2509
Electrophoresis 2015, 36, 2509–2515
Haiyun Zhai1 Kaisong Yuan1 Xiao Yu1 ∗ Zuanguang Chen2 Zhenping Liu1 Zihao Su1 1 College
of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, P. R. China 2 School of Pharmaceutical Science, Sun Yat-sen University, Guangzhou, P. R. China
Received March 12, 2015 Revised June 11, 2015 Accepted June 12, 2015
Research Article
A simple and compact fluorescence detection system for capillary electrophoresis and its application to food analysis A novel fluorescence detection system for CE was described and evaluated. Two miniature laser pointers were used as the excitation source. A Y-style optical fiber was used to transmit the excitation light and a four-branch optical fiber was used to collect the fluorescence. The optical fiber and optical filter were imported into a photomultiplier tube without any extra fixing device. A simplified PDMS detection cell was designed with guide channels through which the optical fibers were easily aligned to the detection window of separation capillary. According to different requirements, laser pointers and different filters were selected by simple switching and replacement. The fluorescence from four different directions was collected at the same detecting point. Thus, the sensitivity was enhanced without peak broadening. The fluorescence detection system was simple, compact, low-cost, and highly sensitive, with its functionality demonstrated by the separation and determination of red dyes and fluorescent whitening agents. The detection limit of rhodamine 6G was 7.7 nM (S/N = 3). The system was further applied to determine illegal food dyes. The CE system is potentially eligible for food safety analysis. Keywords: Capillary electrophoresis / Fluorescence detection system / Four-branch optical fiber / Y-style optical fiber DOI 10.1002/elps.201500265
1 Introduction CE is a separation and analysis technique, which gets its name from the process whereby charged particles migrate across an electric field. It has attracted global attention due to low cost, short analysis time, high separation efficiency, and requirement of only minute amounts of sample. To date, CE has been widely used in pharmaceutical [1], food [2] and biological sample analyses [3]. Optical detection [4], electrochemical detection [5], and other methods such as MS [6] or NMR [7] have been employed for CE to analyze various samples and to improve the sensitivity of the process. Particularly, LIF is one of the most popular detection techniques, due to its high sensitivity and S/N. Many light sources have been used for LIF, including lamps [8], lasers [9], and LEDs [10]. The most commonly used were lasers, such as Ar+ laser, He-Cd laser, and He-Ne laser [11,12]. They have been widely used for LIF due to a high excitation capacity and emission of collimated light combined with a small beam size. In other words, it provides a powerful excitation source without any additional lens. However, most
Correspondence Dr. Haiyun Zhai, College of Pharmacy, Guangdong Pharmaceutical University, Guangzhou 510006, P. R. China E-mail: [email protected] Fax: +86-20-3935-2129
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
of these commercially available lasers are high cost, large in size, and high in energy consumption. A smaller and less expensive option is LEDs, and they have been highlighted in recent research [13–16]. Unfortunately, the intensity of LED light is rather feeble and thus causes low sensitivity, and the broad bandwidth (20–100 nm) leads to high baseline noise in the electropherogram. Besides, it is also difficult to focus the LED light beam onto a small spot to match the inner diameter of a fused-silica capillary. This is due to LED’s wide divergence angle of light. In comparison, laser diodes are more appropriate for microscale fluorescence detection [17] owing to low cost, small size, high intensity, low power consumption, and simple power supply requirements. It also offers high excitation power, collimated light in a small beam size, high stability, and long lifetimes. Moody et al. used a red (820 nm) laser diode as the excitation source to detect indocyanine green [18], and the LOD was 0.388 M. Melanson et al. used a violet (405 nm) laser diode [19] for indirect LIF detection, with an LOD of 0.1 M for porphyrin tetrakis-(4-sulfophenyl) porphine. Besides being an excitation light source, the optical system of an LIF detector is also a key factor in sensitivity and device fabrication. Orthogonal [20] and confocal designs [21] are
∗
Additional corresponding [email protected]
author:
Dr.
Xiao
Yu
E-mail:
Colour Online: See the article online to view Figs. 1–4 in colour.
www.electrophoresis-journal.com
2510
H. Zhai et al.
most commonly used for the on-column LIF detection for CE. However, they both require accurate mechanical alignment of optical elements. It is also sensitive to mechanical vibrations and drifts, and only a small number of detection configurations are suitable. Recently, optical fibers that have been employed to transmit excitation light [22–24] or to collect fluorescence [25, 26] play an important role in the improvement of optical devices. It allows construction of compact and sensitive detectors suitable for miniaturized systems, including microfluidic devices or capillary techniques. Due to its being bendable and leak-proof, optical fibers can reduce the size of an optical system, flexibly locate some optical elements, eliminate the interference of surrounding light, and keep the detector electronics away from the CE unit to insulate from disturbance by high-voltage power. In multiple detection modes, optical fibers also have been used in the combination of detectors and in single points of detection. Since each different detection mode has its own unique advantage, the combination of different detection modes will allow the simultaneous detection of compounds with different properties present in complex samples. Most combinations involve combining two common modes of detection: absorbance-fluorescence [27], and conductivity-fluorescence [28]. Yang et al. transmitted excitation light and collected the emission by using optical fibers [29]. Thus, a device with compact structure, low cost, and high stability was constructed. However, the inner diameter of optical fiber is generally so small (probably a few hundred microns) that limited fluorescence can be collected, ultimately restricting the enhancement of detection sensitivity. Given the emission of fluorescence in all directions, collecting fluorescence in only one direction is bound to decrease the fraction of emitted light, which also influences further increase of detection sensitivity. In other words, when part of the emission light is lost or fails to reach the photomultiplier tube (PMT), that lowers fluorescence collection efficiency with a concomitant decrease in the sensitivity of detection. To this end, Beckman Coulter, Inc. (Fullerton, USA) [30], made a commercialized design by inserting a spherical mirror close to and partly around the detection window to reflect the emission light into the optical fiber. The mirror may therefore reflect quite a few of the scattered excitation lights into the detector at the same time. Also, this kind of spherical mirror reflects light in an immense area randomly. Moreover, finding a proper position to place the mirror, or fabricating and fixing the mirror to the device, is rather tedious. Yang et al. used an optical fiber, with the flat end modified to a spherical end, to collect fluorescence [31], similar to the collection of fluorescence signals by using the microscope objective. Such an arrangement may involve some loss of light owing to the limited acceptance angle and small core diameter of fiber, and modifying the spherical end remains tricky. In this study, we addressed the above-mentioned problems by replacing the single optical fiber with a four-branch optical fiber while collecting the emission light. After collection of the emission light from four directions by the branches part, several times of the emission light can be collected com C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 2509–2515
pared with that using the single optical fiber without peak broadening. A green or a violet laser pointer, was used as the excitation source. The optical fibers were easily aligned to the detection window of the separation capillary through the guide channels of a home-made PDMS detection cell. The analytical performance of this CE-LIF system was demonstrated by separating and detecting orchil and fluorescent whitening agent in food safety analysis.
2 Materials and methods 2.1 Chemicals and CE apparatus Rhodamine 6G, safranine T, rose bengal, phloxine B, fluorescent brightener VBL, and fluorescent brightener FBA 351 were purchased from Aladdin Reagent Company (Shanghai, China). PDMS and Sylgard 184 elastomer curing agent were obtained from Dow Corning (San Diego, CA, USA). Other analytical-grade solvents were purchased from Guangzhou Reagents Company (Guangzhou, China), and food samples such as chili, red dates, and wheat flour were purchased from a local supermarket. Experiments were performed on a homemade CE system which can provided constant-potential direct currents of 0 to 20 kV and 0 to −20 kV for separation. Fused-silica capillaries (75 m id) were purchased from Ruifeng Chromatographic Co., Ltd. (Hebei, China). PMT was purchased from Hamamatsu Photonics Trade (H10722-01, China) Co., Ltd.
