 1 mL of the red wine extract through a mini-cartridge containing magnesium sulfate and PSA. The magnesium sulfate adsorbed residual water remaining in the acetonitrile extract, and the PSA sorbent retained matrix coextractives, including organic acids, sugars, phenols, and pigments. The cleanup method based on a filter-and-clean concept took less than 1 min per sample, thereby providing higher throughput than the traditional dSPE procedure. Cleaned extract was injected directly into an LC–MS-MS system for analysis. The analytical run required only 16 min and the target pesticides were chromatographically well resolved.  

Good sensitivity and selectivity were achieved for the clean extracts obtained using the rapid mini-cartridge filter and LC–MS-MS detection. Good linearity, low LODs and LOQs, and satisfactory accuracy and precision data were obtained, indicating that this method was suitable for the analysis of pesticide residues in red wine samples. Six commercially available red wine samples were tested with the newly developed and validated method. Carbendazim was present in three red wine samples, although the detected concentrations were far below the European and Japanese regulated levels in grapes. 

Thomas August and Lisa Snyder are acknowledged for the arrangement of the UCT products needed for this study. Catherine Messinger and Evelyn Scanlon are thanked for providing red wine samples. Dr. Brian Kinsella is thanked for proofreading the manuscript and providing valuable discussions and suggestions. 

http://www.oiv.int/oiv/info/enoivbilan2010

http://www.mayoclinic.com/health/red-wine/HB00089

Article

Sample preparation in this study involved QuEChERS extraction followed by a filter-based cleanup method that is easier and faster than dispersive solid-phase extraction.

Determination of Pesticides in Red Wine by QuEChERS Extraction, Novel Cleanup with Rapid Mini-cartridge Filtration, and LC–MS-MS Detection

Gas chromatography (GC) resolution and sensitivity are often limited by matrix effects in "real world" samples that originate outside of a laboratory. GC separation and detection may encounter interference from nonvolatile constituents, side-effects of large sample volumes, and the chemical activity of matrix compounds and derivatization residues. A number of sample preparation techniques increase analyte concentrations and detector response while reducing matrix effects, including classical liquid–liquid extraction, chemical derivatization, and sample preconcentration, as well as headspace, thermal desorption, and large-volume sample injection.   

Solid-phase microextraction (SPME) is a relatively new sample extraction technique — first described in the 1990s by Pawliszyn (1,2) — that brings some unique capabilities to bear on the chromatographic analysis of dilute solutions in difficult matrices, both liquid and gaseous. Essentially, SPME consists of two discrete steps: solute absorption from the sample matrix into a thick layer of silicone or related adsorptive material, followed by transfer of the absorbed analytes into a chromatography inlet system by thermal or liquid desorption.  

SPME has been applied to both GC and liquid chromatography (LC) separations. It eliminates the need for large-volume sample transfer into a GC column by concentrating analytes into the fiber coating while leaving the bulk of the solvent and nonvolatile residues behind. SPME uses orders of magnitude less solvent and has significant potential to greatly reduce or eliminate solvent consumption and the concomitant issues of used solvent disposal as part of sample preparation.  

Chromatographers should not confuse SPME with solid-phase extraction (SPE), a related predecessor with similar applications. The principal difference is that SPE is carried out with a relatively large sorptive wafer the size of a small filter paper and requires post-sorption liquid-phase extraction of analytes; SPME is accomplished with a small fiber or tube coated with sorptive material. SPME is applied to both gas-phase and liquid-phase extraction, whereas SPE is limited to extraction from liquid-phase samples. This 

 installment primarily discusses SPME using absorptive polymeric coatings, although the principles also apply to adsorptive SPME onto active solid layers. 

A related technique, stir-bar sorptive extraction (SBSE) uses a magnetic stir bar coated with a thick layer of absorptive polymer. The stir bar is exposed to sample solution for a time, during which solutes are absorbed into the polymer coating. Subsequently the stir bar is removed, dried, and then thermally desorbed for GC injection, or the absorbed analytes can be back-extracted with a different solvent. SBSE uses a larger volume of absorbent than SPME and, thus, is more efficient at extracting analytes with less absorbent solubility. Therefore, SBSE is generally more sensitive than SPME.  

