- Email: [email protected]
Microfluidic device capable of sensing ultrafast chemiluminescence
Talanta 78 (2009) 998–1003
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Microfluidic device capable of sensing ultrafast chemiluminescence Young-Teck Kim a , Seok Oh Ko b,1 , Ji Hoon Lee c,∗ a
Department of Packaging Science, Clemson University, Clemson, SC 29634, USA Department of Civil Engineering, Kyung Hee University, 1, Seochun-Dong, Giheung-Gu, Yongin-Si 409-701, Republic of Korea c Luminescent MD, LLC, 579 Odendhal Avenue, Gaithersburg, MD 20877, USA b
a r t i c l e
i n f o
Article history: Received 28 October 2008 Received in revised form 7 January 2009 Accepted 8 January 2009 Available online 20 January 2009 Keywords: 1,1 -Oxalyldiimidazole (ODI) derivative chemiluminescence Peroxyoxalate Liquid core waveguide (LCW) Microfluidic device Lab on a chip
a b s t r a c t Based on the principle of liquid core waveguide, a novel microfluidic device with micro-scale detection window capable of sensing flashlight emitted from rapid 1,1 -oxalyldi-4-methylimidazole (OD4MI) chemiluminescence (CL) reaction was fabricated. Light emitted from OD4MI CL reaction occurring in the micro-dimensional pentagonal detection window (length of each line segment: 900.0 m, depth: 50.0 m) of the microfluidic device with two inlets and one outlet was so bright that it was possible to take an image every 1/30 s at the optimal focusing distance (60 cm) using a commercial digital camera. Peaks obtained using a flow injection analysis (FIA) system with the micro-scale detection window and OD4MI CL detection show excellent resolution and reproducibility without any band-broadening observed in analytical devices having additional reaction channel(s) to measure light generated from slow CL reaction. Maximum height (Hmax ) and area (A) of peak, reproducibility and sensitivity observed in the FIA system with the microfluidic device and OD4MI CL detection depends on (1) the mole ratio between bis(2,4,6-trichlorophenyl) oxalate and 4-methyl imidazole yielding OD4MI, (2) the flow rate to mix OD4MI, H2 O2 and 1-AP in the detection window of the microfluidic device, and (3) H2 O2 concentration. We obtained linear calibration curves with wide dynamic ranges using Hmax and A. The detection limit of 1-AP determined with Hmax and A was as low as 0.05 fmole/injection (signal/background = 3.0). © 2009 Elsevier B.V. All rights reserved.
1. Introduction Since the first introduction of capillary electrophoresis on a chip with fluorescence detection [1], an enormous number of microfluidic devices with various optical detections (e.g., UV–vis absorbance [2–4], fluorescence [5–7], chemiluminescence [8–12]) have been developed because of the advantages of these miniaturized devices, including rapid separation of complex sample mixtures, portability, reagent/solvent economy, low cost, and broad applications in diverse fields such as biochemistry, environmental toxicology, genetics, and medical diagnostics. The primary objective of fabricating microfluidic devices is to develop totally minimized analytical systems capable of diagnosing and monitoring various diseases or quantifying environmental toxic molecules including biological and chemical warfare agents. Thus, appropriate optical detection for microfluidic devices should have good sensitivity and selectivity as well as be cost effective, small and simple. Application of UV/vis absorbance detection for microfluidic devices with very short optical pathlengths (>30 m)
∗ Corresponding author. Tel.: +1 301 393 9091; fax: +1 301 393 9092. E-mail address: [email protected] (J.H. Lee). 1 Author Seok Oh Ko has equivalently contributed on this paper as a Corresponding Author. 0039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.01.004
is generally limited because of the low sensitivity [13]. Laser induced fluorescence (LIF) detection for microfluidic devices is widely applied because of the high sensitivity. However, LIF may not be suitable light source as a detection of the totally minimized analytical systems because it is expensive, complicate and relatively large [13]. Light-emitting diode (LED) fluorescence detection for microfluidic devices has been developed to solve the problems of LIF detection [7,13]. However, sensitivity of LED detection is not as good as that of LIF detection because of a higher level of background signal generated from the wide half-bandwidth of light emitted from LED [13]. Chemiluminescence (CL) emitted by a chemical reaction is more sensitive and selective than UV/vis and fluorescence generated by a light source because of the low background [13]. Thus, various kinds of microfluidic devices with CL detection have been fabricated [8–12]. However, microfluidic devices with CL detection reported so far are more complicate and larger than those with UV/vis and fluorescence detections because the former needs long reactions channels and wide detection windows to measure optimum emission intensity generated by slow CL reactions (e.g., luminol [12], peroxyoxalate [11]). For example, microfluidic devices with extra flow elements to measure CL do not have better resolution due to the wide band-broadening [13]. Also, it is difficult to measure slow and relatively dim CL using micro-scale detection windows. The smallest dimension of detection window for microfluidic device with CL detection was 2.0 mm × 3.0 mm [13].
