Narrow, Open, Tubular Column for Ultrahigh-Efficiency Liquid

Technical Note Cite This: Anal. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/ac

Narrow, Open, Tubular Column for Ultrahigh-Efficiency LiquidChromatographic Separation under Elution Pressure of Less than 50 bar Yu Yang,† Huang Chen,† Matthew A. Beckner,† Piliang Xiang,† Joann Juan Lu,† Chengxi Cao,‡ and Shaorong Liu*,† †

Department of Chemistry and Biochemistry, University of Oklahoma, Norman, Oklahoma 73019, United States Laboratory of Analytical Biochemistry and Bio-separation, State Key Laboratory of Microbial Metabolism, School of Life Science and Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, People’s Republic of China

Anal. Chem. Downloaded from pubs.acs.org by UNIV OF SOUTH DAKOTA on 09/07/18. For personal use only.

‡

S Supporting Information *

ABSTRACT: We report that we can achieve extremely high separation efficiencies using a narrow, open, tubular (NOT) column for liquidchromatographic separations, and we can carry out these separations under an elution pressure of no more than 50 bar. To improve the separation efficiency in packed-column liquid chromatography, one of the most effective approaches is to reduce the monodispersed-particle sizes. A direct consequence of reduced particle size is an increased elution pressure. High efficiencies have been obtained in ultrahigh-performance liquid chromatography (UPLC) using 1−2 μm or even submicron particles, and high elution pressures (greater than 1000 bar) are commonly used to carry out these separations. Open, tubular (OT) columns have been predicted to be the most efficient columns for high-efficiency liquidchromatographic separations, as long as the column diameter is sufficiently small (1−2 μm). However, high efficiencies have not yet been publically reported, possibly because of the challenges (such as picoliter-volume detection, nanocapillary-column preparation, low sample loadability, etc.) of utilizing 1−2 μm diameter capillaries. In this paper, we show how we overcame these problems and achieved extremely high separation efficiencies using a 2 μm inner diameter capillary. We see 200+ apparent peaks with a peak capacity of 810 within 54 min when separating a sample from trypsin-digested cytochrome C, and we count 440 apparent peaks with a peak capacity of 1640 within 172 min when separating a sample from pepsin/trypsin-digested Escherichia coli cell lysate.

T

studies3−6 have predicted that OT columns offer the best means of achieving high separation efficiencies for liquid chromatography (LC). The theory has also predicted that the optimal inner diameter (i.d.) of the OT column is in the range of 1 to 2 μm. However, open, tubular liquid chromatography (OTLC) has rarely been exploited using columns in this i.d. regime because chromatographers doubt that such a “singlepore” with limited sample loadability could lead to high-quality chromatographic separations. There are also challenges (such as picoliter-volume detection, nanocapillary-column preparation, low sample loadability, etc.) to performing separations utilizing such narrow capillaries. To enhance the sample loadability, Jorgenson et al.7 roughened a borosilicate-glass capillary with hydrochloric acid leaching, rendering a porous silica layer by selectively removing the nonsilica components of the glass. This process increased the internal surface area of the capillary roughly 30fold over that of a geometrically smooth capillary. Pesek and

here has been a trend historically toward using smallerparticle packing materials.1 The first liquid-chromatographic separations were performed using columns packed with large particles, and eluents were driven by gravity. To improve the separation efficiency, columns packed with small particles were used and consequently eluents had to be forced through these columns via pressure pumps, leading to the most widely used analytical separation technique: high-pressure liquid chromatography (HPLC). Common HPLC systems often use columns 25 cm in length and 4.6 mm in diameter packed with particles 5 μm in diameter and operate often within a pressure limit of approximately 400 bar (6000 psi). In 1997, 1.5 μm nanoporous silica particles were packed into a capillary column2 for fast and high-efficiency separations, and an ultrahigh pressure (1400 bar) had to be applied to drive an eluent through this column. This work has established today’s ultrahigh-pressure liquid chromatography (UHPLC or UPLC). Looking from a fundamental point of view, the objective of reducing the monodisperse-particle size is to decrease the pore sizes among the particles and shorten the mass-transfer time in the stationery phase. An efficient way to realize this is by using a narrow-bore, open, tubular (OT) column. In fact, theoretical © XXXX American Chemical Society