				SPME applications cover a broad range that includes flower scents (3), chemical warfare agents (4), pharmaceutical process impurities (5), the determination of organochlorine pesticides in Chinese teas (6), volatile compounds in acidic media (7) and cheese (8), volatile phenols in wine (9), environmental pollutants in water samples (10), chloroanisoles in cork stoppers (11), volatile aliphatic amines in air (12), and phenylurea herbicides in aqueous samples (13). 			

SPME relies on the extraction of solutes from a sample into an absorptive layer on the SPME fiber. After a sampling period — during which extraction ideally reaches equilibrium — the SPME fiber and captured solutes within its absorptive layer are transferred into an inlet system that then desorbs the solutes into a gas (for GC) or liquid (for LC) mobile phase. Success relies on choosing conditions so that the desired solutes favor the SPME absorptive layer as much as possible in the presence of bulk sample, and then subsequently the absorbed solutes are released as quickly and completely as possible for chromatographic analysis. Secondary trapping and release of desorbed solutes after SPME is sometimes required when the initial desorption is too slow and the resulting broadened peaks do not yield full utilization of column resolving power. This may be accomplished with a discrete thermal trap or with column stationary phase trapping by injecting onto a cold column and subsequently temperature programming for solute elution. It is also possible to make multiple SPME extractions and desorptions into a secondary trap for enhanced sensitivity. 

 Figure 1: Cross-sectional diagram of SPME extraction devices. Shown are (a) a fiber device with external sorptive coating and (b) a tube device with sorptive coating on the inside. 			

For bulk samples contained in vials or otherwise easily accessed, SPME is performed conveniently with a short absorptive film–coated fiber, as shown in Figure 1a. A short tube coated on the inside with an absorptive layer (Figure 1b) may be used as well for samples that are amenable to pumping. The choice of the absorptive layer chemistry and film thickness strongly influences the degree of absorption and the subsequent efficiency of desorption. SPME fibers are available in syringe-integrated assemblies that are conveniently handled either manually or by robotic autosampling systems. Used fibers often can be cleaned up for reuse by solvent rinsing or baking. 

For the extraction step with an externally coated SPME absorptive layer, the layer is exposed to sample in the liquid (Figure 2a) or gas (Figure 2b) phase. In the case of SPME within a coated tube, liquid or gaseous sample is circulated through the tube until equilibrium is reached. Stopped-flow flow-through sampling is also possible. Over time, the amounts of the solutes in the SPME layer reach an equilibrium level with their surroundings, which represents the maximum solute amounts that can be absorbed and withdrawn under a given set of sampling conditions. The amount of solute in the SPME layer at equilibrium (

) can be approximated by the following equation: 

 is an aggregate solute distribution constant between the SPME absorptive layer and the sample, 

is the solute concentration in the sample before performing SPME sampling. 

Figure 2: SPME extraction from a sealed vial. Shown are diagrams representing sampling (a) from the liquid phase and (b) from the headspace (gas phase). 			

Equation 1 assumes that the sample volume is much greater than the volume of the SPME layer. SPME coatings typically have thicknesses of about 10–100 µm — around 10 times the film thickness range normally encountered in capillary GC columns. The volume, 

, of a 1-cm long by 100-µm thick annular coating on a 0.56-mm o.d. (24-gauge) fiber (as shown in Figure 1a) is approximately 2 µL: 

The volume of an SPME layer with the same thickness coated inside a tube (Figure 1b) with inner radius 

 would be the same. The assumption in equation 1 is valid for sample volumes of over 100 times the SPME layer volume, or more than about 0.2 mL for the thickest SPME layer of 100 µm. Thinner SPME layers, with smaller volumes, would imply correspondingly smaller minimum sample volumes. 