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
Recently, we reported that the maximum intensity, Imax , and time required to reach the maximum emission, max , for analytical sample monitored in 1,1 -oxalyldiimidazole (ODI) derivative CL reaction are much higher and faster than their values measured in peroxyoxalate (PO) CL reaction [14,15]. For example, values for Imax and max observed with emission of 1-AP in ODI derivative CL reaction are 61.1 times larger and 15.8 times faster than their respective values obtained from emission of 1-aminopyrene (1-AP) in PO-CL reaction [15]. Ultrafast max (>0.3 s) measured in ODI derivative CL reaction was apparently independent of physical properties (e.g., oxidation potential, dielectric constant, viscosity) of fluorescent molecules and solvents even though max obtained in PO-CL was clearly dependent on the factors [15–17]. Additionally, using the rapid ODI derivative CL reaction, it was possible to quantify analytical molecules dissolved in water without considering critical interferences (e.g., decomposition of CL reagents, side reaction of analytes and CL reagent) observed in relatively slow PO-CL [16,17]. Using the advantages of ODI derivative CL reaction in analytical chemistry, we fabricated a new and simple microfluidic device not having extra flow elements. 1-AP selected as an analyte in this research can emit strong light in ODI derivative CL reaction [15–17]. Also, 1-AP formed under various biological and environmental conditions was quantified and monitored [18–21]. For example, 1-nitropyrene is one of the most mutagenic and carcinogenic nitrated polycyclic hydrocarbons contained in diegel exhaust particular matter [18,19]. 1-AP, as a biomarker, formed in the in vivo metabolism of 1-nitropyrene in living cells was quantified [18,19]. 1-AP formed from the reduction reaction of 1-nitropyrene was quantified to indirectly monitor 1nitropyrene having poor fluorescence and CL quantum efficiency [20,21]. In this paper, we report that the novel microfluidic device with ODI derivative CL detection can quantify trace level of 1-AP with better resolution and reproducibility.
2. Experimental 2.1. Chemicals 1-Aminopyrene, bis(2,4,6-trichlorophenyl)oxalate (TCPO), 4methylimidazole (4MImH), and H2 O2 (50%) were purchased from Aldrich. Spectroscopic grade solvents (acetonitrile, ethyl acetate, and acetone) were purchased from Burdick & Jackson.
