Journal of Chromatography A, 1383 (2015) 47–57

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Efficiency of short, small-diameter columns for reversed-phase liquid chromatography under practical operating conditions夽 Yan Ma a , Alexander W. Chassy a , Shota Miyazaki b , Masanori Motokawa b , Kei Morisato b,c , Hideyuki Uzu b , Masayoshi Ohira b , Masahiro Furuno b , Kazuki Nakanishi c , Hiroyoshi Minakuchi d , Khaled Mriziq e , Tivadar Farkas f , Oliver Fiehn a,g , Nobuo Tanaka a,b,∗ a

University of California, Davis, 451 Health Sciences Drive, Davis, CA 95616, USA GL Sciences Inc., 237-2 Sayamagahara, Iruma, Saitama 358-0032, Japan c Kyoto University, Department of Chemistry, Kitashirakawa, Sakyo-ku, Kyoto 606-8502, Japan d Kyoto Monotech Co., 1095 Shuzei-cho, Kamigyo-ku, Kyoto 602-8155, Japan e AB SCIEX, 1201 Radio Road, Redwood City, CA 94065, USA f Phenomenex, Inc., 411 Madrid Avenue, Torrance, CA 90501, USA g University of Jeddah, Biochemistry Department, Saudi Arabia b

a r t i c l e

i n f o

Article history: Received 9 October 2014 Received in revised form 5 January 2015 Accepted 7 January 2015 Available online 14 January 2015 Keywords: Column efficiency 1 mm I.D. column Monolithic silica Core–shell particles Extra-column effect Weak-wash solvent

a b s t r a c t Prototype small-size (1.0 mm I.D., 5 cm long) columns for reversed-phase HPLC were evaluated in relation to instrument requirements. The performance of three types of columns, monolithic silica and particulate silica (2 ␮m, totally porous and 2.6 ␮m, core–shell particles) was studied in the presence of considerable or minimal extra-column effects, while the detector contribution to band broadening was minimized by employing a small size UV-detector cell (6- or 90 nL). A micro-LC instrument having small system volume ( 1–3 (depending on the column), in acetonitrile/water mobile phase (65/35 = vol/vol) at 0.05 mL/min, with the instrument specified above. The column efficiency was lower by up to 30% than that observed with a 2.1 mm I.D. commercial column. The small-size columns also provided 8000–8500 theoretical plates for well retained solutes with a commercial ultrahigh-pressure liquid chromatography (UHPLC) instrument when extra-column contributions were minimized. While a significant extra-column effect was observed for early eluting solutes (k < 2–4, depending on column) with methanol/water (20/80 = vol/vol) as weak-wash solvent, the use of methanol/water = 50/50 as wash solvent affected the column efficiency for most analytes. The results suggest that the band compression effect by the weak-wash solvent associated with partial-loop injection may provide a practical means to reducing the extra-column effect for small-size columns, while the use of an instrument with minimum extra-column effect is highly desirable. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Small-size columns for high performance liquid chromatography (HPLC) attracted considerable attention in 1970s and 1980s [1–13]. Some groups attempted column miniaturization by adopting a capillary format [1,3–5], while others tried to reduce the

夽 Presented at the 30th International Symposium on Chromatography (ISC 2014), Salzburg, Austria, 14 - 18 September 2014. ∗ Corresponding author at: GL Sciences Inc., Iruma, Saitama 358-0032, Japan. Tel.: +81 4 2934 2123; fax: +81 4 2934 3412. E-mail address: [email protected] (N. Tanaka). http://dx.doi.org/10.1016/j.chroma.2015.01.014 0021-9673/© 2015 Elsevier B.V. All rights reserved.

size of conventional columns. The importance of minimizing extra-column band broadening was described in the reports. Because of the practical limitations in minimizing the instrumental contribution to analyte band broadening for small-volume columns, and because low-dispersion liquid chromatographs were not commercially available, it was more advantageous to explore the performance of small-diameter columns in longer formats leading to extensive research aiming at generating ultrahigh numbers of theoretical plates in the early stage of size reduction of packing materials [4–6,8,9]. During the past decades, the dimensions of high-speed, highefficiency columns have been reduced from 4 to 4.6 mm I.D. and 15–30 cm length to 2.1 mm I.D. and 5–10 cm length. This reduction

48

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

in column dimensions was made possible by the development of advanced packing materials (sub-2 ␮m totally porous particles and sub-3 ␮m core–shell particles) that maintained high plate counts and allowed for high-speed separations. All along, the difficulties associated with the use of small-size columns in HPLC in providing high column efficiency have been recognized and emphasized [14–20]. The reduction of extra-column effects has become even more desirable than before, actually a necessity, because solute bands eluted from small-size columns packed with very small, high-efficiency materials could be less than a few ␮L in volume, corresponding to a band variance less than a few ␮L2 [14–19]. Therefore, peaks eluted from small-size columns are very sensitive to extra-column band broadening. The reduction in column diameter from 4.6 mm to 3 mm, and from 3 mm to 2.1 mm is accompanied by a reduction in cross section by a factor of 2.35 and 2.04, respectively. However, a further reduction from 2.1 to 1 mm I.D. results in a more significant decrease in cross sectional area by a factor of 4.41. While most UHPLC instruments with system-caused band variance of several ␮L2 allow for adequate operation of short, 2.1 mm I.D. columns maintaining high chromatographic efficiency, such instruments are inadequate for the operation of short 1 mm I.D. columns [15,16]. In spite of this severe practical limitation, analysts are still interested in using columns of very small volume due to their potential in terms of significant reduction in solvent consumption, reduction in sample size or improvement in detection sensitivity both with UV/vis and MS detectors, along with effective dissipation of the heat generated by the high velocity mobile phase percolating through such columns at high linear velocity. This study on the performance of small volume columns had two objectives: first, to evaluate the current limits in chromatographic efficiency of such columns, and second, to analyze the major factors imposing these limits. In these terms, the intrinsic efficiency of small volume columns is of interest. Can such columns be prepared today to provide the same performance as larger size columns (which are less affected by the current instrument limitations)? Should the focus be on making better columns or have they already reached their limit of performance? Relatively low column efficiency has been reported for a column of small dimensions [15,16]. In early days, electrochemical detection was sometimes employed to reduce extra-column volume (Vextra ) and extra-column band 2 variance (extra ) when using columns of small size [10,12]. The use of popular and widely applicable UV detection is of much practical importance with small-size columns, although sensitivity can be an issue [1,4]. The studies referenced above reported observed column 2 efficiency (Nobserved in Eq. (1), where VR and total stand for retention volume and band variance of a solute observed, respectively) (as opposed to true column efficiency) while recognizing the neg2 ative effect of extra-column band variance (extra in Eq. (2), where 2 describes band variance acquired in a column) and the need column for reducing it. Nobserved =