The presence of a gaseous headspace over a liquid sample causes a portion of each solute to partition into the headspace in competition with the process of extraction into the SPME layer. This effect results in a reduction of solute mass in the SPME layer — relative to having no headspace present at all — that depends on both the headspace volume and the partition ratios between the headspace and the liquid sample as well as between the headspace or the liquid sample and the SPME layer. These relationships are somewhat complex and they produce dependencies of the extracted solute mass on the relative liquid and headspace volumes. From a practical point of view it becomes very important to maintain constant sample and headspace volumes across multiple samples to keep such multiphase influences consistent. 

A finite time span is required to reach solute equilibrium between the sample and the SPME layer, which ideally will occur before the extracted solutes are withdrawn for desorption into a chromatograph. As solute molecules are removed from the sample into the SPME layer, additional solute molecules must diffuse into their place at the SPME–sample interface. The process of absorption is limited by the rate at which solute molecules can replenish the transition layer near the SPME interface. Stirring the liquid phase thoroughly helps reduce this time considerably by maximizing exposure of the SPME layer to the sample and greatly reducing the influence of solute liquid-diffusion rates on SPME uptake. However, stirring the liquid sample does nothing to increase the diffusion rate of absorbed solutes inside the SPME layer itself, which then becomes a limiting factor for the rate required to reach equilibrium. SBSE can benefit from decreased sorption times with rapid, turbulent stirring of the sample solution. 

When sampling from the headspace gas instead of the liquid in a two-phase sample system, solute must first cross the liquid–gas interface before encountering the SPME layer. Interestingly, it makes little difference to the ultimate equilibrium solute amounts in the SPME layer whether the sample is obtained from the liquid or the gas. However, the time to equilibrium may be influenced strongly by the choice of the sampling phase. Nonpolar and volatile solutes that strongly favor the headspace phase will come to equilibrium more rapidly if the SPME layer is exposed to the headspace, and the solutes that favor the liquid phase will equilibrate more rapidly directly from the liquid phase. 

Ultimately, chromatographers must characterize the equilibrium times for each solute of interest. If equilibrium is reached in a reasonable time period, perhaps less than 30 min, then they should use a sampling time at least that long. On the other hand, if an unreasonably long time is required, in terms of the time available for sampling and analysis, then it is possible to perform SPME without reaching equilibrium. In that case, the operator must ensure that the same SPME sampling time is used for each sample, and that time should be as long as possible. 

The sample ionic strength, temperature, and other factors that influence the partition coefficients — both between the liquid sample and headspace and between the SPME layer and the liquid — also must be kept under careful control for good sample-to-sample consistency. Adding salt to an aqueous sample often will shift the partition coefficients for nonpolar solutes in favor of the SPME layer as well as speed up the time to attain equilibrium. 

The chemistry of the sorptive SPME layer plays a significant role in enhancing or discriminating against classes of compounds. For the most part, SPME layers for GC absorb solutes in a manner related to their behavior as GC stationary phases: Polar SPME layers such as those containing polyesters or acrylates will enhance polar constituents and discriminate against nonpolar materials. Adsorptive layers with active carbon constituents will more strongly retain volatile components than layers made of nonpolar dimethylsilicones. However, some thought must be made in consideration of desorption as well, a very strongly held solute may prove too difficult to pry off the SPME layer for analysis. Nonpolar solvents may act to shift the effective polarity of the SPME layer as they are also absorbed, and changing the sample's solvent usually will influence the SPME behavior. This can be advantageous in method development, as a means to shift equilibrium as desired, but the solvent should remain fixed once selected. 

The next step after sampling is to transfer the SPME layer and absorbed analytes away from sample exposure and into conditions for desorption into the chromatography mobile phase. For LC–SPME tube sampling with a multiposition valve connected to the SPME tube there is no need to physically move the SPME layer, but new solvating conditions must be established in place that promote solute desorption. An SPME fiber device, on the other hand, is removed from the sample container and transported over a distance to where the solutes are to be desorbed — the GC inlet or a solvent wash position. Removal from the sample environment immediately starts the absorbed solute concentrations shifting away from their in-sample values to lower levels as solutes naturally desorb into their surroundings. The rate of natural desorption is fairly low for many solutes, but volatile molecules may experience significant losses. In a laboratory situation, the transfer time from sample vial to instrument can be short enough so that losses are insignificant. Losses can be minimized during extended transport and storage by sealing the SPME layer into a small enclosure and then ensuring that the contents are included with the rest of the sample upon desorption. In addition to volatile sample losses, an SPME layer easily can pick up nonsample components from the ambient air, especially during extended transportation from remote sites. Enclosing the SPME layer will also prevent the influx of such contaminants. A number of commercially available SPME devices incorporate such a sealing system. 

