Monolithic capillary column with an integrated electrochemical detector

Journal of Chromatography A, 1509 (2017) 171–175

Contents lists available at ScienceDirect

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

Short communication

Monolithic capillary column with an integrated electrochemical detector Martina Komendová a , Radovan Metelka a , Jiˇrí Urban a,b,∗ a b

Department of Analytical Chemistry, Faculty of Chemical Technology, University of Pardubice, Studentská 573, 532 10, Pardubice, Czech Republic Department of Chemistry, Faculty of Science, Masaryk University, 625 00, Brno, Czech Republic

a r t i c l e

i n f o

Article history: Received 1 June 2017 Accepted 18 June 2017 Available online 19 June 2017 Keywords: Dopamine Electrochemical detection Integrated column Polymer monoliths Urine

a b s t r a c t The carbon fiber and silver microwire were used as working and pseudoreference electrode, respectively, and inserted into the ending of capillary to prepare monolithic capillary column with an integrated electrochemical detector. Prepared capillary devices offered stable and robust results with relative standard deviations of retention, resolution, and detection signal lower than 1.5, 5.5, and 5.0%, respectively. To further increase sensitivity of developed electrochemical microdetector, multiple pulse amperometry detection mode has been used. Optimized integrated device provided reliable chromatographic separation of mixture of neurotransmitters with calibration curve for dopamine linear from 0.5 to 20.0 mg L−1 and an instrumental limit of detection as low as 24 pg of injected dopamine. Finally, developed capillary column was applied to successful determination of dopamine in a human urine. By using both calibration curve and standard addition method, the dopamine level was determined to be 0.74 ± 0.03 mgL−1 and 0.71 ± 0.02 mgL−1 , respectively. Triplicates of dopamine analysis provided relative standard deviations lower than 2.7% for intraday analyses, while interday relative standard deviations were lower than 3.6% for five consecutive days. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Electrochemical detection in capillary and micro highperformance liquid chromatography (HPLC-EC) of electroactive compounds is easily feasible nowadays [1]. Miniaturization of detection cell is required to suppress peak broadening of analyte, whereas enhanced current densities are necessary to achieve high sensitivity [2]. For that reasons, carbon fiber is often the electrode material of choice. Single carbon fibers [3–6] or strand of carbon fibers [7,8], microband electrodes [9] or interdigitated arrays [10–13] were utilized as working electrodes for capillary or micro HPLC of catecholamine neurotransmitters and related compounds. Polymer monoliths were introduced in 1990s as an alternative stationary phases to spherical particles generally used in conventional chromatographic columns. Monoliths with dominant flow-through pores offer application mainly in a fast gradient elution of synthetic and natural polymers [14], although by using

Abbreviations: HPLC-EC, high performance liquid chromatography with electrochemical detection. ∗ Corresponding author at: Department of Chemistry, Faculty of Science, Masaryk University, Brno, 625 00 Czech Republic. E-mail address: [email protected] (J. Urban). http://dx.doi.org/10.1016/j.chroma.2017.06.057 0021-9673/© 2017 Elsevier B.V. All rights reserved.

various experimental protocols [15–19], monolithic stationary phases allowing separation of small molecules are being developed. So far, only several attempts have been made to develop polymer-based monolithic stationary phase for the separation of neurotransmitters. Recently, we have prepared an online solidphase extraction with liquid chromatography method based on polymer monoliths for the determination of dopamine [20] since undesired changes in its metabolism result in serious illnesses such as depression, schizophrenia, Parkinson disease, and tumors [21,22]. Herein, we describe novel and simple approach to prepare monolithic capillary column with an integrated on-column electrochemical detector. The carbon fiber as working electrode was inserted at the same time with silver microwire as pseudoreference electrode into the ending of monolithic capillary column. To prove the concept, chromatographic separation and amperometric detection of selected neurotransmitters is shown. Since urine is a suitable sample for high-throughput and low cost characterization of nervous system activity [23], the applicability of proposed setup was test on determination of dopamine in urine. On the other hand, proper optimization of experimental conditions allows determination of other electroactive small molecules including amino acids, (poly)phenols, and/or sugars.