2.2 Detection system The optical arrangement for LIF detection is depicted schematically in Fig. 1. A green or a violet laser diode (applied voltage, +3 V; 100 mW; wavelength, 532 nm or 405 nm) was removed from the laser pointers, and was used as the excitation light source, controlled by a toggle switch (1MS1T61B11M2QES Q11, Dailywell Electronics Co., Ltd, Taiwan). The toggle switch is characterized by the presence of some type of handle or lever that opens and closes an electric circuit. Its circuits control the two laser pointers, is as also depicted in Fig. 1. To avoid overheating of the laser pointer, the laser pointer was coated with heat-conducting silicone grease and an aluminum cooled sink (homemade). A cooling fan is also used to quicken the heat dissipation. A Y-style optical fiber with an external diameter of 1000 m (1/2-pigtail-S10030cm-SMA, Beijing Scitlion Technology Co., Ltd, Beijing) coated with a metal sleeve was used to transmit the excitation light to the detection cell. Coupling of laser diodes to optical fiber was depicted in Fig. 1. Cylindrical polymer materials were used to combine the laser diode and the optical fiber effectively. The four screws were used to fine tune both the laser diode and the optical fiber to ensure they were at the same level. Two branches of the Y-style fiber (100 m id) were connected via SMA-905 with the green and violet www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2015, 36, 2509–2515
2511
Figure 1. Schematic illustration of the detection system. (a) green laser pointer; (b) violet laser pointer; (c) Y-style optical fiber; (d) PDMS cell; (e) four-branch optical fiber; (f) SMA connector; (g) screws; (h) metal cover of the PMT; (i) optical filter; (j) PMT; (k) data acquisition board; (l) personal computer; (m) reservoirs; (n) high voltage;(o) toggle switch; (p) coupling of laser diodes to optical fiber; (q) the screws to fine tune the laser diode and the optical fiber.
laser diodes, respectively. The common part of the fiber, as a pigtail, was connected to the detection cell. An optical fiber with four branches (1/6-pigtail-S10030 cm-SMA, Beijing Scitlion Technology Co., Ltd., Beijing) was used to collect the fluorescence in four various directions. Being all pigtails (100 m id; 140 m od), the four branches of the optical fiber were connected with the detection cell. The common part of the fiber was connected via SMA-905 with PMT. The illumination system was also equipped with a longpass filter (LP-580 and LP-430, Shenyang, China) to eliminate the interference from the excitation light. The signal from PMT was acquired by a Data Acquisition Board (NI USB6343, X Series Multifunction DAQ, USA).
2.3 Construction of PDMS detection cell A simplified PDMS detection cell with guide channels was designed and employed to align the optical fibers easily with the detection window of the separation capillary. The PDMS detection cell was fabricated by sealing two identical PDMS layers (the top and bottom layers) with the following procedure as we previously reported [32]. First, 3D stereogram layout of the cell was designed by AutoCAD, based on which a copper mold was fabricated by a commercially available fine carving. The detection cell was fabricated from PDMS produced by replica molding. PDMS elastomer and Sylgard 184 curing agent were mixed (10:1, w/w), and degassed thoroughly by a vacuum pump for 20 min. Subsequently, the mixture was poured into the copper mold and cured in an 80°C oven for 30 min, yielding the PDMS layer. After the optical fibers and capillary were bound to the channels of the two PDMS layers, they were put together between two Perspex sheets and then fixed by four screws instead of PDMS bonding to finish the PDMS cell. As shown in Fig. 2, the PDMS layer is a regular hexahedron with the side length of 17.3 mm and the thinness of C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
5 mm. The capillary channel was 30 mm in length, 375 m in width, and 187 m in depth. The excitation light optical fiber channel was 15 mm in length, 1 mm in width, and 500 m in depth. The five emission light optical fiber channels were all 15 mm in length, 140 m in width, and 70 m in depth. All of the channels intersected the center of the regular hexahedron and each of the channels also intersected the side in the middle. Figure 2 also shows the assembling procedure of the fourbranch optical fiber, Y-style optical fiber, and capillary to the detection cell. Obviously, there were two different kinds of fluorescence collection angles (defined as the angle between the excitation light channel and the emission light channel) in our detector, one (two branches of the emission light optical fiber) being 45° and the other (the other two branches of the emission light optical fiber) being 135°. The distance between the capillary and the branches of the optical fiber, with the fluorescence collection angle of 45°, was 200 m. When the angle was 135°, the distance was 0 m.
2.4 Sample preparation Stock solutions of rhodamine 6G, safranine T, rose bengal, phloxine B, fluorescent brightener VBL and fluorescent brightener FBA 351 (10 g mL−1 ) were prepared for the separation study by dissolving 1.0 mg of the compounds with methanol in 100 mL volumetric flasks. The series of all working standard solutions were then obtained by diluting the stock solutions with methanol to give the desired concentrations. All the food samples were broken into pieces by a pulverizer and extraction method was established according to a related method [33]. To determine the chili sample, 2.0 g prehomogenized chili powders were placed into a 50 mL centrifuge tube, vortexshaken for 3 min with 25 mL of methanol, and then stirred in an ultrasonic bath for 10 min. The supernatant was collected www.electrophoresis-journal.com
2512
H. Zhai et al.
Electrophoresis 2015, 36, 2509–2515
Figure 2. Schematic illustration of the fabrication for detection cell. (a) excitation optical fiber channel; (b)(d)(e)(f)(g) emission optical fiber channel; (c) capillary channel; (a’) excitation optical fiber; (b’)(d’)(f’)(g’) four branches of the emission optical fiber; (c’) capillary.
and made up to 2 mL with methanol. The extract solution was filtered through a syringe filter in a tube for CE use. So the food sample solvent was methanol. Wheat flour sample was prepared with the procedure the same as that of chili.
2.5 CE procedure Prior to injection, the capillary was pretreated by being rinsed with 1.0 M HNO3 for 20 min. Between two consecutive injections, the capillary was sequentially flushed with 0.1 M NaOH, water, and separation buffer (5 min each time). Sample solutions were injected into the capillary by gravity for 10 s, with the sample vials lifted to a height of 20 cm. An applied voltage of 19 kV or −19 kV was applied to sample separation. The total length of the capillary was 50 cm and the effective length was 40 cm. For the determination of rhodamine 6G, safranine T, rose bengal, and phloxine B, separation was performed by using 100 mM Na2 B4 O7 (pH 9.2) as the running buffer. The wavelength of exciting light was 532 nm, with a filter (LP 580) applied. For the detection of fluorescent brightener VBL and fluorescent brightener FBA 351, CTAB was selected and combined in reversed migration-micellar electrokinetic chromatography (RM-MEKC). This is characterized by anionic micelles moving faster than the EOF to optimize separation [34]. To increase the solubility of organic analytes, ACN has also been added into the running buffer [35]. Thus, 100 mM Na2 B4 O7 -0.2 mM CTAB-15% acetonitrile (pH 9.2) was used as the running buffer. The wavelength of the exciting light was 405 nm, also with a filter (LP-430) applied.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3 Results and discussion 3.1 Design of LIF system To construct a sensitive, simple, low-cost, and facile optical arrangement for fluorescence detection in CE, several special designs were made in this study. The fluorescence detection system included laser pointers, a Y-style optical fiber, a four-branch optical fiber, long-pass filters, PMT, and a PDMS detection cell, which demonstrated a special optical configuration compared to conventional fluorescence detection. Sensitivity or S/N is generally augmented by enhancing signals while reducing noise. Signal intensity is influenced by excitation light intensity, detection distance, and the collection efficiency of the emission light, etc. A laser pointer, which is intense and highly stable, is an appropriate excitation source. A green or a violet laser pointer was used in the system to extend the detection range of the sample. A Y-style optical fiber was used to transmit two branches of excitation light, with one branch for the green and the other one for the violet. We also used a four-branch optical fiber to collect the emission light in four different directions, so several times of the fluorescence were collected and the signal intensity was increased. As to noise, consideration should be given to the detection distance, the fluorescence collection angle, and the effectiveness of the filter. After being collected by the four-branch optical fiber, the emission light enters PMT, before which an optical filter should be used to filter most reflected and scattered lights. Also, detection distance affects both sensitivity and noise. For a simple, low-cost, and compact device, a laser pointer is a good choice because it is cheap, commercially available, and small sized. A more compact optical system is allowed
www.electrophoresis-journal.com
Electrophoresis 2015, 36, 2509–2515
Figure 3. Effect of different number of branches. Labeling a’-g’ are the same as Fig. 2. Sample: 130 nM rhodamine 6G; operation conditions: injection for 10 s at height of 20 cm; sepration voltage: 19 kV; running buffer: 100 mM Na2 B4 O7 (pH 9.2); excitation/emission wavelength: 532 nm/580 nm.