Once in place at a chromatograph, the SPME layer must then be exposed to conditions that cause the absorbed solutes to desorb with as close to 100% efficiency as possible, and in a sufficiently short time compatible with the chromatography mode in use. In the case of an SPME layer coated inside a tube, for LC analysis, a simple multiposition valve arrangement can switch from sample liquid flow to the mobile phase. Stopping the mobile-phase flow in the SPME tube allows time for solute desorption to come to equilibrium between the SPME layer and the liquid mobile phase before the desorbed materials are introduced into the column. 

Desorption from fiber-based external SPME layers is more conveniently achieved by insertion into a standard GC capillary inlet system in much the same way as a syringe. For LC desorption and analysis, devices are available that wash the external SPME layer with mobile phase and pass the desorbed solutes into the LC injection loop. 

Several trade-offs arise in the course of thermally desorbing an SPME layer for GC analysis. First, the desorption temperature must be high enough so that the solutes leave the SPME layer rapidly. A desorption that is too slow may lead to peak broadening and tailing unless additional arrangements are made for trapping solutes at the beginning of the GC column before temperature-programmed elution. Conversely, too high of an inlet temperature may induce thermal decomposition and introduce some contaminants into the column from septum bleed as well as from the SPME layer itself. 

During sample desorption from an SPME fiber into a split–splitless inlet, the inlet split flow should be turned off so that all of the solutes may enter the column without splitting, as in splitless injection. It is unlikely that enough sample will be absorbed on an SPME layer to necessitate sample splitting. A narrow-bore inlet liner (often called a "splitless" liner) helps produce better peak shapes as well by limiting the volume into which the solutes may expand. After the SPME device has been withdrawn from an inlet splitter the split flow may be turned back on to purge the inlet of any remaining materials and prevent some degree of peak tailing. 

Programmed-temperature vaporizer (PTV) inlet systems are well suited to SPME desorption because of their smaller internal volumes. For SPME use, they should be operated at the same elevated constant temperatures as conventional split–splitless inlets because PTV heat-up rates (on the order of a several hundred degrees per minute) may not be fast enough to produce sufficiently narrow peaks without some form of additional stationary-phase trapping. 

The primary advantages of SPME are its ability to decouple sampling from the matrix effects that would distort the apparent sample composition or disturb the chromatographic separation, its simplicity and ease of use, and its reduced or nonexistent solvent consumption. These characteristics combine to make SPME an attractive alternative to classical headspace or thermal desorption sampling, solid-phase extraction, and classical liquid–liquid extraction. 

As with a number of related sample preparation and injection techniques such as headspace GC or thermal desorption, SPME lends itself well to handling difficult sample matrices, but with the added benefit of low cost and simplicity. SPME doesn't require elaborate and expensive instrument accessories for occasional use and yet seems to be capable of delivering very good manual results in the hands of skilled users. This cannot necessarily be said for manual headspace or thermal desorption sampling. Autosamplers are also available that perform repetitive unattended SPME sampling. 

SPME does require careful optimization and consistent operating conditions for success, but this is true of related techniques as well. Any poorly characterized sampling technique has no valid use in the analytical laboratory, and the burden of developing an SPME method is no greater than for the other techniques. SPME has a significant place in the analyst's arsenal of sample preparation techniques.  