VR2 2 total

=

[V0 (1 + k)]

2 2 2 total = column + extra

Nintrinsic =

VR2 2 column

2 total

measurement of extra-column effect [19,20]. Clearly, the current limitations in HPLC system performance need to be overcome by reducing both system volume and the extra-column variance associated with it. This need is very much the focus of HPLC instrument manufacturers. In the meantime, analysts have only limited choices to improve the performance of a column-instrument combination. Numerous reports have proposed practical solutions to reducing extra-column variance such as shortening connecting tubing, using small volume injectors and/or detector cells, and implementing oncolumn detection. Alternatively, injection techniques associated with compression of the injected band taking place in weak sample solvent have been reported [21–24]. In this report the performance of three prototype columns of 1 mm I.D., 5 cm length packed with sub-2 ␮m particles, 2.6 ␮m core–shell particles, or monolithic silica, will be discussed along with the practical evaluation conditions. 2. Experimental 2.1. Instrument and columns An Eksigent ExpressLC® -Ultra (AB SCIEX, Redwood City, CA, USA) was equipped with a 0.2 ␮L sample loop and a UV detector with a chip cell (volume: 90 nL, path length: 5 mm). An Acquity UPLC (Waters, Milford, MA, USA) with a 2 ␮L loop was used in combination with a UV detector (MU701, GL Sciences, Tokyo, Japan) having a capillary cell (volume: 6 nL, path length: 3 mm). Three prototype columns of 1 mm I.D., 5 cm length were examined. They were packed with (1) Kinetex C18, 2.6 ␮m core–shell particles (Phenomenex, Torrance, CA, USA), (2) InertSustain C18, 2 ␮m particles (GL Sciences), and (3) glass-clad monolithic silica column modified with dimethyloctadecylsilyl moieties (MonoTower, GL Sciences). The monolithic silica was prepared and glass-clad according to the procedure described previously [25]. Two columns were supplied for each particulate packing material which showed very similar results in a preliminary test. 2.2. Chemicals Acetonitrile of LC–MS grade was purchased from Fisher Scientific (Hampton, NH, USA). Water was purified with a Milli-Q Quantum EX from Millipore (Bedford, MA, USA). Reversed-phase column performance test mixture containing (1) acetanilide, (2) acetophenone, (3) propiophenone, (4) butyrophenone, (5) benzophenone, (6) valerophenone, (7) hexanophenne, (8) heptanophenone, and (9) octanophenone at 100 ␮g/mL each, in acetonitrile/water = 65/35 was obtained from Agilent (Little Fall, DE, USA). A triazine pesticides standard mixture containing ametryn, atrazine, prometon, prometryn, propazine, simazine, and terbutryn in methanol at a concentration of 100 ␮g/mL (Supelco, Bellefonte, PA, USA) was also employed. Thiourea and uracil were purchased from Sigma–Aldrich (St. Louis, MO, USA).

2

(1) (2) (3)

Intrinsic column efficiency (Nintrinsic in Eq. (3)) can be estimated based on observed efficiency and extra-column variance following Eqs. (1)–(3), in principle. Extra-column band broadening has been studied based on the peak variance observed when the chromatographic column is replaced with a zero-dead-volume union. A noninvasive method for the determination of extra-column variance and intrinsic plate height has been proposed for the

2.3. Chromatographic measurements Columns were connected to the injector and to the detector of Express-LC with 0.025 mm I.D., 10 cm long PEEKSil tubing (SGE, Austin, TX, USA). In the case of the Acquity instrument, columns were connected either directly to the injector, or to a column port via a Viper tube (0.13 mm I.D., 6.5 cm, Thermo Fisher Scientific, Germering, Germany). In the latter case, a heat exchange tube was also a part of the chromatographic system. On the detector side PEEK tubing of 0.0635 mm I.D., 20–90 cm length (IDEX Health & Science, Oak Harbor, WA, USA), was used to connect the column to the detector. The longer PEEK tubing was necessary to suppress air gaps from reaching the UV detector cell when partial-loop injection was applied in the absence of a column with the Acquity instrument.

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

Extra-column volume and band variance values were measured in the presence of a zero-dead-volume union (ZDU) (GL Sciences) replacing the column. Premixed mobile phase, acetonitrile/water = 65/35 made up by volume, was used in the flow rate range 0.02–0.15 mL/min with the Acquity LC. The ExpressLC was operated at 20–50 ␮L/min. Partial-loop injection of 0.1 ␮L was employed unless noted otherwise. The Acquity instrument was operated with various weak-wash solvents such as methanol/water = 50/50, 20/80, or acetonitrile/water = 65/35 to examine the effect of the coexisting solvent in the sample loop of the injector. With Express-LC Ultra, 0.125 ␮L injection was used at a flow rate of 50 ␮L/min, because the instrument was unable to inject a sample volume corresponding to less than 1/400 min using metering injection (this amounts to 0.125 ␮L). All measurements were carried out in duplicate with the data-acquisition frequency set to 60 Hz. 3. Results and discussion 3.1. Extra-column effects The extra-column volume (Vextra ) and extra-column band variance values are known to be determined by the configuration of the HPLC instrument, while band variance is also flow rate dependent. These values were measured for the two instruments used in this work (a micro-LC and a UHPLC) in the flow rate range relevant to 1 mm I.D. columns. The results are listed in Table 1. The instrumentinduced band variance and the extra-column volume were calculated by a moment-based method for both instruments configured for minimum extra-column volume (Express LC Ultra and Acquity UPLC) with a detector connected directly to the injector with a zerodead-volume union (ZDU). In contrast, variance values were estimated based on the peak width at half height when the detector was connected to the column port of the Acquity instrument since symmetrical peaks were obtained, as shown in Fig. 1. Both the detector cell made of 0.05 mm I.D. fused silica tubing and the 0.15 mm diameter channel chip cell allowed for good peak symmetry. The micro-LC instrument Express-LC Ultra, designed for the operation of 0.2–1.0 mm I.D. columns, has an extra-column volume of less than 1 ␮L, and provides very small band dispersion, calculated to be less than 0.02 ␮L2 , although the maximum flow rate was limited to 0.05 mL/min. Greater extra-column volume and considerable extra-column band variance values were observed with a UHPLC instrument, Acquity UPLC, especially when using the column port. The extra-column band variance values obtained with the Acquity were similar to those reported previously [19], indicating the significant effect of the heat exchanger placed between the injector and the column port, which has a volume of about 10–11 ␮L. This tubing seems to be necessary to achieve complete mixing of the solvents supplied by the two pump heads. Because of this relatively large volume, a small sample plug (only a fraction of 1 ␮L) will experience significant dispersion, reaching the column inlet as a wide band. The instrument configuration optimized for higher efficiency obtained by bypassing the heat exchanger, or direct attachment of the column to the injector with the Viper tube, could produce incomplete mixing of the mobile phase components delivered by the two pump heads, which in turn can cause peak distortion or split peaks not only in gradient elution, but also in isocratic elution if the mobile phase components are mixed at high pressure. The uncertainty associated with measuring extra-column band variance and the true performance of a small-sized column has been previously recognized and discussed recently [19,23,24,26]. Ideally, column efficiency should be measured with an instrument inducing negligible extra-column band broadening, if such an instrument is available. Unfortunately, as mentioned later, it seems difficult to