Received: June 11, 2018 Accepted: August 24, 2018 Published: August 24, 2018 A

DOI: 10.1021/acs.analchem.8b02634 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry Matyska8 reported that by etching the surface of a 400 μm i.d. fused-silica capillary using ammonium hydrogen difluoride,9 the surface area could be increased by up to 1000-fold. These porous surfaces were suitable for bonding of silane stationary phases and increased sample loadability was obtained. Alternatively, a porous polymer layer could be formed on the interior wall of a capillary to increase the surface area for sample loading.3,5,10,11 For example, Yue et al.12 prepared a layer of poly(styrene-divinylbenzene) on the wall of a 4.2 m × 10 μm i.d. capillary and loaded ∼100 fmol of angiotensin I and ∼50 fmol of insulin for separation. High-peak-capacity results have been obtained12,13 using these columns. With the application of microfabrication technologies for producing analytical devices, scientists can now precisely and reproducibly make complex two-dimensional structures for chromatographic separations.14,15 In particular, Desmet et al.16 utilized micromachined, radially elongated pillars for liquidchromatographic separations. Because the pillars were perfectly arranged so that the all pores had the same size and length, each pore worked as the tube (but with turns) of an OT column. Meanwhile, because each pillar served a similar function to that of a particle in a packed-column and because there were multiple pores, this technique resembled packedcolumn LC. The authors basically merged OTLC and packedbed LC. Using such a micromachined column (4 × 1 mm area filled with radially elongated pillars) for LC separation, these researchers obtained efficiencies of 160 000 theoretical plates for unretained analytes and 70 000 theoretical plates for a retained Coumarin derivative. In this paper, we show how we obtain extraordinarily high efficiencies using a 2 μm i.d. open capillary column. The interior wall of the capillary is derivatized with trimethoxy(octadecyl) silane. An Agilent 1200 gradient pump combined with a six-port sample-injection valve and a flow splitter is used to implement the sample injection and gradient elution. A laser-induced fluorescence detector was built for on-column detection of resolved analytes; all analytes are labeled with ATTO-TAG FQ before sample injection. When separating a sample prepared by trypsin-digesting cytochrome C, we obtain 220 apparent peaks and estimate a peak capacity of 810 within 68 min. For separating a mixture of peptides prepared by pepsin/trypsin-digesting Escherichia coli cell lysates, we count 440 apparent peaks and estimate a peak capacity of 1640 within 172 min.