(2) C.L. Arthur, D.W. Potter, K.D. Buchholz, S. Motlagh, and J. Pawliszyn, 

(3) P. Barták, P. Vednár, L. Cáp, L. Ondráková, and Z. Stránsky, 

(4) G.L. Hook, G. Kimm, G. Betsinger, P.B. Savage, A. Swift, T. Logan, and P.A. Smith, 

(5) R.P. Frost, M.S. Hussain, and A.R. Raghani, 

(6) L. Cai, J. Xing, L. Dong, and C. Wu, 

(7) W.A. Araújo, C.A. Lacerda, E.A. Cappelaro, and F.M. Lanças, 

(8) O. Pinho, C. Peres, and I.M.P.L.V.O. Ferreira,

(9) R. Castro Mejas, R. Natera Marn, M.A. de Valme Garca Moreno, and C. Garca Barroso,

(11) F. Bianchi, M. Careri, A. Mangia, and M. Musci, 

GC Connections

SPME decouples sampling from matrix effects that would distort the apparent sample composition or disturb the chromatographic separation. It is also easy to use and reduces solvent consumption.

Solid-Phase Microextraction

The primary objective of an analytical-scale separation is to produce a chromatogram with sharp, well resolved, symmetrical peaks to yield the required analytical information. The goal of preparative-scale high performance liquid chromatography (HPLC) is to produce a quantity of pure compound as easily as possible in the most economical way, ultimately depositing the sample into a sample collector vessel before recovery from the eluent. 

Preparative chromatography can be performed with an analytical column (and system) to produce a few micrograms of material, up to process scale, providing a ton quantities of sample, which use 1-m-long, 200-mm i.d columns. The larger the quantity of analyte required, the further the technique is removed from analytical chromatography, both in terms of scale and ideology; the bigger the scale, the more "nonchromatographic" parameters have to be considered. Some key downstream considerations in preparative chromatography include  					

The amount of material recovered will directly govern the size of the preparative HPLC column chosen and the concentration and number of injections of crude sample required to achieve the desired target analyte mass.

An analytical method that requires a complex gradient may prove difficult to replicate on a large scale.

Large amounts of water, buffer salts, or modifiers in the eluent may cause problems at a later stage if their removal requires complex and costly processing.

The stationary-phase chemistry and particle size should be scalable. 

Elevated temperatures may prove costly to reproduce on a large scale.

Many preparative separations undertaken within the laboratory begin life with a separation carried out at analytical scale. One must develop an analytical method in which the selectivity of the target analytes is maximized. In this way, the amount of analyte that can be loaded onto the column (and hence recovered from the column) per injection is optimized. Elution of analytes in more highly organic fractions is preferred from a sample recovery perspective as these fractions are more quickly and cost effectively processed. 

Studies of the "loadability" of an analytical column are typically carried out using a "standard" 150 mm × 4.6 mm analytical column with 5-μm packing material. As the analyte amount on column increases, the peak shape will deteriorate until resolution is lost between the target analytes; this capacity determines the column loadability for a "touching band" type preparative separation. It is possible to "heart cut" overlapping bands within a separation (where selectivity cannot be further optimized for example), and the fractions can be reinjected to obtain increasingly pure material — however, this approach is much more cumbersome. In separations where selectivity is not the limiting factor, the loading capacity of the column is typically defined by breakthrough, the point at which increasing the amount of analyte on column does not increase peak height or area because the excess analyte cannot bind to the stationary-phase surface. 

When the sample of interest has good solubility in the mobile phase, concentration overload is the technique of choice and sample concentration is increased while the injected sample volume remains constant. Column efficiency (as dictated by particle size) has little effect on concentration overloading, and the selectivity of the separation tends to be the dominant factor. When increasing the sample concentration, it may be necessary to use cosolvents within the diluent (dimethyl sulfoxide and dimethylformamide are popular) to improve sample solubility. If the diluent is more highly eluotropic than the eluent, however, problems can arise with peak shape and with precipitation during post-injection mixing with the eluent (which limits loadability). 

When the sample of interest has limited solubility in the mobile phase, then volume overload is the technique of choice and sample volume is increased while the sample concentration remains constant. Volume overloading is heavily influenced by stationary-phase particle size and column diameter.  Most preparative chromatography methods use a mixture of concentration and volume overloading to obtain the maximum amount of analyte on column per injection. 

After the analytical-scale method has been developed and the loading factor has been estimated, the method can be scaled up using various simple calculations and estimation methods. Scalable factors include eluent flow rate, column internal diameter, gradient profile, sample volume loaded, solvent consumption, total fraction volume, and yield. A detailed treatment of various approaches to method scale-up can be found in the accompanying CHROMacademy Essential Guide on-line article.     