999
2.2. Fabrication of lab on a chip As shown in Fig. 1, we designed a new microfluidic device capable of sensing flash light emitted from ultrafst ODI derivative CL reaction. The microfluidic device has two inlets, one outlet and a pentagonal chamber. The micro-dimensional pentagonal chamber was designed to apply as a high-efficiency mixing and reaction channel as well as a detection window. The microfluidic device we designed was fabricated as follows. The lab on a chip shown in Fig. 1 was fabricated at the Institute of Advanced Machinery & Design Korea Bio-IT Foundry Center in Seoul, Republic of Korea. Six microchips like that (26.0 mm × 19.0 mm) shown in Fig. 1(b) were fabricated using two round quartz plates (diameter: 15.2 cm, thickness: 0.5 mm) to observe and quantify OD4ML CL in organic solvent mixture. The diameter and depth of each reservoir on the bottom quartz plate were 2.0 mm and 50.0 m. Holes (diameter: 0.8 mm) for delivering CL reagents or analyte into each reservoir and taking out waste from the chip were drilled into the cover plate (see Fig. 1(c)). Depth, length, and width of in-flow channel between the reservoir and the pentagonal chamber shown in Fig. 1(d) and (e) were 50 m, 1.0 cm, and 450 m, respectively. The depth of the pentagonal chamber, shown in Fig. 1(e), was 50 m, while its line segment, shown in Fig. 1(f), was 900 m. Depth, length, and width of out-flow channel between the waste reservoir and the CL detection area shown in Fig. 1(d) were 50 m, 1.0 cm, and 900 m, respectively. The surface of the bottom plate mediated with 1% HF was bonded with the cover plate under constant high pressure (0.16 MPa) at room temperature for 24 h. Three PEEK tubings (Inner and outer diameters: 250 m and 1/32 in., length: 60 cm) purchased from VICI were connected with two inlets and an outlet of the microfluidic device using commercial epoxy glue (Permatex). The other side of the two PEEK tubings connected to the inlets of the microfluidic device was each connected to a 3.0 ml Luer-Lok syringe (BD Medical). Using a high-resolution microscope (Labophot-2, Nikon) with a color video camera (VK-C370, Hitachi) and a syringe pump (975, Harvard Apparatus) having four syringe holders, we confirmed that aqueous and organic solvents inserted through the two inlets (both sides in Fig. 1(b)) were mixed smoothly and completely on the detection area shown in Fig. 1(e) and (f) and flowed out into the outlet (center in Fig. 1(b)) under various flow rates (26.0–390 l/min).
Fig. 1. Fabrication of lab on a chip capable of sensing ultrafast ODI derivative CL based on the principle of liquid core waveguide. (a) Design to fabricate 6 microchips on two round quartz plates (diameter: 15.2 cm, thickness: 0.5 mm), (b) design to fabricate microchip (26.0 mm × 19.0 mm), (c) reservoir (diameter: 2.0 mm, depth: 50.0 m) and hole (diameter: 0.8 mm), (d) reservoir and in-flow (width: 450.0 m, depth: 50.0 m, length: 1.0 cm) or out-flow (width: 900.0 m, depth: 50.0 m, length: 1.0 cm) channel, (e) two in-flow and out-flow channels and pentagonal CL detection area (length of each line segment: 900.0 m, depth: 50.0 m), (f) pentagonal CL detection area.
1000
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
2.3. Measurement CL emission with microfluidic device Fig. 2 shows the flow injection analysis (FIA) system with the microfluidic device and 1,1 -oxalyldi-4-methylimidazole (OD4MI), one of ODI derivatives, CL detection. It consists of a syringe pump (975, Harvard Apparatus) capable of pumping 1–4 syringes at the same flow rate, an injection valve (D) having 250 nl loop (CN2-4340, Valco Instruments Co, Inc. (VICI)), a microfluidic device (F) we fabricated, detection system (G) having a photomultiplier tube (PTI, Inc.), and a personal computer (H) for data collection and analysis. In order to protect light from the exterior, F and G were placed in a chamber (E, PTI, Inc). One (A) of the two syringes contains acetonitrile and the other syringe (B) has OD4MI formed from the reaction between TCPO and 4MImH in ethyl acetate at room temperature (21–23 ◦ C). 1-AP solution containing H2 O2 was injected into D through D-1 with a microsyringe (10.0 l). Constant loading time for the sample solution was 30 s. The sample solution filled in the 250 nl loop connected between D-3 and D-6 was inserted into the detection area in F through D-5 and one of the two inlets of F. OD4MI was inserted into the detection area in F through the other inlet of F. Flow rate range of CL reagents and sample inserted into the microfluidic device was 26.0–390 l/min. CL emitted with the fast OD4MI CL reaction in the detection area was detected with G at 427 nm, which is the fluorescence emission wavelength of 1-AP in the solvent of acetonitrile and ethyl acetate mixture. CL measured with G was transferred to H for data collection and analysis. In order to quantify 1-AP under the optimum condition of the FIA system with the microfluidic device and OD4MI CL detection, we observed three parameters (maximum height (Hmax ), half-width (W1/2 ), and area (A)) of each peak under chemical and physical conditions. Experimental results under each reaction condition were analyzed with Origin 7.5 (OriginLab Corporation). 3. Results and discussion 3.1. Imaging of CL emitted from OD4MI CL reaction in the detection area of the microfluidic device Using the movie mode (resolution: 640 pixels × 480 pixels, 30 frame/s) of a digital camera (Powershot A640, Canon), instead of the PMT (G) and the computer (H) shown in Fig. 2, we obtained CL images emitted with the ultrafast OD4MI CL reaction in the pentagonal chamber of the microfluidic device. Distance between the
Fig. 3. Images of CL emitted in the detection area of microfluidic device in FIA system. Integration time for each image: 1/30 s, [1-AP] = 20 fmol/injection, [TCPO] = 5.0 mM, [4MImH] = 60.0 mM, [H2 O2 ] = 1.0 M, Flow rate for 1-AP and OD4MI CL reagents: 100 l/min.