one end was removed for about 1 cm in length. A 25 G X 7/8″ hypodermic needle was used as a guide to facilitate the insertion of this capillary into a pressure chamber (made of transparent acrylic, see the SI for details) containing 50 μL of a 1 M NaOH solution. The other end of the capillary (with the polyimide coating) was placed into a 0.5 mL sealed vial containing DDI water. Nitrogen at a pressure of 500 psi was applied to wash the capillary with NaOH at 100 °C for 2 h; ∼50 nL of the solution passed through the column. The NaOH solution was then replaced with DDI water to rinse the capillary for another hour; ∼25 nL of the solution passed through the column. The washing setup (pressure chamber, capillary, and waste container) was moved out of the oven, and the same setup was used to rinse the capillary with acetonitrile for about 30 min (∼10 nL of acetonitrile passed through the column) at ambient temperature and dry the capillary with nitrogen overnight. The above washing setup was moved inside a dry glovebox. A coating reagent of 50 μL of trimethoxy(octadecyl) silane and 50 μL of toluene was prepared in the dry glovebox and placed inside the pressure chamber. The capillary end with the polyimide coating was dipped into a 0.5 mL sealed vial with toluene. The polyimide-removed end of the capillary was introduced to the coating reagent inside the pressure chamber. This setup was then moved into an oven and the coating reagent was flushed through the capillary at 50 °C for 16 h. The coating reagent was then replaced with toluene to rinse the capillary for about an hour. The column was ready to use after being dried with nitrogen. Peptide-Sample Preparation. To prepare tryptic digests of cytochrome C, 100 μL of a 10 mg/mL cytochrome C stock solution was diluted to 1 mg/mL with 25 mM NH4HCO3, mixed with 1 μL of 1 M DTT, and kept at room temperature for 1 h. Then, the mixture was reacted with 10 μL of 0.2 mg/ mL trypsin solution at 37 °C for 8 h. To prepare tryptic digests of E. coli lysates, 1 mL of the E. coli lysate (the solution was estimated to contain ∼10 mg/mL total protein) was mixed with 5 μL of 1 M NaAc/HAc buffer (pH 4) and 1 μL of pepsin (1 μg/mL) and incubated at 37 °C for 1 h. The above solution (100 μL) was diluted with 900 μL of 25 mM NH4HCO3 and mixed with 1 μL of 1 M DTT at room temperature for at least 1 h. Then, 10 μL of 0.2 mg/mL trypsin solution was added into the above mixture, and the mixture was incubated at 37 °C for 24 h. Analyte Fluorescence Labeling. Following the instruction provided with the ATTO-TAG FQ Amine-Derivatization Kit, a 10 mM ATTO-TAG FQ stock solution was prepared by dissolving 5.0 mg of ATTO-TAG FQ in 2.0 mL of methanol and stored at −20 °C before use. A 10 mM working KCN solution was prepared by diluting a 0.2 M KCN stock solution with a 10 mM borax solution (pH 9.2). Amino acid stock solutions (each containing one amino acid at 1 mM) were prepared by dissolving individual amino acids in DDI water and filtered with 0.22 μm filters. A volume of 1.0 μL of the amino acid stock solution was mixed with 10 μL of the 10 mM KCN working solution and 5 μL of the 10 mM FQ solution in a 0.25 mL vial. This mixture was maintained at room temperature for 1 h in a dark environment before it was ready to test. The resulting FQ-labeled amino acid was diluted with 10 mM NH4HCO3 solution prior to analysis. To label tryptic digests of cytochrome C or E. coli lysate, 10 μL of the peptide solution was mixed with 10 μL of the 10 mM KCN working solution and 10 μL of the 10 mM FQ solution.

■

EXPERIMENTAL SECTION Materials and Reagents. Amino acids, cytochrome C, sodium hydroxide, ammonia bicarbonate, acetonitrile, toluene and trimethoxy(octadecyl) silane were purchased from SigmaAldrich (St. Louis, MO). ATTO-TAG FQ Amine-Derivatization Kit was purchased from Thermo Fisher Scientific (Waltham, MA). Trypsin was purchased from Promega (Madison, WI). Pepsin was purchased from MP Biomedicals (Santa Ana, CA). All solutions were prepared in ultrapure water (Nanopure Ultrapure Water System, Barnstead, Dubuque, IA) and filtered through a 0.22 μm filter (VWR, Missouri, TX) that was degassed before use. The fused-silica capillaries used for making the narrow, open, tubular (NOT) columns (2 μm i.d.; 150 μm outer diameter, o.d.) were produced by Polymicro Technologies, a subsidiary of Molex (Phoenix, AZ). Preparation of NOT Column. A capillary, 80 cm long, 2 μm i.d., and 150 μm o.d., was cut, and the polyimide coating at B