172

M. Komendová et al. / J. Chromatogr. A 1509 (2017) 171–175

3. Results and discussion 3.1. Working and reference microelectrodes

Fig. 1. Preparation of integrated electrochemical detector: a) working (carbon fiber, up) and pseudoreference (silver microwire, down) electrodes glued to the silverplated wires, b) insertion of microelectrodes into the end of fused-silica capillary with monolithic stationary phase, and c) microelectrodes inside the end of the capillary. Diameter of carbon fiber and silver microwire is 7 ␮m and 25 ␮m, respectively. Internal diameter of fused-silica capillary is 320 ␮m.

Carbon fibers are characterized by enhanced current densities due to a non-linear diffusion [24]. However, surface treatment applied during their production may have adverse effect on the resulting electrochemical properties of carbon fiber electrodes, therefore, we selected unsized carbon fibers, without any surface modification, as working electrodes for HPLC-EC of catecholamine neurotransmitters. The prerequisite for electrochemical detection in general is the stability of reference part of the detector. A number of designs of reference electrodes were already presented in HPLC-EC with carbon fiber working electrodes. They mostly involve smaller variants of silver/silver chloride electrode of classical construction with a porous frit and a reference solution. Even when reference electrodes of reduced dimensions are used, they are still much larger than the carbon fibers and special flow cells have to be constructed to place them in the vicinity of working electrodes. We present a completely different approach when the microwire reference electrode is an integral part of the detector directly inside the separation capillary together with the carbon fiber working electrode. We have utilized a solid silver microwire with diameter of 25 ␮m as a pseudoreference electrode. Its robust character allows to non-complicated insertion of the microwire into the capillary and the electrode surface is reproducible owing to manufacture process. Constant potential is properly maintained as confirmed by the cyclic voltammograms of dopamine in Figure SI-1 using two-electrode arrangement with carbon fiber as working electrode and silver microwire as pseudoreference electrode. The course of limiting currents in voltammograms at higher concentration of dopamine is less stable, owing to the worse potential stability of silver wire pseudoreference electrode at currents approaching units of microamperes, where polarization of pseudoreference electrode might take place.

2. Experimental part 3.2. Integrated capillary electrochemical detector 2.1. Preparation of monolithic column with integrated electrochemical detector Monolithic capillary columns based on di(ethylene glycol) dimethacrylate crosslinker were prepared according the protocol published previously [17]. Homogenous polymerization mixture was filled in capillaries and an air plug (2 mm) was left at the end of the capillary to allow space for further integration of microelectrodes. Then, both ends of the capillary were sealed with stoppers and the capillary was placed in a thermostated bath where polymerization reaction proceeded at 60 ◦ C for 20 h. After that, monolithic capillary columns were flushed first with acetonitrile, followed by a particular mobile phase. Carbon fiber as working electrode and silver microwire as pseudoreference electrode with diameter 7 ␮m and 25 ␮m, respectively, were attached to contact silver-plated wires using conductive silver paint and allowed to dry at room temperature. Afterwards, the microelectrodes were cut to desired length with a razor blade and fixed in parallel on ceramic slides using cyanoacrylate adhesive (Fig. 1a). The capillary monolithic column was carefully slid to required length onto the working and reference electrode with the aid of micromamipulator MNO-202ND (Narishige, Tokyo, Japan) under microscopic observation using S8APO optical stereomicroscope (Leica, Wetzlar, Germany) (Fig. 1b). The position of capillary with embedded microelectrodes on ceramic support was fixed with cyanoacrylate adhesive afterwards (Fig. 1c). Detailed description of materials, instrumentation, and sample preparation is available in the Supporting Information.