by using optical fiber that can flexibly locate optical elements. To fix the capillary and optical fibers, a simplified detection PDMS cell with guide channels was designed, with which the optical fibers were easily aligned to the detection window. Lenses, which need extra fixing to devices and which occupy space for light focusing, were not used herein. In addition, as depicted in Fig. 1, a miniature optical filter (diameter, 1.5 cm) was inserted into the PMT (in the front: a metal cover fixed by screws) instead of fixing extra devices, to further miniaturize the system.
3.2 Optimization of the detector 3.2.1 Effects of branch numbers To verify the effectiveness of our four-branch optical fiber, the effects of branch numbers on the fluorescence emission intensities and noise were evaluated (Fig. 3). With increasing number of branches from one to four, the emission intensities rose continuously. Meanwhile, the noise increased, at a significantly slower speed than the signal did though. In other words, increasing the number of branches raised S/N of the fluorescence detector. Different directions of the emission optical fibers contributed differently to the increase of S/N. The branches of optical fibers with the fluorescence collection angle of 135° contributed more to S/N than that with 45° did, which can mainly be ascribed to various detection distances between the capillary and the emission optical fiber instead of various detection angles. Since the fifth branch of optical fiber with the fluorescence collection angle of 180° can further increase S/N, a five-branch optical fiber was also tested. However, in this case, the collected signal exceeded the acceptable range of PMT. Probably, when the excitation and emission optical fibers were in the rectilinear direction, massive excitation light traveled through the capillary, and entered the emission C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
2513
Figure 4. Effect of detection distance. Labeling a’-g’ are the same as Fig. 2. Sample: 130 nM rhodamine 6G; operation conditions are the same as Fig. 3.
optical fiber directly. Though the long-pass filter located before PMT could block more than 90% of the excitation light, only a small proportion of the strong excitation light received by PMT sufficed to induce a high noise, or even exceeded the full-scale value of PMT. Hence, four-branch optical fiber, which gave the highest S/N in our fluorescence detector, was finally selected. 3.2.2 Effect of detection distance The effects of distances between the capillary and the optical fibers on the fluorescence emission intensities and noise were also studied (Fig. 4). Excitation light may produce a certain degree of light reflection and scattering on the wall of the capillary, which was one of the important components to the noise. As shown in Fig. 4, the detection cell can be divided into two areas (zone A and zone B) by using the capillary as the separatrix, with far more noise in zone A than that in zone B. Therefore, for the branches with 45° fluorescence collection angles (branches in zone A), increasing the noise contributed more to S/N than increasing the signal did when the detection distance was less than 200 m. On the contrary, when the detection distance was more than 200 m, decrease of the signal was more severe than decrease of the noise. For the branches with 135° fluorescence collection angles (branches in zone B), reducing the signal contributed more to S/N than reducing the noise did when the detection distance was more than 0 m. Obviously, detection distance evidently affected the sensitivity. Thus, the detection distances of 200 and 0 m were selected for the branches in zone A and zone B, respectively.
3.3 Performance of the detection system The performance of the LIF detector, with the green (532 nm) laser pointer as the excitation source and the longpass filter (580 nm) reducing the background signal, was evaluated by analyzing rhodamine 6G (exmax = 525 nm, www.electrophoresis-journal.com
2514
H. Zhai et al.
Electrophoresis 2015, 36, 2509–2515
Table 1. Performance of the present laser-induced fluorescence detector
Analytes
LOD (nM)
Linearity range (nM)
Correlation
Theoretical plates
rhodamine 6G safranine T rose bengal phloxine B fluorescent brightener VBL fluorescent brightener FBA 351
7.7 109.0 63.3 25.2 57.2 68.9
15–850 150–1000 100–1000 50–850 150–1000 150–1000
0.999 0.999 0.998 0.999 0.998 0.998
40 974 40 903 349 893 334 441 14 154 57 390
Figure 5. Electrophorogram of dyes. (A) Peak identities: (a) rhodamine 6G; (b) safranine T; (c) rose Bengal; (d) phloxine B; operation conditions were the same as Fig. 3; (B) Peak identities: (a) fluorescent brightener VBL; (b) fluorescent brightener FBA 351; operation conditions: injection for 10 s at height of 20 cm; sepration voltage: -19 kV; running buffer: 100 mM Na2 B4 O7 -0.2 mM CTAB-15% acetonitrile.
emmax = 550 nm), safranine T (exmax = 522 nm, emmax = 588 nm), rose bengal (exmax = 550 nm, emmax = 570 nm), and phloxine B (exmax = 532 nm, emmax = 560 nm). In the meantime, fluorescent brightener VBL (exmax = 344 nm, emmax = 429 nm) and fluorescent brightener FBA 351 (exmax = 352 nm, emmax = 427 nm) were also used to evaluate the performance of the LIF detector, with the violet (405 nm) laser pointer as the excitation source and the long-pass filter (430 nm) cutting off the background signal. The S/N was determined from the peak height divided by the standard deviation of the baseline noise (n = 7). As shown in Table 1, LODs are in the range of 7.7–109.0 nM (S/N = 3) under the optimized conditions, giving similar or higher sensitivity than those reported previously [29, 36–38]. Typical electropherogram of these fluorophores is shown in Fig. 5, exhibiting good repeatability with relative standard deviations (RSDs) of 2.3–3.4% and 2.8–2.9% for peak height and migration time, respectively. Fluorescent dyes, which belong to illegal food additives, have become a global concern recently [33,39–41]. They might be harmful if eaten by human, irritate the skin, eyes and respiratory tract, and even lead to teratogenesis and cancer. Hence, it is meaningful to establish an effective and sensitive method for the determination of fluorescent dyes in food. Qi et al. [42] developed an HPLC coupled with fluorescence detection method to rapidly, simply, and sensitively determine rhodamine B in chili-containing products. However, CE is C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
superior to these methods, and it requires significantly less injection volume (nanoscale) compared to HPLC does (microscale) and saves organic reagents [43]. To show the potential applicability of this proposed instrumental setup, rhodamine 6G, safranine T, rose bengal, phloxine B, fluorescent brightener VBL, and fluorescent brightener FBA 351 were determined and separated. Rhodamine 6G, safranine T, rose bengal, and phloxine B have been illegally added in food like chili for color improvement. Also, fluorescent brightener VBL and fluorescent brightener FBA 351 have been used to whiten food such as wheat flour. The recoveries of the spiked samples ranged between 97.0 and 103%, with RSDs ranging from 1.6 to 3.6%. LODs (S/N = 3) of the six fluorescent dyes were in the range of 3.68–64.4 g kg−1 , which met the limits of 500–1000 g kg–1 for all carcinogenic dyes set by the European Community [44].