49

obtain unambiguous extra-column band variance contributions for a column having the dimensions discussed in this work because of the complexity of the injection process employed. Column band variance values calculated for peaks eluting from small columns that provide 9000 theoretical plates are shown in Table 2 for various analyte retention factors (two discrete values were assumed for column porosity, as listed). By comparing the band dispersion values expected for columns of small dimensions (e.g. 1 or 2.1 mm I.D. and 5 cm length) with the extra-column band variance values reported in Table 1 for the micro-LC and UHPLC, the critical importance of reducing extra-column volume and band variance for the operation of a short 1 mm I.D. column becomes evident. The UHPLC instrument used in its optimized configuration could reduce the efficiency value of a 1 mmID, 5 cm long column by ca. 40–50% for peaks eluting with k = 2–4 within the flow rate range studied. The use of the column port could reduce the efficiency by ca. 50% even for well-retained analytes (k = 4–8), while the micro-LC would preserve more than 90% of column efficiency even for early eluting analytes (in the region k = 1). In contrast, a larger volume column (such as 2.1 mm I.D., 5 cm length) would perform far better on both instruments, given that the relative magnitude between the expected (i.e. calculated) band variance shown in Table 2 and the observed extra-column band variance shown in Table 1 is more favorable in any of the configurations described above as long as analyte k > 4. 3.2. Column efficiency observed with an instrument with minimum extra-column band broadening The three columns showed similar efficiencies of about 8500 theoretical plates, especially for the late eluting solutes present in the reversed-phase check-out sample, when operated with the ExpressLC-Ultra instrument that was specifically designed for the operation of small-size columns with its 90 nL detection cell. (See Fig. 2, where the numbers of theoretical plates are provided for peaks, 2, 6, and 8 for InertSustain C18 and for peaks, 3, 6, and 9 for monolithic silica C18 and Kinetex C18 in order to compare column efficiency values for the same solutes and for solutes of closer retention factors as well.) In spite of the use of the same instrument and the similar column efficiency observed for late-eluting solutes, a significant difference in the number of theoretical plates was found among these columns for early eluting solutes, particularly for uracil representing an unretained solute band. This observation needs attention. The intrinsic column efficiency of the three columns seems to be similar, as consistently observed for late eluting solutes. Band variance values should be proportional to the retention volume squared, if constant column efficiency is assumed for the entire chromatogram. The addition of extra-column band variance, which is independent of the peak volume and only function of instrument configuration and flow rate, to the intrinsic column-generated variance of each peak will result in a greater effect on the number of theoretical plates for early eluting solutes which are eluted as small-volume bands. Because the elution volume of early eluting peaks, particularly the peak at t0 , is dictated by the volume of mobile phase present in a column, or to the total column porosity, the monolithic silica column, N2595, with the highest porosity among the three columns should result in the greatest band volume for the t0 peak, thus be least affected by the extra-column variance. As a matter of fact, the porosity of the columns studied here is in the order monolithic silica C18, N2595 > Kinetex C18 > InertSustain C18. Indeed, the observed column efficiency values follow this same order for the t0 peaks. The observed extra-column band variance values on the Express LC Ultra, at the flow rate of 0.05 mL/min, are ca. 25–40% of the calculated band variance values for the t0 peak on a column expected

50

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

Table 1 Extra-column volume and band variance values observed with the UHPLC and micro-LC instruments at various mobile phase flow rates and connection configurations. Flow rate (mL/min)

Acquity

Express LC Ultra

Injector-column (1)a

0.020 0.030 0.050 0.075 0.100 0.125 0.150

Injector-column (2)b

Column port (1)c

Column port (2)d

Injector-Columne

Vextra (␮L)

2 extra (␮L2 )

Vextra (␮L)

2 extra (␮L2 )

Vextra (␮L)

2 extra (␮L2 )

Vextra (␮L)

2 extra (␮L2 )

Vextra (␮L)

2 extra (␮L2 )

3.9 4.0 4.3 ND ND ND ND

0.44 0.49 0.62 ND ND ND ND

6.3 6.4 6.7 6.9 7.2 7.5 7.8

0.42 0.53 0.72 0.82 1.04 1.16 1.25

14.2 14.4 14.7 14.9 15.1 15.3 15.5

1.1 1.3 2.1 2.6 3.2 3.7 4.0

16.8 17.0 17.0 17.4 17.7 17.8 18.1

1.2 1.6 2.4 3.2 3.7 4.2 4.6

0.50 0.60 0.78 ND ND ND ND

0.012 0.014 0.020 ND ND ND ND

The acquisition time delay was not considered for Vextra values. a Injector-0.13 mm ID Viper tube (6.5 cm)-ZDU-0.0635 mmID PEEK tube (20 cm). Calculated by moment method. b Injector-0.13 mm ID Viper tube (6.5 cm)-ZDU-0.0635 mmID PEEK tube (90 cm). Calculated by moment method. c Column port-0.13 mm ID Viper tube (6.5 cm)-ZDU-0.0635 mmID PEEK tube (20 cm). d Column port-0.13 mm ID Viper tube (6.5 cm)-ZDU-0.0635 mmID PEEK tube (90 cm). e Injector-0.05 mm ID PEEKsil tube (10 cm)-ZDU-0.05 mmID PEEKsil tube (10 cm). Calculated by moment method.