At first, we have explored a repeatability of the microelectrodes integration inside prepared monolithic capillary columns. Table 1 summarizes repeatability data for three independently prepared integrated monolithic capillary columns with electrochemical detectors. Chromatographic analysis provided variability of retention, resolution, and detection signal in the range of relative standard deviations of 0.5–1.4%, 3.9–5.5%, and 3.3–4.9%, respectively, and confirmed robust preparation of both monolithic stationary phase and an electrochemical detector. Monolithic stationary phase used in this work provides dualretention mechanism controlled by the composition of the mobile phase [17]. Hence, we have selected mobile phases with low concentrations of acetonitrile and high phosphate buffer concentrations allowing both chromatographic analysis in a reversed-phase retention mechanism and sensitive electrochemical detection. We have tested an effect of buffer molarity on the signal of dopamine and 3,4-dihydroxybenzylamine that is generally used as an internal standard. With increase in the buffer molarity from 0.1 to 0.2 and 0.5 M, the corresponding signal decreased from 1.7 to 1.4 and 0.2 nA for dopamine and from 1.1 to 1.0 and 0.9 nA for 3,4-dihydroxybenzylamine, respectively. Concentrations of buffer lower than 0.1 M did not have any significant effect on the electrochemical signal. We have also determined an effect of mobile phase composition on the chromatographic resolution of dopamine and 3,4-dihydroxybenzylamine. Unfortunately, none of the mobile phases containing acetonitrile and phosphate buffer enabled

M. Komendová et al. / J. Chromatogr. A 1509 (2017) 171–175

173

Table 1 Repeatability and stability of the detectors preparation. Mean

SD

RSD, %

Integrated Detector 1

I, nA kIS kD RD,IS

Run 19.56 2.08 2.37 1.18

20.33 2.04 2.33 1.21

21.19 2.01 2.30 1.29

20.36 2.05 2.33 1.23

0.67 0.03 0.03 0.05

3.27 1.42 1.13 3.90

Integrated Detector 2

I, nA kIS kD RD,IS

23.11 1.97 2.25 1.39

21.20 1.94 2.27 1.52

23.82 1.94 2.22 1.33

22.71 1.95 2.24 1.41

1.11 0.01 0.02 0.08

4.87 0.77 0.86 5.53

Integrated Detector 3

I, nA kIS kD RD,IS

22.30 1.88 2.15 1.15

20.33 1.86 2.12 1.18

21.19 1.86 2.15 1.27

21.27 1.87 2.14 1.20

0.81 0.01 0.01 0.05

3.79 0.58 0.63 4.26

Mobile phase: 99% 0.1 M phosphate buffer and 1% acetonitrile, flow-rate 3 ␮L/min, I, nA − current signal for dopamine standard 5 · 10−4 M, injection volume 100 nl, kIS and kD − retention factor of 3,4-dihydroxybenzylamine (internal standard) and dopamine, respectively, RD,IS − resolution of dopamine and 3,4-dihydroxybenzylamine calculated by using equation RD,IS = 1.18 · (tR,D − tR,IS )/(w0.5,D + w0.5,IS ), SD − standard deviation, RSD − relative standard deviation. Table 2 Effect of amperometry mode on the detection of dopamine. Injection

IDC , nA

IDC+C , nA

IPAD , nA

1 2 3 4

2.64 2.21 1.39 1.01

2.39 2.24 2.53 2.57

14.16 14.03 14.53 15.08

Mobile phase: 99% water + 0.1% TFA and 1% acetonitrile + 0.1% TFA, flow-rate 3 ␮L/min, back-pressure 7.8 MPa, dopamine standard 5 · 10−4 mol/L, injection volume 100 nl. IDC − direct current amperometry, IDC+C − direct current amperometry with a 30 s electrochemical pretreatment of the microelectrodes at −1 V, IPAD − pulse amperometry.

Fig. 2. Cyclic voltammograms of dopamine in 1% acetonitrile + 0.1% TFA (blue line) and in 0.1 M phosphate buffer (orange line). Potential scan from −0.5 to 1.0 V, Polarization speed 50 mV/s. Working electrode − unsized carbon fiber, pseudoreference electrode − silver microwire. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