Preparative HPLC equipment differs from analytical scale only in the capability to deliver very high flow rates (100 mL/min is not unusual in a laboratory-scale semipreparative separation), and typically the inclusion of a fraction-collection device for automated sample recovery. Preparative fractions are typically collected by means of a diverter valve that can be triggered either by time settings or detector signal. If the fraction collection is time based, then one must take care to avoid analyte retention time drift. Where the collection is triggered by mass (using mass spectrometry detection), a response threshold, or a rate of change in detector response, the delay volume (time) between the detector and the fraction collector nozzle must be carefully calibrated to ensure precious sample is not lost. 

The Essentials

An overview of the principles and equipment required for scaling analytical separations to the preparative scale and considerations for the scale of analyte required.

Introduction to Preparative HPLC

The solvent chosen for injection can affect the appearance of chromatographic peaks.

A recent observation in the laboratory of one of the authors (B.A.) of this column highlights a problem that has been mentioned in "LC Troubleshooting" discussions before, but its importance justifies additional discussion about the selection of the injection solvent to use with a liquid chromatography (LC) method. In this particular case, it was observed that the peak areas were constant with two different injection solvents, but the peak heights were not. Similar results were observed with methods for two different compounds (A and B). Let's see what the likely cause is for this problem. 

We can't share the complete details of the methods because they are proprietary, but the key elements follow. A 10 ppm concentration of sample is prepared in distilled water or in 100% methanol, and 20 μL of this solution is injected. A 250 mm × 4.6 mm C18 column packed with 5-μm diameter particles is used with UV absorbance detection at 210 nm. For compound A (method A), a mobile phase of 60:40 methanol–water is used at a flow rate of 1.1 mL/min; for compound B (method B), the mobile phase is 70:30 methanol–water with a flow rate of 1.4 mL/min. The data are summarized in Table I. Both compounds (A and B) are well retained with 

-values of  approximately 2.6 in each case. You can see that the peak area differences between injection in methanol and water are ≤0.1% for both methods. However, the peak heights are approximately 30% larger for the water injections.  

Table I: Chromatographic results for two methods A and B 

The chromatograms of Figure 1 illustrate the influence of the injection solvent on the appearance of peaks in the chromatogram for another sample (1). In each case, the reversed-phase column was operated with an 18% acetonitrile–buffer mobile phase, where the buffer was 81:1 water–acetic acid. Injection volumes of 30 μL were made in various injection solvent compositions.  

Figure 1: Chromatograms for caffeine (first peak of each chromatogram) and salicylamide (second peak). Injections of 30 Î¼L of ~1 mg/mL of each analyte dissolved in (a) acetonitrile, (b) mobile phase, (c) 81:1 waterâacetic acid, and (d) water. Column: 250 mm Ã 4 mm, 10-Î¼m dp; mobile phase: 18:81:1 acetonitrileâwaterâacetic acid; flow rate: 1 mL/min. Adapted from reference 1. 

First, consider the cases where the injection solvent is no stronger than the mobile phase (Figures 1b–1d). In each case, the retention times of each peak are approximately the same. Although we don't have data for the peak areas in each case, we'll assume they are the same, because the same volume of the sample at the same concentration was used in each case. Note, however, that the peak heights differ. Water is the weak mobile-phase solvent in this example, and the less water that is present in the injection solvent, the shorter the peak heights are. These observations are consistent with the problem presented at the beginning, where peak areas and retention times were constant, but peak heights dropped when less water was used in the injection solvent. 

Next, notice what happens when the sample of Figure 1a was injected in 30 μL of 100% acetonitrile — the peaks are broader, distorted, and at shorter retention times. This example continues the pattern of Figures 1b–1d, where the peaks are broader and shorter as less water is used in the injection solvent.  