Fig. 1. Peaks observed by injecting 0.1 ␮L of thiourea solution (0.05 mg/mL in 60% acetonitrile) at 50 ␮L/min. Instrument and connections before and after column: (a) Express LC Ultra, two PEEKsil tubing (0.025 mm I.D., 10 cm), with time scale expanded by five times, (b) Acquity (i), injector-Viper (0.13 mm I.D., 6.5 cm)-ZDU-PEEK (0.0635 mm I.D., 20 cm), (c) Acquity (ii), injector-Viper (0.13 mm I.D., 6.5 cm)-ZDU-PEEK (0.0635 mm I.D., 90 cm), (d) Acquity (iii), column port-Viper (0.13 mm I.D., 6.5 cm)-ZDU-PEEK (0.0635 mm I.D., 20 cm). (The difference in peak intensity between Fig. 1a and b–d, also in Figures appearing later, is due to the difference in detector sensitivity.).

to produce 9000 plates (i.e. having an intrinsic plate count of 9000). The observed band variance values of each t0 peak were 0.09, 0.18, and 0.10 ␮L2 for Kinetex C18, InertSustain C18, and monolithic silica C18 (N2595), respectively. By correcting the observed band variance value of each non-retained peak eluting off the Kinetex C18 and InertSustain C18 columns for extra-column variance (0.02 ␮L2 ), the resulting column efficiency for these peaks is still

lower than that for late eluting peaks. In other words, the extracolumn effect cannot fully explain the loss and relative differences in efficiency for their t0 peaks between the three columns. Similarly, differences were observed in net efficiency values for other early eluting solutes, when the extra-column band variance values were subtracted from the observed variance values. The number of theoretical plates observed for the non-retained and

Table 2 Band variance values calculated for columns of 1 and 2.1 mm I.D. and for solutes eluting with k = 0–10. Two different porosity values were considered. Band variance, ␴2 column (␮L2 )

Column

Retention factor, k Column diameter (mm)

Column length (cm)

Total porosity

Column dead volume (␮L)

Number of theoretical plates

0

1

2

4

6

8

10

1.0 1.0 2.1 2.1

5 5 5 5

0.55 0.67 0.55 0.67

21.6 26.3 95.2 116.0

9000 9000 9000 9000

0.05 0.08 1.01 1.49

0.21 0.31 4.03 5.98

0.47 0.69 9.06 13.45

1.29 1.92 25.2 37.4

2.54 3.77 49.3 73.2

4.19 6.22 81.6 121.0

6.27 9.30 121.8 180.8

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

51

Fig. 2. Chromatograms obtained with ExpressLC-Ultra. Column: (a and b) Monolithic silica C18 N2595, (c and d) Kinetex C18, (e and f) InertSustain C18. Mobile phase: acetonitrile/water = 65/35. Flow rate: 50 ␮L/min. Sample: (a, c, and e) uracil (0.125 ␮L solution of 0.16 mg/mL in 65% acetonitrile), and (b, d, and f) reversed-phase check-out sample (0.125 ␮L in 65% acetonitrile). See experimental section for peak identity. The numbers of theoretical plates (N) and retention factor (k) are indicated for uracil (t0 ) for all columns, for propiophenone (3), valerophenone (6), and octanophenone (9) for Monolithic silica C18 and Kinetex C18, and for acetophenone (2), 6, and heptanophenone (8) for InertSustain C18 after correction for the extra-column volume.

for the early eluting peaks after correction for instrument-induced band dispersion may very well reflect the intrinsic column efficiency for each case (of column and analyte pair) because the results were obtained using the same conditions for all three columns. It seems rational to suggest that these differences in intrinsic column efficiency found for analytes of various retention factors may be due to some previously overlooked column feature such as non-optimal structure of its inlet and outlet and/or its frits. The importance of the structure of column inlet has been reported [27]. The prototype 1 mmI.D. InertSustain column had 1.575 mm diameter frits (inlet and outlet) that were too large for the diameter of the column. Such unfavorable column feature may bring about its own contribution to band broadening, and should contribute to the same extent to the widths of all bands exiting the column. Such a contribution could not be accounted for based on what we could call as net extra-column effect, but will still affect the efficiency of a column in a similar way as the extra-column effect. High efficiency for early eluting solutes can be a feature of monolithic silica columns that may be free of structural complexity at the column end during preparation. (See further discussion on column efficiency in Sections 3.4 and 3.5.) These results imply that the efficiency of small-size columns may not be accurately estimated by only taking into account the extra-column effect assuming constant intrinsic plate height values for all solutes. In the case of 2.1 mm I.D. columns, the expected band variance for the t0 peak is ca. 20 times greater (assuming the same porosity as for 1 mm I.D. column), than that of 1 mm I.D. columns at similar column efficiency, making this effect much less significant.

column port. Such an operation mode is known to produce higher column efficiency for small-size columns than the case where the column is connected to a column port [14]. The extra-column volume taken into account in the calculation of the number of theoretical plates was of several ␮L (see Table 1). Peaks eluting late (with k > 5) showed column efficiency values of about 8000–9000 theoretical plates similar to those obtained with ExpressLC. However, the t0 peak as well as the early eluting peaks (k < 0.7) showed severe deformation or broadening. The decrease in the number of theoretical plates observed with all peaks eluting with k < 3 was more obvious than that in Fig. 2, reflecting the significantly greater contribution of extra-column effect with the Acquity system. Consistent column efficiency values observed with solutes with k > 3 imply that the plate heights obtainable with these columns with the instrument configuration employed are ca. 5.5–6 ␮m, comparable with those obtained with the Express LC Ultra (which was found to have significantly smaller extra-column volume). The extra-column band variance experimentally obtained (Table 1) and band variance values calculated for a column producing 9000 theoretical plates (Table 2) predict that the column efficiency of such a column can decrease by about 50–25% for solutes with k = 2–4 at 0.05 mL/min on the Acquity UPLC instrument. The actual decrease in the number of theoretical plates for the peaks in Fig. 3 was found to be much less than predicted from these values. This difference may be explained based on the belief that using the extra-column variance values generated per Section 3.1 may be overcompensating for extra-column effects. The overcompensation could also be due to the difference in the method of calculation of band variance for the extra-column peak and retained peaks [28].

3.3. Column efficiency observed with the UHPLC instrument 3.3.1. Direct attachment of a column to the injector Fig. 3 shows chromatograms obtained with Acquity UPLC with a column connected directly to the injector using a 0.13 mm I.D., 6.5 cm long Viper tube, bypassing the tube from the injector to the

3.3.2. Column efficiency observed with a UHPLC instrument used in its standard configuration Fig. 4 shows chromatograms with the column connected to the column port of an Acquity UPLC. Large extra-column band variance (associated with broad extra-column peaks, as described

52

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

Fig. 3. Chromatograms obtained with Acquity UPLC with 1 mm I.D., 5 cm long columns connected directly to the injector. Column: (a and b) Monolithic silica C18 N2595, (c and d) Kinetex C18, (e,f) InertSustain C18. Mobile phase: acetonitrile/water = 60/40. Flow rate: (a–d) 50 ␮L/min, (e and f) 75 ␮L/min. Sample: (a, c, and e) 0.1 ␮L of thiourea solution (0.05 mg/mL in 60% acetonitrile), and (b, d, and f) 0.1 ␮L of reversed-phase check-out sample in 65% acetonitrile. See Fig. 2 for peak annotation.