DOI: 10.1021/acs.analchem.8b02634 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

the stage were locked. The capillary was thoroughly rinsed with an eluent (e.g., 50% acetonitrile with 10 mM NH4HCO3) before conducting NOTLC separation For separating amino acids (typically 0.1 μM), the NOT column had a 2 μm i.d., a total length of 48 cm (44 cm effective), and an o.d. of 150 μm. The eluent consisted of 75% mobile phase A (10 mM NH4HCO3 solution) and 25% mobile phase B (80% acetonitrile with 10 mM NH4HCO3). The Agilent pump combined with the flow splitter provided an elution pressure of 650 psi at the head of the NOT column. The external loop of the injection valve had a volume of 6 μL. The flow rate in the restriction capillary was about 7.5 μL/min (measured), and it took ∼50 s for the 7.5 μL sample to pass across the head of the NOT column. The eluent flow velocity inside the NOT column was ∼1 mm/s, corresponding to a flow rate 0.19 nL/min. The injected sample volume was estimated to be ∼150 pL. The samples of enzyme-digested cytochrome C and E. coli lysate were separated using a 2 μm i.d., 80 cm long (75 cm effective) trimethoxy(octadecyl) silane coated NOT column; 10 mM NH4HCO3 in DDI water was used as mobile phase A, and 80% acetonitrile in 10 mM NH4HCO3 was used as mobile phase B. The elution pressure was ∼500 psi. The external loop of the injection valve had a volume of 6 μL. The flow rate in the restriction capillary was ∼5.6 μL/min (measured), and it took ∼64 s for the 6 μL sample to pass across the head of the NOT column. Under these conditions, an eluent linear flow velocity of 0.6 mm/s was measured inside the NOT column, and the injected volume was estimated to be ∼120 pL. For separating trypsin-digested cytochrome C, the gradient started with 100% mobile phase A, and mobile phase B increased from 0 to 50% within 50 min and from 50 to 80% from 50 to 55 min. For separating pepsin/trypsin-digested E. coli lysates, the gradient also started with 100% mobile phase A, whereas mobile phase B increased linearly from 0 to 100% within 180 min.

After a 1 h reaction in the dark at room temperature, the peptides were ready for dilution (with 10 mM NH4HCO3) and separation. Apparatus. Figure 1 presents the experimental apparatus used in this work. A gradient pump (Agilent 1200 quaternary

Figure 1. Schematic diagram of experimental apparatus. The NOT column had a 2 μm i.d. and was trimethoxy(octadecyl) silane derivatized. The gradient pump was an Agilent 1200 capillary pump. The injection was a VICI six-port valve. The flow splitter was built using an Upchurch micro-T. A 10 cm long and 20 μm i.d. capillary was used as a restrictor. A 200 μm i.d. and 360 μm o.d. capillary was used to connect the injection valve and the micro-T. Inside the flow splitter, a small portion (head) of the 2 μm i.d. and 150 μm o.d. NOT column was inserted into a 200 μm i.d. and 360 μm o.d. connection capillary to facilitate sample injection. At 5 cm from the effluent outlet of the NOT column, a detection window was made by removing the polyimide coating. A laser-induced fluorescence detector17 underneath the NOT column was used to monitor the resolved analytes.

pump, Santa Clara, CA) was used for driving a mobile phase through a six-port valve (VICI Valco, Houston, TX), via a flow splitter with a 20 μm i.d. and 20 cm long restriction capillary, to a NOT column. At 5 cm from the tip of the NOT column, the polyimide coating was removed, forming a detection window. The detection end of the column was affixed to a capillary holder on an x−y−z translation stage so that the detection window could be aligned for maximum fluorescence output. The confocal laser-induced-fluorescence (LIF) detector was described previously.17 Briefly, an argon-ion laser (LaserPhysics, Salt Lake City, UT) generated a 488 nm laser beam. Then, the laser beam was directed by a dichroic mirror (Q505LP, Chroma Technology, Rockingham, VT) and focused onto the detection window of the narrow capillary via an objective lens (20× and 0.5 NA, Rolyn Optics, Covina, CA). The fluorescence emission was collimated by the same lens; passed through the same dichroic mirror, an interference band-pass filter (532 nm, Life Technologies, Carlsbad, CA), and a 1 mm pinhole; and finally was collected by a photosensor module (H5784-04, Hamamatsu Photonics, Hamamatsu, Japan). A data-acquisition card (USB-1208FS, Measurement Computing, Norton, MA) was used to measure the response from the photosensor module as a voltage signal. The data were collected and analyzed by a homemade LabView program (National Instruments, Austin, TX). NOTLC Separation. To align the NOT column on the LIF detector, a 10 μM fluorescein solution was pressurized through the column, and the column was roughly aligned with the detector as shown in Figure 1. The fluorescein solution was constantly flushed through the column until the alignment was done. Then, the LIF detector was turned on, and the fluorescence signal was monitored. By tuning the column position via the x−y−z translation stage until the maximum fluorescence output was obtained, the x, y, and z positions of