microfluidic device and the camera to maintain optimum focus was 60.0 cm. The peak shown in Fig. 3 was obtained with FIA system shown in Fig. 2, and the experimental condition for obtaining the peak was the same as that for obtaining the images. As shown in Fig. 3, brightness of CL emitted in the pentagonal chamber was dependent on the concentration of 1-AP. In other words, (1) CL emission was not observed in the absence of 1-AP, (2) brightness of CL was enhanced with the increase of 1-AP concentration, and (3) brightness of CL began to dim with the decrease of 1-AP. The CL emission in the presence of relatively high concentration of 1-AP was observed in the pentagonal chamber as well as outside the pentagonal chamber (out-flow channel and outlet) because the half-life of bright CL emitted in OD4MI CL reaction was as long as 1.7 (±0.2) s [15,16]. The outlet diameter (2.0 mm) of the microfluidic device is wider than the width (900 m) of the flow channel. As a result, CL was observed in the outlet while CL in the flow channel next to the outlet was not apparent. Based on the results observed in the section, we studied in detail the effects of OD4MI CL reagents and flow rate in FIA system with PMT instead of a commercial digital camera. 3.2. 4MImH effect in FIA system with microfluidic device and OD4MI CL detection
Fig. 2. Diagram of the flow injection analysis (FIA) system with a microfluidic device and OD4MI CL detection. (A) Solvent (acetonitrile), (B) OD4MI, (C) syringe pump, (D) injection valve ((1) sample, (2) vent/waste, (3 and 6) sample loop, (4) carrier/solvent, (5) microfluidic device), (E) chamber, (F) microfluidic device, (G) photomultiplier tube (PMT) and (H) data collection and analysis.
Scheme 1 shows the mechanism of OD4MI CL reaction. High-energy intermediates (X) capable of transferring energy to fluorescent compounds (F) are produced when OD4MI molecules formed from the 1:2 chemical reactions between TCPO and 4MImH react with H2 O2 [14,15]. Finally, ultrafast and strong CL emitted from the interaction between X and fluorescent compound (F), based on the chemically initiated electron exchange luminescence (CIEEL) mechanism [18], is observed. Even though OD4MI molecules are formed from the 1:2 reactions between TCPO and 4MImH, as shown in Scheme 1, we reported that appropriate mole ratio in the chemical reaction between TCPO and 4MImH for the fast formation of OD4MI should be larger than 1:2 [14,22]. This is necessary in order to enhance quantum efficiency and sensitivity of OD4MI CL with the elimination of interference effects observed from other POCL reactions (e.g., TCPO-CL reaction) occurring competitively with OD4MI CL reaction [14,22]. This is because 4MImH molecules act
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
1001
Scheme 1. OD4MI CL reaction.