in Section 3.1) resulted in much smaller numbers of theoretical plates for early-eluting solutes compared to the late-eluting ones. In the calculation of the number of theoretical plates and retention factors with this set up, it was necessary to take into account Vextra values that accounted for ca. 50% of the elution volume of the unretained solute. The observed column efficiency was lower than that indicated in Fig. 3, and also than that observed with the micro-LC equipment shown in Fig. 2. Loss in column efficiency was

more severe for the early eluting peaks shown in Fig. 4 than in Figs. 2 and 3, which was presumably caused by a greater extracolumn effect. The loss in column efficiency for early eluting solutes of k up to about 5 is clearly visible in Fig. 4, while late eluting solutes showed 15–20% lower column efficiency than what was obtained with the instrument generating smaller extra-column effect. Consistent column efficiency observed with solutes with k > 5 indicates plate

Fig. 4. Chromatograms obtained with Acquity UPLC with columns connected to the column port. Column: (a and b) Monolithic silica C18 N2595, (c and d) Kinetex C18, (e and f) InertSustain C18. Mobile phase: acetonitrile/water = 65/35. Flow rate: (a–d) 50 ␮L/min, (e and f) 75 ␮L/min. Sample: (a, c, and e) 0.1 ␮L of thiourea solution (0.05 mg/mL in 60% acetonitrile), and (b, d, and f) 0.1 ␮L of reversed-phase check-out sample in 65% acetonitrile. See Fig. 2 for peak annotation.

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

height values of ca. 6–7 ␮m for the solutes studied, clearly larger than those obtained with Express LC Ultra or with Acquity with the column connected directly to the injector. The extra-column band variance experimentally obtained (Table 1) and band variance values calculated for a column producing 9000 theoretical plates (Table 2) predict that the efficiency of such a column can decrease by 80, 60, and 30% for solutes with k = 2, 4, and 8, respectively, at 0.05 mL/min. The actual decrease in the number of theoretical plates for the peaks in Fig. 4 was found to be much less than predicted. (See Sections 3.4 and 3.5 for further discussion of column efficiency.) Another interesting observation based on Figs. 3 and 4 is that some solute bands (peaks 2 and 3 in Fig. 3, and peaks 1–6 in Fig. 4) are narrower than the analyte band reaching the column inlet (“extra-column band” shown in Fig. 1c and d, respectively). Smaller band width of some early eluting peaks than that of the extra-column band observed in Figs. 3 and 4, and smaller decrease in column efficiency than expected from Tables 1 and 2 could be the result of band compression (i.e. analyte refocusing due to retention on the sorbent surface) induced by the weak-wash solvent (50% methanol) co-existing in the sample loop with the sample plug whenever partial-loop injection is performed. The loop volume, 2 ␮L, is fairly large compared to injected sample volume (0.1 ␮L), leaving some weak-wash solvent in the loop. The volume of this weak-wash solvent, can be equivalent to ca. 4–5% of column void volume. Therefore, with only 0.1 ␮L sample, the sample introduction process at the column head could be considerably different from the case with a larger-sized column or with full-loop injection. The results will be further discussed below.

3.3.3. Effect of wash solvent on column efficiency with partial loop injection. A sample injected in a weak solvent is expected to lead to band compression, as a function of the retention factor of each analyte [21]. Successful enhancement of this phenomenon has been reported in reversed-phase mode by co-injecting weak-wash solvent along with the sample plug [22–24]. In the present study, partial-loop injection was applied to carry out small-volume injection (0.1 ␮L) using a 2 ␮L loop. In such a case, about 1 ␮L of weak-wash solvent remains in the sample loop together with the sample plug and pre- and post-sample air gaps (0.5 ␮L each). In Figs. 3 and 4, several early eluting peaks showed narrower band width than the extra-column band in the same mobile phase. Similar results were obtained with triazines, as shown in Fig. S1 (see Supplementary Material). In both cases, the sample was dissolved in a solvent having either the same composition as or higher eluting strength than the mobile phase. Therefore, significant band compression due to weak sample solvent cannot be expected. The positive effect of the weak-wash solvent on column efficiency for partial-loop injection can be clearly seen in Fig. 5. When acetonitrile/water = 65/35 (of the same eluting strength as the mobile phase) was used as weak-wash solvent, instead of methanol/water = 50/50, significantly lower column efficiency was observed for most peaks. In contrast, an increase in column efficiency was observed with methanol/water = 20/80 as weak-wash solvent. Retention factors of acetophenone and propiophenone, the second and third peaks observed for the check-out sample in 20% and 50% methanol, and in 65% acetonitrile as mobile phase are listed in Table 3. The retention factors of the first two phenones in 20% and 50% methanol were 30–50 times and 3–5 times larger, respectively, than those in 65% acetonitrile. Considerable band compression can be expected with such weak-wash solvent upon injection [22,23], if one considers that the volume of weak-wash solvent, ca. 1 ␮L, in a loop accounts for about 4–5% of column void volume, although

53

Fig. 5. Chromatograms obtained with Acquity UPLC with the column connected to the column port and with weak-wash solvents of various compositions. Column: Kinetex C18. Mobile phase: acetonitrile/water = 65/35. Flow rate: 50 ␮L/min. Weakwash solvent: (a) 65% acetonitrile, (b) 50% methanol, and (c) 20% methanol. Sample: 0.1 ␮L of reversed-phase check-out sample in 65% acetonitrile containing uracil at 0.16 mg/mL. See experimental section for peak identity.

the sample band would not be evenly distributed along the sample plug provided from the sample loop. Actually, the resulting volume of a sample band at the inlet of the column seems to be much greater with Acquity, espe2 values listed in cially when the column port was used, if extra Table 1 are considered to be the band variance at the column inlet. (Note that the detector contribution to band broadening is very small in the present study.) In other words, sample injection with weak solvent was conveniently carried out by employing partial loop injection which essentially attained the condition for performance-optimizing-injection sequence (POISE) [22–24]. Thus under practical conditions using partial-loop injection in this work, an attempt to obtain intrinsic column efficiency by subtracting the extra-column band variance value from the observed band variance was not made. The subject needs further study, because this kind of sample introduction method may be able to get around the problem of large extra-column effect associated with the use of small-size columns with common UHPLC equipment, although limitations of such an approach were seen with the disturbed peaks near the t0 peak and the decrease in peak intensity for late eluting solutes with the weak solvent for wash. 3.4. Comparison of column efficiency based on van Deemter plots Fig. 6a–d shows plots of plate height (H) observed against linear velocity of mobile phase (u) for monolithic silica C18 and Kinetex C18 under the conditions shown in Figs. 2 and 3, for selected solutes, namely C2 H5 COPh, C4 H9 COPh, and C7 H15 COPh. Results with CH3 COPh, C4 H9 COPh, and C6 H13 COPh are shown for InertSustain C18. The same solute and also solutes having similar retention factors were included for the plots for InertSustain C18 as for the other columns. The maximum number of theoretical plates obtained for the two particulate columns, ca. 8500–8800 (corresponding to a plate height of 5.7–5.9 ␮m) is lower by at least 10–15% than what was previously reported for a 2.1 mm I.D. column [22,29] or measured in this study in spite of the fact that the columns were packed with particles of the same properties according to their manufacturers. (See Fig. S2 of supplementary materials for the results obtained with a 2.1 mmI.D. column.) The plate height values correspond to reduced plate height (H/domain size, or H/particle diameter) of