4 Concluding remarks A novel fluorescence detection system for CE was designed and prepared. This system possessed a special optical configuration compared to those of conventional fluorescence detection systems, which adopted two laser pointers as excitation source to meet the demand of analytes with different excitation wavelengths, and a Y-style optical fiber was used to transport the excitation light. In order to increase the www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2015, 36, 2509–2515
2515
detection sensitivity, the system utilized a four-branch optical fiber to collect the emission light in different directions. Moreover, a PDMS cell was designed and built to combine capillary and fibers readily. Then, an optical filter was inserted into PMT to filter the reflected and scattered lights. To further optimize the sensitivity, four major influencing factors, i.e. number of optical fiber branches, detection distance, as well as performance of laser diodes and excitation filter, were studied and optimized. This design rendered the system easily home-built, low-cost and compact, accompanied by sensitive responses to the selected fluorescent dyes. Furthermore, this device was also applied to analyze these dyes in food samples. In summary, this system is promising for food safety analysis.
[16] Hurth, C., Lenigk, R., Zenhausern, F., Appl. Phys. B-Lasers O. 2008, 93, 693–699.
The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (NSFC, Grant No. 21005021) and Guangdong Provincial Science and Technology Project (No.2013B040402010 and No.2010B0203106010).
[24] Yang, X. P., Qian, F., Xie, L. X., Yang, X. C., Cheng, X. M., Choi, M. M. F., Electrophoresis 2014, 35, 762–769.
[17] Park, J., Kim, S., Microchem. J. 2011, 9, 103–107. [18] Moody, E. D., Viskari, P. J., Colyer, C. L., J. Chromatogr. B 1999, 729, 55–64. [19] Melanson, J. E., Boulet, C. A., Lucy, C. A., Anal. Chem. 2001, 7, 1809–1813. [20] Nagaraj, S., Karnes, H. T., Instrum. Sci. Technol. 2000, 28, 119–129. [21] Yang, F., Li, X. C., Zhang, W., Pan, J. B., Chen, Z. G., Talanta 2011, 84, 1099–1106. [22] Yang, X. P., Yuan, H. Y., Wang, C. L., Zhao, S. L., Xiao, D., Choi M. M. F., Electrophoresis 2007, 28, 3105–3114. [23] Mazurczyk, R., Vieillard, J., Bouchard, A., Hannes, B., Krawczyk, S., Sensor. Actuat. B-Chem. 2006, 11, 11–19.
[25] Geng, X. H., Wu, D. P., Guan, Y. F., Talanta 2012, 88, 463–467. [26] Butler, P. A. G., Mills, B., Hauser, P. C., Analyst 1997, 122, 949–953. [27] Caslavska, K., Gassmann, E., Thormann, W., J. Chromatogr. A 1995, 709, 147–156.
The authors have declared no conflict of interest.
[28] Tan, F., Yang, B. C., Guan, Y. F., Anal. Sci. 2005, 21, 583–585.
5 References ˜ [1] Nevado, J. J. B., Penalvo, G. C., Dorado R. M. R., Robledo, V. R., J. Pharmaceut. Biomed. 2014, 92, 211–219.
[29] Yang, X. B., Yan, W. P., Bai, H. M., Lv, H. F., Liu, Z. H., Optik 2012, 123, 2126–2130. [30] Kork, B. J. D., Jong, G. J. D., Somsen, G. W., Anal. Chim. Acta 2013, 76, 13–33.
[2] Dabrowska, M., Starek, M., Food Chem. 2014, 142, ˛ 220–232.
[31] Yang, B. C., Guan, Y. F., Talanta 2003, 59, 509–514.
[3] Zhu, Z. F., Lu, J. J., Liu, S. R., Anal. Chim. Acta 2012, 709, 21–31.
[32] Zhai, H. Y., Li, J. M., Chen, Z. G., Su, Z. H., Liu, Z. P., Yu, X., Microchem. J. 2014, 114, 223–228.
¨ [4] Gotz, S., Karst, U., Sensor. Actuat. B-Chem. 2007, 123, 622–627.
[33] Wang, W. J., Zhu, X. S., Yan, C. G., Food Chem. 2013, 141, 2207–2212.
[5] Mark, J. J. P., Scholz, R., Matysik, F. M., J. Chromatogr. A 2012, 1267, 45–64.
´ Y., Rodr´ıguez, R., Manes, ˜ [34] Pico, J., Trac-Trend. Anal. Chem. 2003, 22, 133–151.
[6] Wang, T., Zhang, Q., Zhang, Y., Kang, J., J. Chromatogr. A 2014, 133, 188–193.
[35] Yang, Y. Z., Boysen, R. I., Hearn, M. T. W., J. Chromatogr. A 2004, 1043, 91–97.
[7] Li, Y., Lacey, M. E., Sweedler, J. V., Webb, A. G., J. Magn. Reson. 2003, 162, 133–140.
[36] Kort, B. J. D., Jong, G. J. D., Somsen, G. W., Electrophoresis 2010, 31, 2861–2868.
[8] Yan, Q., Chen, R. S., Cheng, J. K., Anal. Chim. Acta 2006, 555, 246–249.
ˇ a, ´ M., Akther, M., Maaskant, P., Brabazon, D., [37] Vaculovicov Macka, M., Anal. Chim. Acta 2015, 871, 85–92.
[9] Ban, E., Song, E. J., J. Chromatogr. B 2013, 929, 180–186.
[38] Banerjee, A., Pais, A., Papautsky, I., Klotzkin, D., Ieee Sens. J. 2008, 8, 621–627.
[10] Bui, D. A., Hauser, P. C., Anal. Chim. Acta 2015, 853, 46–58.
[39] Xiao, N., Deng, J., Huang, K. H, Ju, S., Hu, C. H., Liang, J., Spectrochim. Acta A 2014, 128, 312–318.
[11] Lin, Y. F., Wang, Y. C., Chang, S. Y., J. Chromatogr. A 2008, 1188, 331–333. ´ ´ ´ P., Ryvolova, ´ M., Havel, J., Preisler, [12] Vrabel, P., Taborsk y, J., J. Lumin. 2006, 118, 283–292. [13] Macka, M., Piasecki, T., Dsgupta, P. K., Ann. Rev. Anal. Chem. 2014, 7, 183–207.
[40] Yoshioka, N., Ichihashi, K., Talanta 2008, 7, 1408–1413. [41] Jiang, D., Chen, L., Fu, W. S., Qiu, H. Q., J. Sep. Sci. 2015, 38, 605–611. [42] Qi, P., Lin, Z. H., Li, J. X., Wang, C. L., Meng, W. W., Hong, H., Zhang, X. W., Food Chem. 2014, 16, 98–103.
86,
´ [43] Perez-Ruiz, T., Mart´ınez-Lozano, C., Sanz, A., Bravo, E., Chromatographia 1998, 48, 263–267.
[15] Bui, D. A., Hauser, P. C., Anal. Chim. Acta 2015, 853, 46–58.
[44] EC, Directive of the European Parliament and of the council 94/36/EC of June 30, 1994 on colours for use in foodstuffs. Off. J. 1994, L237, 13.
[14] Prikryl, J., Foret, 11951–11956.
F.,
Anal.
Chem.