a baseline separation of this critical pair. Therefore, we have replaced 0.1 M phosphate buffer with water containing 0.1% TFA. Indeed, chromatographic resolution of dopamine and 3,4dihydroxybenzylamine improved from 1.0 in mobile phase with 1% acetonitrile in 0.1 M phosphate buffer to 1.2 for the mobile phase of 1% acetonitrile in 0.1% TFA and allowed baseline separation of both compounds. Fig. 2 shows cyclic voltammetry curves of dopamine measured in 1% acetonitrile in water with an addition of 0.1% TFA and in 0.1 M phosphate buffer and confirms that a replacement of phosphate buffer by 0.1% TFA did not cause a significant loss of current intensity for the detection of dopamine. Determination of dopamine with prepared microelectrodes is less vulnerable to pH change of the solution and provides comparable values of limiting current in both phosphate buffer and acidified aqueous acetonitrile. The shift of dopamine redox signal can be attributed to the formation of surface layers by the reaction of silver with the components of the used mobile phase, i.e. phosphate and trifluoracetate anions, which could influence the potential of the silver pseudoreference electrode. However, such potential shift has only minor effect on the current values recorded at +1.0 V. 3.3. Improving a sensitivity of an electrochemical detection Passivation of electrode surface by quinone derivatives and other oxidation products is expected during prolonged electro-

chemical oxidation of catecholamines. Peak currents decreased to more than a half of the original value in amperometric mode after four injections of 5 · 10−4 M dopamine (Table 2, IDC ). To overcome such behavior, unfavorable for long-term and repetitive application of microelectrode detector, the electrochemical cleaning of carbon fiber at −1.0 V for 2 min after the separation was found necessary to remove products formed during an electrochemical detection. Data in Table 2 (IDC+C ) confirm the positive effect of such treatment; peak currents for 5 · 10−4 M dopamine were stable enough during four consecutive injections even at such high concentration of analyte. To further increase the sensitivity of detection, multiple pulse amperometry, where very short time sequences (hundreds of miliseconds) of detection potential alternate with a potential that is used for cleaning the electrodes, was utilized with periodic changes between +1.0 V as detection and −1.0 V as cleaning step, respectively. Sampling of current was performed 200 ms after application of detection potential. As shown in Table 2 (IPAD ), sensitivity of dopamine determination increased more than five times for multiple pulse amperometry detection when compared to direct amperometry even when pre-injection cleaning step was performed. Moreover, the repeatability of peak heights was significantly improved by this potential switching, which helps to reduce the fouling of the working electrode. 3.4. Determination of dopamine in human urine An ultimate goal of this work was to develop a monolithic capillary column with on-column electrochemical detection for possible analysis of neurotransmitters. Fig. 3 provides an example of neurotransmitters standards isocratic separation on developed integrated capillary column. Further, the applicability of proposed on-column electrochemical detector was tested using HPLC-EC determination of dopamine. Miniaturized electrochemical detector provided linear response of

174

M. Komendová et al. / J. Chromatogr. A 1509 (2017) 171–175

Stability of an integrated capillary was further confirmed by intra and inter day variation. Triplicates of dopamine analysis provided RSD of 1.1–2.7% for intraday analyses, while RSD for interday dopamine determination (n = 3) was 1.7–3.6% for five consecutive days. 4. Conclusions

Fig. 3. Separation of mixture of neurotransmitters on monolithic capillary column with an integrated electrochemical detector. Mobile phase 5% acetonitrile + 0.1% TFA, flow-rate 7 ␮L/min, back-pressure 8.1 MPa, column length 170 mm. Injection volume 100 nl. Analytes: (1) epinephrine, (2) dopamine, (3) serotonin, (4) homovanillic acid, (5) 5-hydroxyindol-3-acetic acid, (6) 3,4-dihydroxyphenyl acetic acid.