54

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

Table 3 Retention factors and Rf values of alkyl phenyl ketones in 20% and 50% methanol, or 65% acetonitrile (weak-wash solvent had the same composition as the mobile phase). Column

Mobile phase 65% CH3 CN

InertSustain C18 Kinetex C18 N2595

50% CH3 OH

20% CH3 OH

k1

1/(1 + k1)

k2

1/(1 + k2)

k1

1/(1 + k1)

k2

1/(1 + k2)

k1

1/(1 + k1)

k2

1/(1+k2)

1.37 0.62 0.41

0.42 0.62 0.71

2.19 1.00 0.67

0.31 0.50 0.60

3.90 1.70 1.63

0.20 0.37 0.38

8.48 3.66 3.51

0.11 0.21 0.22

32.8 17.0 15.1

0.030 0.055 0.062

100.4 51.5 45.3

0.010 0.019 0.022

The k1 and k2 stand for the retention factor of Cn H2n+1 COPh, n = 1 and n = 2, respectively.

Fig. 6. The van Deemter plots obtained with Express LC Ultra (a, c, and e) and Acquity UPLC with the column connected directly to the injector (b, d, and f). Column: (a and b) Monolithic silica N2595, (c and d) Kinetex C18, (e and f) InertSustain C18. Mobile phase: (a, c, and e) acetonitrile/water = 65/35, (b, d, and f) acetonitrile/water = 60/40. Solutes: () propiophenone, () valerophenone, () octanophenone. (Results for acetophenone and heptanophenone were employed instead of those for propiophenone and octanophenone for InertSustain C18 column to show results for solutes having closer k values.).

about 3.2, 2.2, and 2.8 for monotlithic column, Kinetex, and InertSustain. The columns of 1 mm I.D., 5 cm length might not have been prepared so well as the larger-sized columns. As a matter of fact, the permeability values of the 1 mmI.D. particulate columns were found to be greater than that of a 2.1 mm I.D. column by ca. 30% indicating lower density of the packed bed. The plate height observed with the monolithic silica column at optimum linear velocity is comparable with that of Kinetex C18 or InertSustain C18 in Fig. 6b, d and f but higher than that observed with a larger-sized monolithic silica column (see Fig. S3 of supplementary materials for the results obtained with a 2.1 mm I.D. column). The increase in plate height at high linear velocity is more pronounced for this column, and the optimum linear velocity is lower than that observed for a monolithic silica or for a particulate column of 2.1 mm I.D. or of similar diameter, especially for early eluting solutes.

The shift in optimum linear velocity to lower values can be attributed to a greater extra-column effect and a larger interstitial void size, or to poorer packing in the case of a particulate column, that result in the greater contribution of mass-transfer to band broadening in the van Deemter equation. The rise in plate height values for poorly retained solutes indicates the significant contribution of extra-column band broadening on Acquity UPLC. The results obtained with the Express-LC Ultra having much smaller systemgenerated band variance, by showing good agreement in the van Deemter plots between early and late eluting analytes for N2595 and Kinetex (Fig. 6a and c), seem to support this interpretation. In the case of InertSustain C18, the results suggest that the deviation was caused by the aforementioned column hardware- or bed structure-related factor, because the deviation for the early eluting bands was not improved by the reduced extra-column effect on Express LC.

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

55

Fig. 7. The van Deemter plots obtained with Acquity with a column connected to the column port. Mobile phase: acetonitrile/water = 65/35. Solutes as indicated in Fig. 6. Column: (a and b) Monolithic silica N2595, (c and d) Kinetex C18, (e and f) InertSustain C18. Weak-wash solvent: (a, c, and e) 50% methanol (methanol/water = 50/50), and (b, d, and f) 20% methanol (methanol/water = 20/80).

Fig. 7 shows van Deemter plots obtained with the Acquity UPLC with a column connected to the column port with two different weak-wash solvents: 50% methanol and 20% methanol. Comparison of the van Deemter plots in Fig. 7 with those in Fig. 6 indicates that (i) early-eluting solutes are severely affected by the system configuration inducing severe extra-column band broadening, (ii) the extra-column effect extends to all peaks included, (iii) optimum linear velocity observed shifted to low flow rate range, especially for early eluting solutes, and most importantly (iv) the weak-wash solvent improves column efficiency in the partial-loop injection mode. In spite of the significant impact of extra-column effect on column efficiency, the use of common UHPLC equipment might be still possible for 1 mm I.D. columns, though prone to the limitations mentioned earlier, depending on the range of retention factors and column efficiency required for the particular application. The reduction in the number of theoretical plates was ca. 50% and 10–15% for solutes with k = 1.5 and k > 5, respectively, when a weak-wash solvent of adequate (i.e. low enough) eluting strength was used. Whenever column efficiency is of primary concern, the use of micro-LC equipment is preferred with 1 mm I.D., 5 cm long columns, while the UHPLC system can be used with columns of this size only when all peaks of interest are well retained (k > 5). Table 4 summarizes the properties of the 1 mm I.D., 5 cm long columns examined in this work. The monolithic silica column provided column efficiency similar to that of 2 ␮m totally porous particles or 2.6 ␮m core–shell particles with permeability similar to that of columns made with ca. 3.5–4 ␮m particles. The separation impedance of monolithic silica was much lower compared to particulate columns as a consequence of its much greater permeability. The hydrophobic retentivity of the monolithic silica C18 column was slightly higher compared to the particulate columns. The retention factor of butyl phenyl ketone on monolithic silica C18 was only 2/3 of that observed on Kinetex, while retention times of

this solute on the two columns are different only 10% due to the larger porosity, or greater t0 value of the monolithic column. 3.5. Relation between height equivalent to a theoretical plate (H) and retention factor (k) It is known that the band width of a peak eluted with a greater retention factor (k) is accompanied by a relatively smaller contribution of the extra-column band broadening to total peak variance or experimentally observed plate height. Fig. 8 provides information on variation in plate height in relation to retention factor and the instrument used. ExpressLC Ultra provided column efficiency least affected by extra-column band broadening showing consistent H for solutes with k > 1 for N2595, k > 2 for Kinetex, and k > 3 for InertSustain at 0.05 mL/min, as shown in Fig. 8a. Optimized Acquity provided steady H only for solutes with k > 4. Whenever the column was connected to the column port of Acquity, constant H values were not observed for solutes with k ≤ 5, although the column efficiency was not necessarily poor, Kinetex and InertSustian providing about 8000 theoretical plates with low-eluting-strength weak-wash solvent for peaks with k > 5. In Fig. 8, the observed plate height values on the monolithic silica C18 column were least affected by extracolumn effect compared to the other columns at similar retention factor. This is due to the greater band volume on the monolithic column at the same k values based on the greater porosity of the column. In combination with the results shown by van Deemter plots in Figs. 6 and 7, the monolithic silica column can provide optimum performance at low flow rates and with small extra-column effect. The severity of the extra-column effect is a function of the band volume or peak variance of a solute observed with a particular column, which in turn is a function of column dead volume (i.e.