2014,
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
Electrophoresis 2015, 36, 2509–2515
Haiyun Zhai1 Kaisong Yuan1 Xiao Yu1 ∗ Zuanguang Chen2 Zhenping Liu1 Zihao Su1 1 College
of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, P. R. China 2 School of Pharmaceutical Science, Sun Yat-sen University, Guangzhou, P. R. China
Received March 12, 2015 Revised June 11, 2015 Accepted June 12, 2015
Research Article
A simple and compact fluorescence detection system for capillary electrophoresis and its application to food analysis A novel fluorescence detection system for CE was described and evaluated. Two miniature laser pointers were used as the excitation source. A Y-style optical fiber was used to transmit the excitation light and a four-branch optical fiber was used to collect the fluorescence. The optical fiber and optical filter were imported into a photomultiplier tube without any extra fixing device. A simplified PDMS detection cell was designed with guide channels through which the optical fibers were easily aligned to the detection window of separation capillary. According to different requirements, laser pointers and different filters were selected by simple switching and replacement. The fluorescence from four different directions was collected at the same detecting point. Thus, the sensitivity was enhanced without peak broadening. The fluorescence detection system was simple, compact, low-cost, and highly sensitive, with its functionality demonstrated by the separation and determination of red dyes and fluorescent whitening agents. The detection limit of rhodamine 6G was 7.7 nM (S/N = 3). The system was further applied to determine illegal food dyes. The CE system is potentially eligible for food safety analysis. Keywords: Capillary electrophoresis / Fluorescence detection system / Four-branch optical fiber / Y-style optical fiber DOI 10.1002/elps.201500265
1 Introduction CE is a separation and analysis technique, which gets its name from the process whereby charged particles migrate across an electric field. It has attracted global attention due to low cost, short analysis time, high separation efficiency, and requirement of only minute amounts of sample. To date, CE has been widely used in pharmaceutical [1], food [2] and biological sample analyses [3]. Optical detection [4], electrochemical detection [5], and other methods such as MS [6] or NMR [7] have been employed for CE to analyze various samples and to improve the sensitivity of the process. Particularly, LIF is one of the most popular detection techniques, due to its high sensitivity and S/N. Many light sources have been used for LIF, including lamps [8], lasers [9], and LEDs [10]. The most commonly used were lasers, such as Ar+ laser, He-Cd laser, and He-Ne laser [11,12]. They have been widely used for LIF due to a high excitation capacity and emission of collimated light combined with a small beam size. In other words, it provides a powerful excitation source without any additional lens. However, most
Correspondence Dr. Haiyun Zhai, College of Pharmacy, Guangdong Pharmaceutical University, Guangzhou 510006, P. R. China E-mail: [email protected] Fax: +86-20-3935-2129
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
of these commercially available lasers are high cost, large in size, and high in energy consumption. A smaller and less expensive option is LEDs, and they have been highlighted in recent research [13–16]. Unfortunately, the intensity of LED light is rather feeble and thus causes low sensitivity, and the broad bandwidth (20–100 nm) leads to high baseline noise in the electropherogram. Besides, it is also difficult to focus the LED light beam onto a small spot to match the inner diameter of a fused-silica capillary. This is due to LED’s wide divergence angle of light. In comparison, laser diodes are more appropriate for microscale fluorescence detection [17] owing to low cost, small size, high intensity, low power consumption, and simple power supply requirements. It also offers high excitation power, collimated light in a small beam size, high stability, and long lifetimes. Moody et al. used a red (820 nm) laser diode as the excitation source to detect indocyanine green [18], and the LOD was 0.388 M. Melanson et al. used a violet (405 nm) laser diode [19] for indirect LIF detection, with an LOD of 0.1 M for porphyrin tetrakis-(4-sulfophenyl) porphine. Besides being an excitation light source, the optical system of an LIF detector is also a key factor in sensitivity and device fabrication. Orthogonal [20] and confocal designs [21] are
∗
Additional corresponding [email protected]
author:
Dr.
Xiao
Yu
E-mail:
Colour Online: See the article online to view Figs. 1–4 in colour.
www.electrophoresis-journal.com
2510
H. Zhai et al.
most commonly used for the on-column LIF detection for CE. However, they both require accurate mechanical alignment of optical elements. It is also sensitive to mechanical vibrations and drifts, and only a small number of detection configurations are suitable. Recently, optical fibers that have been employed to transmit excitation light [22–24] or to collect fluorescence [25, 26] play an important role in the improvement of optical devices. It allows construction of compact and sensitive detectors suitable for miniaturized systems, including microfluidic devices or capillary techniques. Due to its being bendable and leak-proof, optical fibers can reduce the size of an optical system, flexibly locate some optical elements, eliminate the interference of surrounding light, and keep the detector electronics away from the CE unit to insulate from disturbance by high-voltage power. In multiple detection modes, optical fibers also have been used in the combination of detectors and in single points of detection. Since each different detection mode has its own unique advantage, the combination of different detection modes will allow the simultaneous detection of compounds with different properties present in complex samples. Most combinations involve combining two common modes of detection: absorbance-fluorescence [27], and conductivity-fluorescence [28]. Yang et al. transmitted excitation light and collected the emission by using optical fibers [29]. Thus, a device with compact structure, low cost, and high stability was constructed. However, the inner diameter of optical fiber is generally so small (probably a few hundred microns) that limited fluorescence can be collected, ultimately restricting the enhancement of detection sensitivity. Given the emission of fluorescence in all directions, collecting fluorescence in only one direction is bound to decrease the fraction of emitted light, which also influences further increase of detection sensitivity. In other words, when part of the emission light is lost or fails to reach the photomultiplier tube (PMT), that lowers fluorescence collection efficiency with a concomitant decrease in the sensitivity of detection. To this end, Beckman Coulter, Inc. (Fullerton, USA) [30], made a commercialized design by inserting a spherical mirror close to and partly around the detection window to reflect the emission light into the optical fiber. The mirror may therefore reflect quite a few of the scattered excitation lights into the detector at the same time. Also, this kind of spherical mirror reflects light in an immense area randomly. Moreover, finding a proper position to place the mirror, or fabricating and fixing the mirror to the device, is rather tedious. Yang et al. used an optical fiber, with the flat end modified to a spherical end, to collect fluorescence [31], similar to the collection of fluorescence signals by using the microscope objective. Such an arrangement may involve some loss of light owing to the limited acceptance angle and small core diameter of fiber, and modifying the spherical end remains tricky. In this study, we addressed the above-mentioned problems by replacing the single optical fiber with a four-branch optical fiber while collecting the emission light. After collection of the emission light from four directions by the branches part, several times of the emission light can be collected com C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 2509–2515
pared with that using the single optical fiber without peak broadening. A green or a violet laser pointer, was used as the excitation source. The optical fibers were easily aligned to the detection window of the separation capillary through the guide channels of a home-made PDMS detection cell. The analytical performance of this CE-LIF system was demonstrated by separating and detecting orchil and fluorescent whitening agent in food safety analysis.
2 Materials and methods 2.1 Chemicals and CE apparatus Rhodamine 6G, safranine T, rose bengal, phloxine B, fluorescent brightener VBL, and fluorescent brightener FBA 351 were purchased from Aladdin Reagent Company (Shanghai, China). PDMS and Sylgard 184 elastomer curing agent were obtained from Dow Corning (San Diego, CA, USA). Other analytical-grade solvents were purchased from Guangzhou Reagents Company (Guangzhou, China), and food samples such as chili, red dates, and wheat flour were purchased from a local supermarket. Experiments were performed on a homemade CE system which can provided constant-potential direct currents of 0 to 20 kV and 0 to −20 kV for separation. Fused-silica capillaries (75 m id) were purchased from Ruifeng Chromatographic Co., Ltd. (Hebei, China). PMT was purchased from Hamamatsu Photonics Trade (H10722-01, China) Co., Ltd.