Let's think on a conceptual basis about what is going on with the various examples of Figure 1. If the sample is injected with mobile phase as the injection solvent, it does not have to undergo any chemical changes, so analyte molecules should immediately begin to interact with the stationary phase in the column and undergo normal retention behavior. However, if the injection volume is large enough, you can imagine that some band broadening will take place during the injection. For example, if the injection is extremely large, such as 1 mL, the first sample molecules will travel a significant way through the column before the last of the sample molecules arrive at the inlet. The distance between the first and last sample molecules in a given band determine the peak width, so a broad peak would be observed. At the other extreme, an infinitely small injection would have none of this injection-related broadening, so the peak would be much narrower. As we'll see later, we generally can tolerate an injection volume of approximately 15% of the peak volume of the first peak of interest before we find the injection-related band broadening objectionable. 

When the injection solvent is changed, we have another variable in addition to the injection volume mentioned above. In this case, the injected solvent plug must be diluted to the same composition as the mobile phase before "normal" retention can occur. If we think of the injection plug as a ball of solvent at the inlet of the column, it will shrink as it travels through the column, with the outside edges being diluted to the mobile-phase concentration before the center of the ball. While they are inside the ball of the injection solvent, molecules will travel down the column as if that solvent was the mobile phase, whereas those diluted into the mobile phase will now travel at the normal rate in the mobile phase. So if the injection solvent is stronger than the mobile phase, it will tend to sweep analyte molecules through the column more quickly than normal. Because some analyte molecules travel more quickly (in the middle of the injection plug) than others (on the outer edges), injection in a solvent stronger than the mobile phase tends to smear the sample along the column, causing peak distortion. On the other hand, if the injection solvent is weaker than the mobile phase, analyte molecules inside the injection plug will move more slowly than normal, which tends to concentrate them at the column inlet. This often is referred to as on-column concentration and is a technique that can be used to inject large volumes of dilute samples without excessive band broadening. And, as you might expect, the influence of the injection solvent depends on both its volume and how much its composition differs from the mobile phase. A small volume of a strong solvent will be less detrimental than a large volume of the same injection solvent, and at a small enough volume, even a very strong injection solvent will get diluted very quickly, so it will be of little consequence. 

It can be shown (see discussion of reference 2) that if you want to have <1% loss in resolution because of injection-related peak broadening, the injection volume should be <15% of the peak volume of the first peak of interest if mobile phase is used as the injection solvent. If you are willing to tolerate a 10% loss in resolution, you can increase this to <40% of the peak volume. You can determine the peak volume of a peak by measuring its baseline width (or half-height width × 1.7) and multiplying by the flow rate. Alternatively, you can get a good estimate of the peak volume with a few simple calculations that follow. We think it is a good idea to make this calculation and compare the results with the actual measurements as a double-check to make sure there isn't excessive peak broadening because of other causes, such as a bad column. 

We can use the data of Figure 1 as an example. The expected width can be estimated from the column plate number, 

 is the peak width at baseline between tangents drawn to the sides of the peak. Equation 1 can be rearranged to  

, which for real samples under practical separation conditions can be estimated as  

 is the column length (in millimeters) and 

 is the particle diameter (in micrometers). The column of Figure 1 is 250 mm long packed with 10-μm particles, so 

 ≈ 7500. Injected in the weakest solvent, water (Figure 1d), retention times are 4.72 and 9.05 min. With equation 2, this translates to expected peak widths of 0.218 and 0.418 min, respectively. (The usual disclaimer applies here: Values have been rounded for convenience, so if you try to reproduce these calculations, your results may vary somewhat.) Multiply by the flow rate (1 mL/min) to convert this to peak volumes of 218 and 415 μL. 

If we use the <15% rule mentioned at the beginning of this section, maximum recommended injection volumes in mobile phase are then 0.15 × 218 = 33 μL and 0.15 × 418 = 63 μL. If we are interested in the first peak, this will be the limiting case, so injection of 5–30 μL in mobile phase would be expected to cause less than 1% loss in resolution. If we are more liberal with the tolerance for loss of resolution, as might be justified in the case of Figure 1, where resolution >>2, the <40% rule should be adequate, or 0.4 × 218 = 87 μL. Similar calculations can be performed for the data of Table I and we would see that the <15% rule would allow approximately 45-μL injections and the <40% rule would allow approximately 120-μL injections. 