56

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

Table 4 Properties of 1 mm I.D., 5 cm long, columns.

Porosity Particle size (␮m) Column dead volume (␮L) Permeability (m2 ) Number of theoretical plates (optimum) Height equivalent to a theoretical plate (optimum) (␮m) Separation impedance Methylene group selectivity, ␣(CH2 ) Retention factor of k(C4 H9 COPh) (Retention time, tR min)b a b

Kinetex C18

InertSustain C18

Monolithic C18

0.55 2.6 21.5 7.0E−15 8500–9000 5.7 5000 1.581 2.90 (1.77)

0.54 2.0 21.3 2.9E−15 8500–9000 5.7 11,200 1.566 5.64 (2.84)

0.67 (1.8)a 26.3 1.7E−14 8500–9000 5.7 1900 1.595 1.97 (1.65)

Domain size, or combined size of a through-pore and a skeleton. Net retention time in mobile phase of acetonitrile/water = 60/40, with flow rate at 0.05 mL/min.

its porosity), the retention factor of the solute, and the intrinsic efficiency of the column under the conditions employed, including mobile phase linear velocity. Therefore the contribution of the extra-column effect to the observed column efficiency is column dependent. The greatest effect was observed for the InertSustain column, and the smallest with the monolithic silica column at low linear velocity. If all solutes should provide similar plate height in the absence of extra-column effect, subtracting a discrete value for extra-column band variance from the total band variance values should be able to correct for the extra-column effect for all solutes, and for all columns. In other words, if the columns under examination were to possess similar intrinsic column efficiency, and a single extracolumn effect is superimposed, all columns should show similar tendency in Fig. 8, when corrected for extra-column effect, according to Eqs. (1)–(3). But this was not the case. The plots obtained with Express LC Ultra could be corrected by assuming the follow2 ing different extra values 0.02, 0.04, and 0.12 ␮L2 for monolithic silica, Kinetex, and InertSustain, respectively, resulting in more-orless constant H values within the range of retention factors studied, as shown in Fig. 8b. The extra-column effect is instrument-configuration dependent, therefore it should have the same effect on the performance of 2 all columns. If a single extra value provided consistent H values for 2 value could be all columns and for all solutes, then such single extra

regarded as an intrinsic characteristic of the instrument in its par2 value does not seem to exist ticular configuration, but such extra in the present case. There seems to be an additional contributor to band broadening that is characteristic to the column and not to the instrument, which is similar to the extra-column effect while different in magnitude. Non-ideal frit/inlet structure or flow path structure can act similarly to the extra-column effect, but cannot be easily accounted for by lumping the two together. In such a case, the intrinsic plate height (i.e. true efficiency of a packed bed) could be elusive. The use of weak wash solvent of low eluting strength improved the column efficiency for early eluting solutes as shown in Fig. 8d–f. Consistent correction of the resulting plots of plate height against retention factor, however, was not possible with any extra-column band variance value due to the different extent of band compression effect of weak wash solvent for each solute during injection. The plot of observed plate height values against k values (retention factor) seems to reflect the column efficiency correctly. The results of this study imply that the evaluation of column efficiency with an instrument with minimum extra-column effect is most straightforward and reliable for a small-sized column. By utilizing such a method, one may be able to evaluate small-sized columns correctly, and obtain information to improve the columns, equipment, and operating conditions including injection and detection.

Fig. 8. Plot of HETP versus retention factor obtained with (a) ExpressLC Ultra, (b) ExpressLC Ultra with corrections for extra-column variance of 0.02, 0.04, 0.12 ␮L2 for monolithic silica C18, Kinetex C18, and InertSustain C18, respectively, (c) Acquity with the column connected to the injector, (d)–(f) with the column connected to the column port with 20% methanol, 50% methanol, and 65% acetonitrile as weak-wash solvent, respectively. Column: () Monolithic silica N2595, () Kinetex C18, () InertSustain C18. Mobile phase: acetonitrile/water = 65/35. Weak-wash solvent: (c,e) methanol/water = 50/50, (d) methanol/water = 20/80, and (f) acetonitrile/water = 65/35. Solutes as indicated in the figure caption of Fig. 2, in reversed-phase check-out sample in 65% acetonitrile.

Y. Ma et al. / J. Chromatogr. A 1383 (2015) 47–57

4. Conclusions 1. Small-sized columns (1 mm I.D., 5 cm length) can be evaluated, and will be practically used, with a suitably designed instrument for high-efficiency operation of such small columns. The instrument needs an adequate flow-rate range and must induce minimal band broadening during injection and detection. A UV detector with a small-sized cell can essentially eliminate the extra-column effect on the detector side. The performance of the capillary-LC equipment tested in this study was excellent. An injector inducing minimal band broadening, further improved column packing techniques, and/or improved column hardware may bring about higher levels of efficiency with small-size columns. 2. The 1 mm I.D., 5 cm long columns were able to provide up to ca. 80–90% of the theoretical plates generated by a 2.1 mm I.D. column on capillary-LC or UHPLC instruments under optimized conditions (except for early eluting peaks). It appears that 1 mm I.D., 5 cm long columns have not been made so far to match the efficiency of larger-sized columns. 3. Small-sized monolithic silica columns can be made to provide similar column efficiency to some advanced particulate columns at optimum flow rate, but less efficient at high flow rates. Monolithic columns may have an advantage over particulate columns if the manufacturing of the latter to high efficiency remains difficult. 4. Extra-columns effects can be handled with 1 mm I.D. columns (despite their small column peak variances), because such columns are used at low flow rate where the magnitude of extracolumn effect was much reduced. It is difficult to assess intrinsic column performance in the case of 1 mm I.D., 5 cm long columns by accounting for extra-column band variance. Accurate measurements of extra-column peak variance are difficult. Plotting observed plate height values against retention factors of solutes seems to be a practical way to estimate the efficiency of a column in the presence of extra-column effects. 5. In some cases, retained peaks showed smaller band width than the width of a band detected without including the column in the flow path due to the band compression effect induced by the presence of weak injector-wash solvent in the sample loop when partial-loop injection is performed. Partial loop injection in the presence of weak solvent may allow for the use of UHPLC instruments with 1 mm I.D., 5 cm long columns with some limitations, while the use of micro-LC instruments is highly desirable. Acknowledgements The authors thank Dr. Subodh Nimkar, AB SCIEX, for his assistance in instrumentation, and Dr. Carl Sanchez of Phenomenex for valuable discussion. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma. 2015.01.014. References [1] D. Ishii, K. Asai, K. Hibi, T. Jonokuchi, M. Nagaya, A study of microhigh-performance liquid chromatography: I. Development of technique for miniaturization of high-performance liquid chromatography, J. Chromatogr. 144 (1977) 157–168.