electric current on the concentration of dopamine standards from 0.5 to 20 mg L−1 with regression line IDA = 0.2474 + 0.2155 · cDA and correlation coefficient of 0.9993, where IDA is current signal in nA and cDA is a concentration of dopamine standard in mg L−1 (Figure SI-2). Limit of detection, cLOD , is generally determined from standard deviation of the baseline noise [7,10–12]. In this work, we have calculated more realistic value from standard error of the regression line. Standard deviation of residuals (difference in between fitted and experimental value) has been used to determine limit of detection cLOD = 3 · ␴r /k, where ␴r corresponds to the standard deviation of residuals and k is a slope of the regression curve. Monolithic capillary column with integrated electrochemical detector provided instrumental limit of detection of 0.24 mg L−1 which corresponds to 23.6 pg of injected dopamine (injection volume 100 nl). It should be pointed out, however, that presented limit of detection corresponds only to instrumental setup with developed integrated monolithic capillary column since SPE method used for dopamine extraction has not been included in the validation of dopamine determination. Finally, dopamine has been determined in a human urine to further prove applicability of developed integrated capillary device, as shows Figure SI-3A. Commercially available SPE cartridges using a pH-dependent ring formation reaction of cis-diol group with boronic acid functionality attached at the surface of SPE material (BondElut PBA, Agilent, Palo Alto, CA, USA) were used to extract dopamine from the sample of human urine. Recovery rate of extraction method determined by using 3,4-dihydroxybenzylamine as an internal standard was 97.08 ± 2.05% (n = 3), while matrix had a minimal effect on dopamine recovery and allowed extraction of 98.53 ± 0.75% (n = 3) of the sample. At first, we have used calibration curve to establish dopamine concentration in an extracted sample (Figure SI-2). Calibration curve provided 0.74 ± 0.03 mg L−1 of dopamine which is comparable to previously determined dopamine levels [20,25]. Furthermore, we have utilized a standard addition method (n = 5) to minimize matrix effects and to determine dopamine in human urine as demonstrates Figure SI-4 B and Figure SI-5. Standard addition method provided 0.71 ± 0.02 mg L−1 of dopamine that is in a good agreement with value determined from calibration curve and confirms robustness of prepared integrated capillary in the determination of dopamine in a real sample.

In this work we have used unsized carbon fiber together with a silver microwire to prepare miniaturized electrochemical detector that has been integrated inside monolithic capillary column. Stable and robust chromatographic analysis and multiple pulse amperometry detection of neurotransmitters has been achieved. Additionally, integrated device has been also successfully used to determine dopamine in a human urine. It should be pointed out that presented instrumental detection limit calculated from a regression line is slightly higher when compared to other HPLC-EC system where limits of detection are determined from the noise of the baseline [26]. However, in this proof-of-concept study we aimed mainly in utilization of straightforward preparation of monolithic stationary phases and simple miniaturization of an electrochemical detection. Advantageous combination of these techniques allows further development of tailored analytical systems. Acknowledgement Financial support by the Czech Science Foundation project1422426S is gratefully acknowledged. Appendix A. Supplementary data Supplementary data associated with cle can be found, in the online http://dx.doi.org/10.1016/j.chroma.2017.06.057.

this artiversion, at

References [1] J.P.C. Vissers, H.A. Claessens, C.A. Cramers, Microcolumn liquid chromatography: instrumentation, detection and applications, J. Chromatogr. A 779 (1997) 1–28. ˇ [2] K. St’ulík, V. Pacáková, Electrochemical Measurements in Flowing Liquids, Ellis Horwood Chichester, 1987. [3] L.A. Knecht, E.J. Guthrie, J.W. Jorgenson, On-clumn electrochemical detector with a single graphte fiber electrode for open-tubular liquid-chromatography, Anal. Chem. 56 (1984) 479–482. [4] M. Goto, K. Shimada, T. Takeuchi, D. Ishii, Electrochemical detector using carbon-fiber working electrode for micro high-performance liquid-chromatography, Anal. Sci. 4 (1988) 17–20. [5] S.P. Kounaves, J.B. Young, Carbon-fiber electrode cell for square-wave voltammetric detection of biogenic-amines in high-performance liquid-chromatography, Anal. Chem. 61 (1989) 1469–1472. [6] M. Goto, K. Shimada, Rapid-scanning electrochemical detector with micro working electrode for micro high-performance liquid-chromatograpy, Chromatographia 21 (1986) 631–634. [7] K. Stulik, V. Pacakova, M. Podolak, Carbon-fiber electrochemical detector for high-perfomance liquid-chromatography, J. Chromatogr. 298 (1984) 225–230. [8] W.L. Caudill, J.O. Howell, R.M. Wightman, Flow-rate independent amperometric cell, Anal. Chem. 54 (1982) 2532–2535. [9] J.M. Zadeii, R. Mitchell, T. Kuwana, Glassy-carbon microband electrochemical detector for high-performance liquid-chromatography, Electroanalysis 2 (1990) 209–215. [10] M. Takahashi, M. Morita, O. Niwa, H. Tabei, Highly sensitive high-performance liquid-chromatography detection of catecholamine with interdigitated array microelectrodes, J. Electroanal. Chem. 335 (1992) 253–263. [11] H. Tabei, M. Takahashi, S. Hoshino, O. Niwa, T. Horiuchi, Subfentomole detection of catecholamine with interdigitated array carbon microelectrodes in HPLC, Anal. Chem. 66 (1994) 3500–3502. [12] O. Niwa, H. Tabei, B.P. Solomon, F.M. Xie, P.T. Kissinger, Improved detection limit for catecholamines using liquid-chromatography electrochemisty with a carbon interdigitated array microelectrode, J. Chromatogr. B 670 (1995) 21–28.