2.2 Detection system The optical arrangement for LIF detection is depicted schematically in Fig. 1. A green or a violet laser diode (applied voltage, +3 V; 100 mW; wavelength, 532 nm or 405 nm) was removed from the laser pointers, and was used as the excitation light source, controlled by a toggle switch (1MS1T61B11M2QES Q11, Dailywell Electronics Co., Ltd, Taiwan). The toggle switch is characterized by the presence of some type of handle or lever that opens and closes an electric circuit. Its circuits control the two laser pointers, is as also depicted in Fig. 1. To avoid overheating of the laser pointer, the laser pointer was coated with heat-conducting silicone grease and an aluminum cooled sink (homemade). A cooling fan is also used to quicken the heat dissipation. A Y-style optical fiber with an external diameter of 1000 m (1/2-pigtail-S10030cm-SMA, Beijing Scitlion Technology Co., Ltd, Beijing) coated with a metal sleeve was used to transmit the excitation light to the detection cell. Coupling of laser diodes to optical fiber was depicted in Fig. 1. Cylindrical polymer materials were used to combine the laser diode and the optical fiber effectively. The four screws were used to fine tune both the laser diode and the optical fiber to ensure they were at the same level. Two branches of the Y-style fiber (100 m id) were connected via SMA-905 with the green and violet www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2015, 36, 2509–2515
2511
Figure 1. Schematic illustration of the detection system. (a) green laser pointer; (b) violet laser pointer; (c) Y-style optical fiber; (d) PDMS cell; (e) four-branch optical fiber; (f) SMA connector; (g) screws; (h) metal cover of the PMT; (i) optical filter; (j) PMT; (k) data acquisition board; (l) personal computer; (m) reservoirs; (n) high voltage;(o) toggle switch; (p) coupling of laser diodes to optical fiber; (q) the screws to fine tune the laser diode and the optical fiber.
laser diodes, respectively. The common part of the fiber, as a pigtail, was connected to the detection cell. An optical fiber with four branches (1/6-pigtail-S10030 cm-SMA, Beijing Scitlion Technology Co., Ltd., Beijing) was used to collect the fluorescence in four various directions. Being all pigtails (100 m id; 140 m od), the four branches of the optical fiber were connected with the detection cell. The common part of the fiber was connected via SMA-905 with PMT. The illumination system was also equipped with a longpass filter (LP-580 and LP-430, Shenyang, China) to eliminate the interference from the excitation light. The signal from PMT was acquired by a Data Acquisition Board (NI USB6343, X Series Multifunction DAQ, USA).
2.3 Construction of PDMS detection cell A simplified PDMS detection cell with guide channels was designed and employed to align the optical fibers easily with the detection window of the separation capillary. The PDMS detection cell was fabricated by sealing two identical PDMS layers (the top and bottom layers) with the following procedure as we previously reported [32]. First, 3D stereogram layout of the cell was designed by AutoCAD, based on which a copper mold was fabricated by a commercially available fine carving. The detection cell was fabricated from PDMS produced by replica molding. PDMS elastomer and Sylgard 184 curing agent were mixed (10:1, w/w), and degassed thoroughly by a vacuum pump for 20 min. Subsequently, the mixture was poured into the copper mold and cured in an 80°C oven for 30 min, yielding the PDMS layer. After the optical fibers and capillary were bound to the channels of the two PDMS layers, they were put together between two Perspex sheets and then fixed by four screws instead of PDMS bonding to finish the PDMS cell. As shown in Fig. 2, the PDMS layer is a regular hexahedron with the side length of 17.3 mm and the thinness of C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
5 mm. The capillary channel was 30 mm in length, 375 m in width, and 187 m in depth. The excitation light optical fiber channel was 15 mm in length, 1 mm in width, and 500 m in depth. The five emission light optical fiber channels were all 15 mm in length, 140 m in width, and 70 m in depth. All of the channels intersected the center of the regular hexahedron and each of the channels also intersected the side in the middle. Figure 2 also shows the assembling procedure of the fourbranch optical fiber, Y-style optical fiber, and capillary to the detection cell. Obviously, there were two different kinds of fluorescence collection angles (defined as the angle between the excitation light channel and the emission light channel) in our detector, one (two branches of the emission light optical fiber) being 45° and the other (the other two branches of the emission light optical fiber) being 135°. The distance between the capillary and the branches of the optical fiber, with the fluorescence collection angle of 45°, was 200 m. When the angle was 135°, the distance was 0 m.
2.4 Sample preparation Stock solutions of rhodamine 6G, safranine T, rose bengal, phloxine B, fluorescent brightener VBL and fluorescent brightener FBA 351 (10 g mL−1 ) were prepared for the separation study by dissolving 1.0 mg of the compounds with methanol in 100 mL volumetric flasks. The series of all working standard solutions were then obtained by diluting the stock solutions with methanol to give the desired concentrations. All the food samples were broken into pieces by a pulverizer and extraction method was established according to a related method [33]. To determine the chili sample, 2.0 g prehomogenized chili powders were placed into a 50 mL centrifuge tube, vortexshaken for 3 min with 25 mL of methanol, and then stirred in an ultrasonic bath for 10 min. The supernatant was collected www.electrophoresis-journal.com
2512
H. Zhai et al.
Electrophoresis 2015, 36, 2509–2515
Figure 2. Schematic illustration of the fabrication for detection cell. (a) excitation optical fiber channel; (b)(d)(e)(f)(g) emission optical fiber channel; (c) capillary channel; (a’) excitation optical fiber; (b’)(d’)(f’)(g’) four branches of the emission optical fiber; (c’) capillary.
and made up to 2 mL with methanol. The extract solution was filtered through a syringe filter in a tube for CE use. So the food sample solvent was methanol. Wheat flour sample was prepared with the procedure the same as that of chili.
2.5 CE procedure Prior to injection, the capillary was pretreated by being rinsed with 1.0 M HNO3 for 20 min. Between two consecutive injections, the capillary was sequentially flushed with 0.1 M NaOH, water, and separation buffer (5 min each time). Sample solutions were injected into the capillary by gravity for 10 s, with the sample vials lifted to a height of 20 cm. An applied voltage of 19 kV or −19 kV was applied to sample separation. The total length of the capillary was 50 cm and the effective length was 40 cm. For the determination of rhodamine 6G, safranine T, rose bengal, and phloxine B, separation was performed by using 100 mM Na2 B4 O7 (pH 9.2) as the running buffer. The wavelength of exciting light was 532 nm, with a filter (LP 580) applied. For the detection of fluorescent brightener VBL and fluorescent brightener FBA 351, CTAB was selected and combined in reversed migration-micellar electrokinetic chromatography (RM-MEKC). This is characterized by anionic micelles moving faster than the EOF to optimize separation [34]. To increase the solubility of organic analytes, ACN has also been added into the running buffer [35]. Thus, 100 mM Na2 B4 O7 -0.2 mM CTAB-15% acetonitrile (pH 9.2) was used as the running buffer. The wavelength of the exciting light was 405 nm, also with a filter (LP-430) applied.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3 Results and discussion 3.1 Design of LIF system To construct a sensitive, simple, low-cost, and facile optical arrangement for fluorescence detection in CE, several special designs were made in this study. The fluorescence detection system included laser pointers, a Y-style optical fiber, a four-branch optical fiber, long-pass filters, PMT, and a PDMS detection cell, which demonstrated a special optical configuration compared to conventional fluorescence detection. Sensitivity or S/N is generally augmented by enhancing signals while reducing noise. Signal intensity is influenced by excitation light intensity, detection distance, and the collection efficiency of the emission light, etc. A laser pointer, which is intense and highly stable, is an appropriate excitation source. A green or a violet laser pointer was used in the system to extend the detection range of the sample. A Y-style optical fiber was used to transmit two branches of excitation light, with one branch for the green and the other one for the violet. We also used a four-branch optical fiber to collect the emission light in four different directions, so several times of the fluorescence were collected and the signal intensity was increased. As to noise, consideration should be given to the detection distance, the fluorescence collection angle, and the effectiveness of the filter. After being collected by the four-branch optical fiber, the emission light enters PMT, before which an optical filter should be used to filter most reflected and scattered lights. Also, detection distance affects both sensitivity and noise. For a simple, low-cost, and compact device, a laser pointer is a good choice because it is cheap, commercially available, and small sized. A more compact optical system is allowed
www.electrophoresis-journal.com
Electrophoresis 2015, 36, 2509–2515
Figure 3. Effect of different number of branches. Labeling a’-g’ are the same as Fig. 2. Sample: 130 nM rhodamine 6G; operation conditions: injection for 10 s at height of 20 cm; sepration voltage: 19 kV; running buffer: 100 mM Na2 B4 O7 (pH 9.2); excitation/emission wavelength: 532 nm/580 nm.