57

[2] P. Kucera, Design and use of short microbore columns in liquid chromatography, J. Chromatogr. 198 (1980) 93–109. [3] T. Takeuchi, D. Ishii, High-performance micro packed flexible columns in liquid chromatography, J. Chromatogr. 213 (1981) 25–32. [4] T. Tsuda, M. Novotny, Packed microcapillary columns in high performance liquid chromatography, Anal. Chem. 50 (1978) 271–275. [5] Y. Hirata, M. Novotny, T. Tsuda, D. Ishii, Packed microcapillary columns with different selectivities for liquid chromatography, Anal. Chem. 51 (1979) 1807–1809. [6] G. Guiochon, Preparation and operation of liquid chromatographic columns of very high efficiency, J. Chromatogr. 185 (1979) 3–26. [7] Y. Zhang, M. Bao, X. Li, P. Lu, High-performance liquid chromatographic columns of small diameter, J. Chromatogr. 197 (1980) 97–108. [8] R.P.W. Scott, P. Kucera, Mode of operation and performance characteristics of microbore columns for use in liquid chromatography, J. Chromatogr. 169 (1979) 51–72. [9] P. Kucera, G. Manius, High-resolution reversed-phase liquid chromatography utilizing microbore column concatenation, J. Chromatogr. 216 (1981) 9–12. ˇ [10] K. Slais, D. Kouˇrilová, Minimization of extra-column effects with microbore columns using electrochemical detection, J. Chromatogr. 258 (1983) 57–63. [11] J. Bowermaster, H. McNair, Microbore high-performance liquid chromatographic columns: speed, efficiency, sensitivity and temperature programming, J. Chromatogr. 279 (1983) 431–438. [12] V. Kahle, M. Krejˇci, Role of extra-column volume in the efficiency of high-speed liquid chromatography, J. Chromatogr. 321 (1985) 69–79. [13] R. Gill, Practical narrow-bore high-performance liquid chromatography system using commercial equipment and 1-mm bore columns packed in the laboratory, J. Chromatogr. 354 (1986) 169–183. [14] F. Gritti, G. Guiochon, On the extra-column band-broadening contributions of modern, very high pressure liquid chromatographs using 2.1 mm I.D. columns packed with sub-2 ␮m particles, J. Chromatogr. A 1217 (2010) 7677–7689. [15] F. Lestremau, D. Wu, R. Szücs, Evaluation of 1.0 mm I.D. column performances on ultra high pressure liquid chromatography instrumentation, J. Chromatogr. A 1217 (2010) 4925–4933. [16] N. Wu, A.C. Bradley, Effect of column dimension on observed column efficiency in very high pressure liquid chromatography, J. Chromatogr. A 1261 (2012) 113–120. [17] D. Spaggiari, S. Fekete, P.J. Eugster, J.-L. Veuthey, L. Geiser, S. Rudaz, D. Guillarme, Contribution of various types of liquid chromatography–mass spectrometry instruments to band broadening in fast analysis, J. Chromatogr. A 1310 (2013) 45–55. [18] F. Gritti, G. Guiochon, Rapid development of core–shell column technology: accurate measurements of the intrinsic column efficiency of narrow-bore columns packed with 4.6 down to 1.3 ␮m superficially porous particles, J. Chromatogr. A 1333 (2014) 60–69. [19] F. Gritti, G. Guiochon, Accurate measurements of the true column efficiency and of the instrument band broadening contributions in the presence of a chromatographic column, J. Chromatogr. A 1327 (2014) 49–56. [20] W. Th. Kok, U.A. Th. Brinkman, R.W. Frei, H.B. Hanekamp, F. Nooitgedacht, H. Poppe, Use of conventional instrumentation with microbore column in highperformance liquid chromatography, J. Chromatogr. 237 (1982) 357–369. [21] J.P.C. Vissers, A.H. de Ru, M. Ursem, J.-P. Chervet, Optimised injection techniques for micro and capillary liquid chromatography, J. Chromatogr. A 746 (1996) 1–7. [22] F. Gritti, C.A. Sanchez, T. Farkas, G. Guiochon, Achieving the full performance of highly efficient columns by optimizing conventional benchmark highperformance liquid chromatography instruments, J. Chromatogr. A 1217 (2010) 3000–3012. [23] C. Sanchez, J.A. Anspach, T. Farkas, Performance optimizing injection sequence for minimizing injection band broadening contributions in high efficiency liquid chromatographic separations, J. Chromatogr. A 1228 (2012) 338–348. [24] C. Sanchez, G. Friedlander, S. Fekete, J. Anspach, D. Guillarme, M. Chitty, T. Farkas, Pushing the performance limits of reversed-phase ultra high performance liquid chromatography with 1.3 ␮m core–shell particles, J. Chromatogr. A 1311 (2013) 90–97. [25] K. Morisato, S. Miyazaki, M. Ohira, M. Furuno, M. Nyudo, H. Terashima, K. Nakanishi, Semi-micro-monolithic columns using macroporous silica rods with improved performance, J. Chromatogr. A 1216 (2009) 7384–7387. [26] F. Gritti, G. Guiochon, Effect of the pressure on pre-column sample dispersion theory, experiments, and practical consequences, J. Chromatogr. A 1352 (2014) 20–28. [27] W. Smits, S. Deridder, G. Desmet, The impact of flow distribution on column performance: a computational fluid dynamics study, J. Chromatogr. A 1369 (2014) 125–130. [28] Y. Vanderheyden, K. Broeckhoven, G. Desmet, Comparison and optimization of different peak integration methods to determine the variance of unretained and extra-column peaks, J. Chromatogr. A 1364 (2014) 140–150. [29] E. Oláh, S. Fekete, J. Fekete, K. Ganzler, Comparative study of new shell-type, sub-2 ␮m fully porous and monolith stationary phases, focusing on masstransfer resistance, J. Chromatogr. A 1217 (2010) 3642–3653.

Efficiency of short, small-diameter columns for reversed-phase liquid chromatography under practical operating conditions.

Prototype small-size (1.0mm I.D., 5cm long) columns for reversed-phase HPLC were evaluated in relation to instrument requirements. The performance of ...
2MB Sizes 0 Downloads 9 Views