M. Komendová et al. / J. Chromatogr. A 1509 (2017) 171–175 [13] C. Terashima, H. Tanaka, M. Furuno, Using an electrochemical detector with a carbon interdigited-array microelectrode for capillary-column liquid chromatography, J. Chromatogr. A 828 (1998) 113–120. [14] F. Svec, J.M.J. Fréchet, Continuous rods of macroporous polymer as high-performance liquid-chromatography separation media, Anal. Chem. 64 (1992) 820–822. [15] A. Greiderer, L. Trojer, C.W. Huck, G.K. Bonn, Influence of the polymerisation time on the porous and chromatographic properties of monolithic poly(1,2-bis(p-vinylphenyl))ethane capillary columns, J.Chromatogr. A 1216 (2009) 7747–7754. [16] I. Nischang, I. Teasdale, O. Bruggemann, Towards porous polymer monoliths for the efficient, retention-independent performance in the isocratic separation of small molecules by means of nano-liquid chromatography, J. Chromatogr. A 1217 (2010) 7514–7522. ˇ ríková, J. Urban, Cross-linker effects on the ˇ [17] M. Stanková, P. Jandera, V. Skeˇ separation efficiency on (poly)methacrylate capillary monolithic columns. Part II. Aqueous normal-phase liquid chromatography, J. Chromatogr. A 1289 (2013) 47–57. ˇ ríková, J. Urban, Highly stable surface modification of hypercrosslinked [18] V. Skeˇ monolithic capillary columns and their application in hydrophilic interaction chromatography, J. Sep. Sci. 36 (2013) 2806–2812. ˇ ríková, Effect of hypercrosslinking conditions on pore size [19] J. Urban, V. Skeˇ distribution and efficiency of monolithic stationary phases, J. Sep. Sci. 37 (2014) 3082–3089.

175

˚ M. Komendová, J. Urban, Development of an online solid-phase [20] S. Janku, extraction with liquid chromatography method based on polymer monoliths for the determination of dopamine, J. Sep. Sci. 39 (2016) 4107–4115. [21] L.B. Zahodne, H.H. Fernandez, Pathophysiology and treatment of psychosis in Parkinson’s disease − A review, Drugs Aging 25 (2008) 665–682. [22] K.D. Alex, E.A. Pehek, Pharmacologic mechanisms of serotonergic regulation of dopamine neurotransmission, Pharmacol. Ther. 113 (2007) 296–320. [23] D.T. Marc, J.W. Ailts, D.C.A. Campeau, M.J. Bull, K.L. Olson, Neurotransmitters excreted in the urine as biomarkers of nervous system activity: validity and clinical applicability, Neurosci. Biobehav. Rev. 35 (2011) 635–644. [24] M. Marazuela, L. Agui, A. Gonzalez-Cortes, P. Yanez-Sedeno, J.M. Pingarron, Microcylinder polymer modified electrodes as amperometric detectors for liquid chromatographic analysis of catecholamines, Electroanalysis 11 (1999) 1333–1339. [25] H. Echizen, R. Itoh, T. Ishizaki, Adenosine and dopamine simultaneously determined in urine by reversed-úhsae HPLC, witho online measurement of ultaviolet absorbance and electrochemical detection, Clin. Chem. 35 (1989) 64–68. [26] J. Van Schoors, K. Maes, Y. Van Wanseele, K. Broeckhoven, A. Van Eeckhaut, Miniaturized ultra-high performance liquid chromatography coupled to electrochemical detection: investigation of system performance for neurochemical analysis, J. Chromatogr. A 1427 (2016) 69–78.