■

RESULTS AND DISCUSSION High-Efficiency Separation. Figure 2A presents a chromatogram for the separation of amino acids. To view the peak profiles, an expanded view of these peaks is exhibited in the inset. Many of the peaks had peak full widths at half maxima (fwhm) of 0.3−0.5 s. Because of the use of a highly mass-permissive, open, tubular column, the elution pressure was low (600 psi). Figure 2B presents a chromatogram for separating trypsin-digested cytochrome C. Using the criteria of a signal-to-noise ratio of 3, we counted 220 peaks. To evaluate the peak capacity, we used the method described previously18 for gradient elution with minor modifications. In the referenced method, peaks were assumed to be present over the entire gradient chromatogram. The peak capacity could simply be calculated from the gradient run time divided by the average peak width. The modification we made was to use the gap between the first peak (the peak of an unretained analyte) and the last peak to replace the gradient run time. For the chromatogram in Figure 2B, we measured the fwhm values of the 10 highest peaks to be ∼2.3 s, corresponding to an average full width (4σ ≈ 1.7 × fwhm) of 3.97 s. We also measured the retention time of an unretained analyte to be at ∼13.5 min and the retention time of last peak to be at 67.2 min. We divided the gap between these two peaks by the average full peak width and obtained a peak capacity of 810 within 54 min. C

DOI: 10.1021/acs.analchem.8b02634 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

Sample Loadability. Sample loadability was a great concern for us initially. This was because we wanted to load adequate analytes so that we could detect the resolved peaks, but at the same time, we did not want to overload the column because that would deteriorate the separation. With the method we developed for preparing the NOT column, we noticed that the column had adequate loadability. Figure 3A

Figure 2. Typical NOTLC chromatograms for high-efficiency separations. (A) Separation of amino acids. The NOT column (2 μm i.d. × 48 cm length, 44 cm effective) was trimethoxy(octadecyl) silane derivatized. The injection volume was ∼7.1 pL. Mobile phase A was 10 mM NH4HCO3 in DDI water, and mobile phase B was acetonitrile. The gradient profile was the following: mobile phase B increased from 0 to 50% from 0 to 1.5 min, stayed at 50% B from 1.5 to 2 min, and then decreased from 50 to 0% from 2 to 2.5 min. The elution pressure was ∼600 psi. The sample contained histidine (1), asparagine (2), glycine (3), tyrosine (4), arginine (5), alanine (6), tryptophan (7), valine (8), isoleucine (9), phenylalanine (10), and leucine (11), each at 6.5 μM. (B) Separation of trypsin-digested cytochrome C. The NOT column (2 μm i.d. × 80 cm length, 75 cm effective) was trimethoxy(octadecyl) silane derivatized. Mobile phase A was 10 mM NH4HCO3 in DDI water, and mobile phase B was 80% acetonitrile in 10 mM NH4HCO3. The volume of sample injected was ∼120 pL. The elution pressure was ∼500 psi. The gradient was as follows: mobile phase B increased from 0 to 50% within 50 min and then from 50 to 80% from 50 to 55 min. (C) Separation of pepsin/ trypsin-digested E. coli lysate. For the gradient, mobile phase B increased linearly from 0 to 100% within 180 min. All other conditions were the same as in (B).