Now, the final test has to do with whether or not all these estimates line up with what is observed in reality. First, let's consider the data of Figure 1. Resolution is directly related to peak width, so a 1% or 10% loss in resolution would correspond to a 1% or 10% increase in peak width, respectively. Also, any increase in peak width should give a corresponding decrease in peak height, so we can indirectly measure any loss of resolution by a reduction in peak height. The 30-μL injections of Figure 1 are approximately the same as the injection volume estimated above for a 10% loss in resolution. You can see that the peak heights for injection of 30 μL of sample in mobile phase (Figure 1b) are 60–70% of those when 30 μL of water is injected (Figure 1d). So some broadening is occurring, as expected — partly because of the injection in mobile phase rather than a weaker solvent. Note that in all cases where the mobile phase or weaker injection solvent was used (Figures 1b–1d), the retention times were unchanged (within normally expected sample-to-sample variability). 

Contrast the results of Figures 1b–1d with those of Figure 1a. In the case of Figure 1a, too strong an injection solvent was used. This resulted in two problems. First, the peaks broadened unacceptably, especially for the first peak, which is severely distorted. Second, some of the sample molecules are swept down the column during the injection-solvent dilution process, giving shorter retention times. These problems could be solved by injecting a much smaller sample volume or using a more dilute injection solvent. 

The data of Table I show the same pattern, but less severely. This is because of at least two factors. First, the injected volumes (20 μL) are less than those of the <15% rule, and the difference between the injection solvent strength and the mobile phase is less. For Figure 1a, 30 μL of acetonitrile is injected into an 18% acetonitrile mobile phase, whereas in Table I, 20 μL of methanol is injected in a 60% or 70% methanol mobile phase. 

We have seen that the combination of injection solvent strength and injection solvent volume must be chosen carefully to avoid unwanted band broadening and the resulting loss of resolution. In the examples of Table I and Figures 1b–1d, the injection conditions were responsible for some loss in resolution (measured indirectly from loss in peak height), but no change in retention was observed. The peaks of Figure 1a, however, were distorted and lost retention because too much of too strong a solvent was injected. 

The <15% and <40% rules discussed above are guidelines that seem to work fairly well. But, as with all guidelines, it is a good idea to give the actual results a reality test. One way was discussed here, where estimates of peak volumes and injection conditions were compared with actual experimental conditions. Another technique is simply empirical — try increasing or decreasing the proposed injection volume by twofold. In the same manner, injection solvent concentrations that are closer to the mobile phase than the 100% strong or weak solvent conditions used in Table I and Figure 1a would be expected to have a less detrimental result. For example, injection in 70% methanol into a 60% methanol mobile phase would be expected to be less of a problem than an injection solvent of 100% methanol. If significant differences in the chromatograms are observed, you should adjust the conditions so that you are not operating too close to the reliability limits of the method. 

(2) L.R. Snyder, J.J. Kirkland, and J.W. Dolan, 

Introduction to Modern Liquid Chromatography, 3rd Ed.

 (Wiley, Hoboken, New Jersey, 2010), section 2.6.1. 

is a PhD student in the laboratory of Dr. Hugh Flowers at Glasgow University in Scotland.

"LC Troubleshooting" Editor John Dolan has been writing "LC Troubleshooting" for

for more than 25 years. One of the industry's most respected professionals, John is currently the Vice President of and a principal instructor for LC Resources, Walnut Creek, California. He is also a member of 

's editorial advisory board. Direct correspondence about this column via e-mail to 

A closer look at how the solvent chosen for injection can affect the appearance of chromatographic peaks.

LCGC North America-10-01-2012

Ultrahigh-Efficiency Protein Separations with Submicrometer Silica Using Slip Flow

Determination of Pesticides in Red Wine by QuEChERS Extraction, Novel Cleanup with Rapid Mini-cartridge Filtration, and LC–MS-MS Detection

Solid-Phase Microextraction

Introduction to Preparative HPLC

The Role of the Injection Solvent

Vol 30 No 10 LCGC North America October 2012 Regular Issue PDF