by using optical fiber that can flexibly locate optical elements. To fix the capillary and optical fibers, a simplified detection PDMS cell with guide channels was designed, with which the optical fibers were easily aligned to the detection window. Lenses, which need extra fixing to devices and which occupy space for light focusing, were not used herein. In addition, as depicted in Fig. 1, a miniature optical filter (diameter, 1.5 cm) was inserted into the PMT (in the front: a metal cover fixed by screws) instead of fixing extra devices, to further miniaturize the system.
3.2 Optimization of the detector 3.2.1 Effects of branch numbers To verify the effectiveness of our four-branch optical fiber, the effects of branch numbers on the fluorescence emission intensities and noise were evaluated (Fig. 3). With increasing number of branches from one to four, the emission intensities rose continuously. Meanwhile, the noise increased, at a significantly slower speed than the signal did though. In other words, increasing the number of branches raised S/N of the fluorescence detector. Different directions of the emission optical fibers contributed differently to the increase of S/N. The branches of optical fibers with the fluorescence collection angle of 135° contributed more to S/N than that with 45° did, which can mainly be ascribed to various detection distances between the capillary and the emission optical fiber instead of various detection angles. Since the fifth branch of optical fiber with the fluorescence collection angle of 180° can further increase S/N, a five-branch optical fiber was also tested. However, in this case, the collected signal exceeded the acceptable range of PMT. Probably, when the excitation and emission optical fibers were in the rectilinear direction, massive excitation light traveled through the capillary, and entered the emission C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
2513
Figure 4. Effect of detection distance. Labeling a’-g’ are the same as Fig. 2. Sample: 130 nM rhodamine 6G; operation conditions are the same as Fig. 3.
optical fiber directly. Though the long-pass filter located before PMT could block more than 90% of the excitation light, only a small proportion of the strong excitation light received by PMT sufficed to induce a high noise, or even exceeded the full-scale value of PMT. Hence, four-branch optical fiber, which gave the highest S/N in our fluorescence detector, was finally selected. 3.2.2 Effect of detection distance The effects of distances between the capillary and the optical fibers on the fluorescence emission intensities and noise were also studied (Fig. 4). Excitation light may produce a certain degree of light reflection and scattering on the wall of the capillary, which was one of the important components to the noise. As shown in Fig. 4, the detection cell can be divided into two areas (zone A and zone B) by using the capillary as the separatrix, with far more noise in zone A than that in zone B. Therefore, for the branches with 45° fluorescence collection angles (branches in zone A), increasing the noise contributed more to S/N than increasing the signal did when the detection distance was less than 200 m. On the contrary, when the detection distance was more than 200 m, decrease of the signal was more severe than decrease of the noise. For the branches with 135° fluorescence collection angles (branches in zone B), reducing the signal contributed more to S/N than reducing the noise did when the detection distance was more than 0 m. Obviously, detection distance evidently affected the sensitivity. Thus, the detection distances of 200 and 0 m were selected for the branches in zone A and zone B, respectively.
3.3 Performance of the detection system The performance of the LIF detector, with the green (532 nm) laser pointer as the excitation source and the longpass filter (580 nm) reducing the background signal, was evaluated by analyzing rhodamine 6G (exmax = 525 nm, www.electrophoresis-journal.com
2514
H. Zhai et al.
Electrophoresis 2015, 36, 2509–2515
Table 1. Performance of the present laser-induced fluorescence detector
Analytes
LOD (nM)
Linearity range (nM)
Correlation
Theoretical plates
rhodamine 6G safranine T rose bengal phloxine B fluorescent brightener VBL fluorescent brightener FBA 351
7.7 109.0 63.3 25.2 57.2 68.9
15–850 150–1000 100–1000 50–850 150–1000 150–1000
0.999 0.999 0.998 0.999 0.998 0.998
40 974 40 903 349 893 334 441 14 154 57 390
Figure 5. Electrophorogram of dyes. (A) Peak identities: (a) rhodamine 6G; (b) safranine T; (c) rose Bengal; (d) phloxine B; operation conditions were the same as Fig. 3; (B) Peak identities: (a) fluorescent brightener VBL; (b) fluorescent brightener FBA 351; operation conditions: injection for 10 s at height of 20 cm; sepration voltage: -19 kV; running buffer: 100 mM Na2 B4 O7 -0.2 mM CTAB-15% acetonitrile.
emmax = 550 nm), safranine T (exmax = 522 nm, emmax = 588 nm), rose bengal (exmax = 550 nm, emmax = 570 nm), and phloxine B (exmax = 532 nm, emmax = 560 nm). In the meantime, fluorescent brightener VBL (exmax = 344 nm, emmax = 429 nm) and fluorescent brightener FBA 351 (exmax = 352 nm, emmax = 427 nm) were also used to evaluate the performance of the LIF detector, with the violet (405 nm) laser pointer as the excitation source and the long-pass filter (430 nm) cutting off the background signal. The S/N was determined from the peak height divided by the standard deviation of the baseline noise (n = 7). As shown in Table 1, LODs are in the range of 7.7–109.0 nM (S/N = 3) under the optimized conditions, giving similar or higher sensitivity than those reported previously [29, 36–38]. Typical electropherogram of these fluorophores is shown in Fig. 5, exhibiting good repeatability with relative standard deviations (RSDs) of 2.3–3.4% and 2.8–2.9% for peak height and migration time, respectively. Fluorescent dyes, which belong to illegal food additives, have become a global concern recently [33,39–41]. They might be harmful if eaten by human, irritate the skin, eyes and respiratory tract, and even lead to teratogenesis and cancer. Hence, it is meaningful to establish an effective and sensitive method for the determination of fluorescent dyes in food. Qi et al. [42] developed an HPLC coupled with fluorescence detection method to rapidly, simply, and sensitively determine rhodamine B in chili-containing products. However, CE is C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
superior to these methods, and it requires significantly less injection volume (nanoscale) compared to HPLC does (microscale) and saves organic reagents [43]. To show the potential applicability of this proposed instrumental setup, rhodamine 6G, safranine T, rose bengal, phloxine B, fluorescent brightener VBL, and fluorescent brightener FBA 351 were determined and separated. Rhodamine 6G, safranine T, rose bengal, and phloxine B have been illegally added in food like chili for color improvement. Also, fluorescent brightener VBL and fluorescent brightener FBA 351 have been used to whiten food such as wheat flour. The recoveries of the spiked samples ranged between 97.0 and 103%, with RSDs ranging from 1.6 to 3.6%. LODs (S/N = 3) of the six fluorescent dyes were in the range of 3.68–64.4 g kg−1 , which met the limits of 500–1000 g kg–1 for all carcinogenic dyes set by the European Community [44].