Figure 3. Sample-loadability-test results. (A) Chromatograms for NOTLC separations of three amino acids. The NOT column (2 μm i.d. × 80 cm length, 75 cm effective) was trimethoxy(octadecyl) silane derivatized. Mobile phase A was 10 mM NH4HCO3 in DDI water, and mobile phase B was acetonitrile. Separation was carried out using 20% mobile phase B and 80% mobile phase A. The injection volume was ∼157 pL. The sample contained 0.04 μM Gly, 0.08 μM Iso, and 0.08 μM Leu in10 mM NH4HCO3 for the bottom chromatogram, and the concentrations were increased by 10 times for the middle chromatogram and 100 times for the top chromatogram. An elution pressure of 650 psi was utilized for these separations. On the basis of the sample concentration and the injection volume, the numbers of molecules injected into the NOT column were calculated and are shown. In each chromatogram, the number of molecules injected had a range because the concentrations of amino acids were different. (B) Effect of injection-sample quantity on peak broadening.

To demonstrate the feasibility of NOTLC for more complex sample separation, we used pepsin to digest an E. coli lysate and trypsin to further digest the resulting sample (see the Experimental Section for details). The digested sample was then separated by NOTLC, and the chromatogram is presented in Figure 2C. Using the same approach for analyzing the chromatogram of trypsin-digested cytochrome C, we counted 440 peaks and estimated a peak capacity of 1640 within 172 min. One-dimension high-peak-capacity separation results were obtained12,19−22 but usually at higher elution pressures and with longer separation times. For example, Han et al.19 performed a separation of digested cytochrome C using a meter-long packed nano-LC column under an elution pressure of 5800 psi and obtained a peak capacity of 800 within 10+ h. Shen et al.21 used 20 kpsi RPLC-MS to analyze human-plasma tryptic digests and generated peak capacities of 1000−1500 in 12+ h. It should be pointed out that because a laser-induced fluorescence detector was used to monitor the resolved peaks, analytes were labeled with ATTO-TAG FQ. Excess labeling dye was added in order to label all binding sites on each peptide, but there would likely be free binding sites, leading to multiple peaks for one peptide. Therefore, each peak in Figure 2B,C may represent only a specific fluorescent molecule (e.g., a peptide labeled with a specific number of fluorescent dye molecules at specific binding sites). On the other hand, a particular peak might include multiple fluorescent molecules if those molecules could not be resolved.

presents three chromatograms made by injecting a sample containing 40 nM to 8 μM amino acids, and decent separations were obtained within this concentration range. The fwhm did increase with the amino acid concentration (see Figure 3B), but all peaks still had fwhm values of 0.5 s or less. The number of molecules injected into the NOT column were calculated on the basis of the sample concentration and the injection volume. It is worth pointing out that eight hundred million molecules correspond to femtomoles of molecules. Apparently, if we need to inject more analytes into the column, we can, but it will be at an expense of resolution or efficiency. Reproducibility. Other concerns included the stability of the coatings and the reproducibility of preparing the coatings, both would contribute to the reproducibility of the separations. Figure S2A presents the results of the coating-stability experiment. The performance of the coating remained virtually the same after 192 runs in more than 5 months; the amino acid retention-time standard deviations were less than 1%, and the peak-area standard deviations were less than 5%. Figure S2B presents the reproducibility results of the coating preparations. The performances of eight different columns prepared at different times by different students were very constant; the retention-time standard deviations were less than 1%, and the peak-area standard deviations were less than 3%. D

DOI: 10.1021/acs.analchem.8b02634 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

■

(15) De Malsche, W.; Eghbali, H.; Clicq, D.; Vangelooven, J.; Gardeniers, H.; Desmet, G. Anal. Chem. 2007, 79, 5915−5926. (16) Desmet, G.; Callewaert, M.; Ottevaere, H.; De Malsche, W. Anal. Chem. 2015, 87, 7382−7388. (17) Weaver, M. T.; Lynch, K. B.; Zhu, Z.; Chen, H.; Lu, J. J.; Pu, Q.; Liu, S. Talanta 2017, 165, 240−244. (18) Neue, U. D. J. Chromatogr. A 2005, 1079, 153−161. (19) Han, J.; Ye, L.; Xu, L.; Zhou, Z.; Gao, F.; Xiao, Z.; Wang, Q.; Zhang, B. Anal. Chim. Acta 2014, 852, 267−273. (20) Luo, Q.; Shen, Y.; Hixson, K. K.; Zhao, R.; Yang, F.; Moore, R. J.; Mottaz, H. M.; Smith, R. D. Anal. Chem. 2005, 77, 5028−5035. (21) Shen, Y.; Zhang, R.; Moore, R. J.; Kim, J.; Metz, T. O.; Hixson, K. K.; Zhao, R.; Livesay, E. A.; Udseth, H. R.; Smith, R. D. Anal. Chem. 2005, 77, 3090−3100. (22) Zhou, F.; Lu, Y.; Ficarro, S. B.; Webber, J. T.; Marto, J. A. Anal. Chem. 2012, 84, 5133−5139.

CONCLUSIONS We have demonstrated extraordinarily high efficiencies for NOTLC separations. Another great feature of NOTLC separation is that it does not require high elution pressures. We had been concerned about the sample loadability, the column-coating durability, and the column-preparation reproducibility. These concerns have been addressed nicely with the approaches we took for preparing the columns and running the separations. We are now working to couple NOTLC with mass spectrometry, and we expect NOTLC to be a powerful analytical technique for chemical and biochemical analyses.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.8b02634. Detailed construction of the pressure chamber, coatingstability and coating-preparation-reproducibility results, detailed interday- and intraday-test results, coating stability under various pH conditions, estimation of the sample injection volume, and complete chromatograms for Figure 2B,C (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: 1 (405) 325 6111. ORCID

Piliang Xiang: 0000-0002-7742-7083 Chengxi Cao: 0000-0002-3873-5112 Shaorong Liu: 0000-0002-4602-0483 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is partially sponsored by the Department of Energy (DE-SC0006351) and the Oklahoma Center for the Advancement of Science and Technology (AR17-022).

■

REFERENCES

(1) Jorgenson, J. W. Annu. Rev. Anal. Chem. 2010, 3, 129−150. (2) MacNair, J. E.; Lewis, K. C.; Jorgenson, J. W. Anal. Chem. 1997, 69, 983−989. (3) Knox, J. H.; Gilbert, M. T. J. Chromatogr. A 1979, 186, 405−418. (4) Knox, J. H. J. Chromatogr. Sci. 1980, 18, 453−461. (5) Guiochon, G. Anal. Chem. 1981, 53, 1318−1325. (6) Jorgenson, J. W.; Guthrie, E. J. J. Chromatogr. A 1983, 255, 335− 348. (7) Jorgenson, J.; Kennedy, R.; St. Claire, R. L., III; White, J.; Dluzneski, P.; de Wit, J. S. J. Res. Natl. Bur. Stand. 1988, 93, 403−406. (8) Pesek, J. J.; Matyska, M. T. J. Chromatogr. A 1996, 736, 255− 264. (9) Onuska, F.; Comba, M.; Bistricki, T.; Wilkinson, R. J. Chromatogr. A 1977, 142, 117−125. (10) Kennedy, R. T.; Oates, M. D.; Cooper, B. R.; Nickerson, B.; Jorgenson, J. W. Science 1989, 246, 57−63. (11) Tijssen, R.; Bleumer, J.; Van Kreveld, M. J. Chromatogr. A 1983, 260, 297−304. (12) Yue, G.; Luo, Q.; Zhang, J.; Wu, S.-L.; Karger, B. L. Anal. Chem. 2007, 79, 938−946. (13) Rogeberg, M.; Wilson, S. R.; Greibrokk, T.; Lundanes, E. J. Chromatogr. A 2010, 1217, 2782−2786. (14) He, B.; Tait, N.; Regnier, F. Anal. Chem. 1998, 70, 3790−3797. E

DOI: 10.1021/acs.analchem.8b02634 Anal. Chem. XXXX, XXX, XXX−XXX