4 Concluding remarks A novel fluorescence detection system for CE was designed and prepared. This system possessed a special optical configuration compared to those of conventional fluorescence detection systems, which adopted two laser pointers as excitation source to meet the demand of analytes with different excitation wavelengths, and a Y-style optical fiber was used to transport the excitation light. In order to increase the www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2015, 36, 2509–2515
2515
detection sensitivity, the system utilized a four-branch optical fiber to collect the emission light in different directions. Moreover, a PDMS cell was designed and built to combine capillary and fibers readily. Then, an optical filter was inserted into PMT to filter the reflected and scattered lights. To further optimize the sensitivity, four major influencing factors, i.e. number of optical fiber branches, detection distance, as well as performance of laser diodes and excitation filter, were studied and optimized. This design rendered the system easily home-built, low-cost and compact, accompanied by sensitive responses to the selected fluorescent dyes. Furthermore, this device was also applied to analyze these dyes in food samples. In summary, this system is promising for food safety analysis.
[16] Hurth, C., Lenigk, R., Zenhausern, F., Appl. Phys. B-Lasers O. 2008, 93, 693–699.
The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (NSFC, Grant No. 21005021) and Guangdong Provincial Science and Technology Project (No.2013B040402010 and No.2010B0203106010).
[24] Yang, X. P., Qian, F., Xie, L. X., Yang, X. C., Cheng, X. M., Choi, M. M. F., Electrophoresis 2014, 35, 762–769.
[17] Park, J., Kim, S., Microchem. J. 2011, 9, 103–107. [18] Moody, E. D., Viskari, P. J., Colyer, C. L., J. Chromatogr. B 1999, 729, 55–64. [19] Melanson, J. E., Boulet, C. A., Lucy, C. A., Anal. Chem. 2001, 7, 1809–1813. [20] Nagaraj, S., Karnes, H. T., Instrum. Sci. Technol. 2000, 28, 119–129. [21] Yang, F., Li, X. C., Zhang, W., Pan, J. B., Chen, Z. G., Talanta 2011, 84, 1099–1106. [22] Yang, X. P., Yuan, H. Y., Wang, C. L., Zhao, S. L., Xiao, D., Choi M. M. F., Electrophoresis 2007, 28, 3105–3114. [23] Mazurczyk, R., Vieillard, J., Bouchard, A., Hannes, B., Krawczyk, S., Sensor. Actuat. B-Chem. 2006, 11, 11–19.
[25] Geng, X. H., Wu, D. P., Guan, Y. F., Talanta 2012, 88, 463–467. [26] Butler, P. A. G., Mills, B., Hauser, P. C., Analyst 1997, 122, 949–953. [27] Caslavska, K., Gassmann, E., Thormann, W., J. Chromatogr. A 1995, 709, 147–156.
The authors have declared no conflict of interest.
[28] Tan, F., Yang, B. C., Guan, Y. F., Anal. Sci. 2005, 21, 583–585.
5 References ˜ [1] Nevado, J. J. B., Penalvo, G. C., Dorado R. M. R., Robledo, V. R., J. Pharmaceut. Biomed. 2014, 92, 211–219.
[29] Yang, X. B., Yan, W. P., Bai, H. M., Lv, H. F., Liu, Z. H., Optik 2012, 123, 2126–2130. [30] Kork, B. J. D., Jong, G. J. D., Somsen, G. W., Anal. Chim. Acta 2013, 76, 13–33.
[2] Dabrowska, M., Starek, M., Food Chem. 2014, 142, ˛ 220–232.
[31] Yang, B. C., Guan, Y. F., Talanta 2003, 59, 509–514.
[3] Zhu, Z. F., Lu, J. J., Liu, S. R., Anal. Chim. Acta 2012, 709, 21–31.
[32] Zhai, H. Y., Li, J. M., Chen, Z. G., Su, Z. H., Liu, Z. P., Yu, X., Microchem. J. 2014, 114, 223–228.
¨ [4] Gotz, S., Karst, U., Sensor. Actuat. B-Chem. 2007, 123, 622–627.
[33] Wang, W. J., Zhu, X. S., Yan, C. G., Food Chem. 2013, 141, 2207–2212.
[5] Mark, J. J. P., Scholz, R., Matysik, F. M., J. Chromatogr. A 2012, 1267, 45–64.
´ Y., Rodr´ıguez, R., Manes, ˜ [34] Pico, J., Trac-Trend. Anal. Chem. 2003, 22, 133–151.
[6] Wang, T., Zhang, Q., Zhang, Y., Kang, J., J. Chromatogr. A 2014, 133, 188–193.
[35] Yang, Y. Z., Boysen, R. I., Hearn, M. T. W., J. Chromatogr. A 2004, 1043, 91–97.
[7] Li, Y., Lacey, M. E., Sweedler, J. V., Webb, A. G., J. Magn. Reson. 2003, 162, 133–140.
[36] Kort, B. J. D., Jong, G. J. D., Somsen, G. W., Electrophoresis 2010, 31, 2861–2868.
[8] Yan, Q., Chen, R. S., Cheng, J. K., Anal. Chim. Acta 2006, 555, 246–249.
ˇ a, ´ M., Akther, M., Maaskant, P., Brabazon, D., [37] Vaculovicov Macka, M., Anal. Chim. Acta 2015, 871, 85–92.
[9] Ban, E., Song, E. J., J. Chromatogr. B 2013, 929, 180–186.
[38] Banerjee, A., Pais, A., Papautsky, I., Klotzkin, D., Ieee Sens. J. 2008, 8, 621–627.
[10] Bui, D. A., Hauser, P. C., Anal. Chim. Acta 2015, 853, 46–58.
[39] Xiao, N., Deng, J., Huang, K. H, Ju, S., Hu, C. H., Liang, J., Spectrochim. Acta A 2014, 128, 312–318.
[11] Lin, Y. F., Wang, Y. C., Chang, S. Y., J. Chromatogr. A 2008, 1188, 331–333. ´ ´ ´ P., Ryvolova, ´ M., Havel, J., Preisler, [12] Vrabel, P., Taborsk y, J., J. Lumin. 2006, 118, 283–292. [13] Macka, M., Piasecki, T., Dsgupta, P. K., Ann. Rev. Anal. Chem. 2014, 7, 183–207.
[40] Yoshioka, N., Ichihashi, K., Talanta 2008, 7, 1408–1413. [41] Jiang, D., Chen, L., Fu, W. S., Qiu, H. Q., J. Sep. Sci. 2015, 38, 605–611. [42] Qi, P., Lin, Z. H., Li, J. X., Wang, C. L., Meng, W. W., Hong, H., Zhang, X. W., Food Chem. 2014, 16, 98–103.
86,
´ [43] Perez-Ruiz, T., Mart´ınez-Lozano, C., Sanz, A., Bravo, E., Chromatographia 1998, 48, 263–267.
[15] Bui, D. A., Hauser, P. C., Anal. Chim. Acta 2015, 853, 46–58.
[44] EC, Directive of the European Parliament and of the council 94/36/EC of June 30, 1994 on colours for use in foodstuffs. Off. J. 1994, L237, 13.
[14] Prikryl, J., Foret, 11951–11956.
F.,
Anal.
Chem.
2014,
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