as a reagent to form OD4MI as well as a catalyst. However, maximum CL intensity (Imax ) observed in the presence of over a certain concentration of 4MImH decreased as excess 4MImH molecules decompose OD4MI quickly before it reacts with H2 O2 to form X [14]. Thus, it is important to determine appropriate mole ratio between TCPO and 4MImH for quantifying trace level of analytes using the FIA system we developed. We prepared five different 4MImH solutions (20.0–100.0 mM in ethyl acetate) based on the previous research results (observing steady-state OD4MI CL using a stopped-flow injection) [15–17]. Table 1 shows that sensitivity of the FIA system shown in Fig. 2 is dependent on 4MImH concentration used to produce OD4MI. The mixture of 1-AP and H2 O2 was injected into the FIA system after OD4MI was formed from the reaction between TCPO and 4MImH for 1 min in the FIA system. Based on our previous batch experimental results obtained with a stopped-flow system [15–17], we expected that relative Hmax and A in the presence of 80 or 100 mM 4MImH would be higher and wider than those in the presence of lower 4MImH concentrations, respectively. However, the concentration of 4MImH to obtain the best Hmax and A observed in the FIA system was 20 mM (TCPO:4MImH = 1:4) and then, Hmax and A decreased proportionally with the increase of 4MImH concentration. This indicates that environmental condition emitting CL in the detection area of the microfluidic device of FIA system is different from that in the stopped-flow injection system. The W1/2 of peak as shown in Table 1 was constant in the presence of various concentrations of 4MImH because it is dependent on the fixed flow rate (100.0 l/min) of FIA system. However, the
results shown in Table 1 imply that the appropriate mole ratio between TCPO and 4MImH to enhance the sensitivity of OD4MI CL depends on the flow rate of 1-AP and OD4MI CL reagents in FIA system.
Table 1 4MImH effect in FIA system with the microfluidic device and OD4MI CL detection at room temperature.
TCPO:4MImH = 1:8 26.0 2.42 (±0.088) 52.0 3.10 (±0.073) 100.0 3.38 (±0.065) 200.0 1.63 (±0.069) 390.0 0.83 (±0.015)
16.78 (±0.45) 11.15 (±0.24) 7.42 (±0.21) 4.83 (±0.23) 3.23 (±0.14)
4.67 (±0.100) 3.58 (±0.074) 2.63 (±0.039) 0.83 (±0.010) 0.29 (±0.007)
TCPO:4MImH = 1:4 26.0 8.29 (±0.141) 52.0 6.49 (±0.105) 100.0 4.43 (±0.078)
16.48 (±0.49) 11.16 (±0.24) 7.56 (±0.14)
14.19 (±0.265) 7.00 (±0.134) 3.41 (±0.064)
4MImHa
Hmax (×104 )
W1/2 (s)
A (×105 )
20.0 40.0 60.0 80.0 100.0
4.43 (±0.078) 3.38 (±0.065) 2.22 (±0.059) 1.39 (±0.062) 0.96 (±0.069)
7.56 (±0.14) 7.42 (±0.21) 7.54 (±0.26) 7.52 (±0.39) 7.66 (±0.37)
3.41 (±0.064) 2.63 (±0.039) 1.97 (±0.041) 1.22 (±0.052) 0.76 (±0.047)
Flow rate: 100.0 l/min [1-AP] = 4.34 fmol/injection in acetonitrile, [H2 O2 ] = 1.0 M in acetone, [TCPO] = 5.0 mM in ethyl acetate. Mean value and standard deviation values for Hmax and W1/2 and A obtained under each experimental condition were determined with values measured with the injection of 1-AP five consecutive times at 30 s intervals. a [mM].
3.3. Flow rate effect in FIA system with the microfluidic device and OD4MI CL detection Table 2 shows the results observed under different flow rates of 1-AP and OD4MI CL reagents in the presence of ODI molecules formed under two different mole ratios (1:4 and 1:8) of TCPO and 4MImH. The mixture of 1-AP and H2 O2 was injected into the FIA system after OD4MI was formed from the reaction between TCPO and 4MImH for 1 min in the system. Hmax measured in OD4MI CL reaction in the presence of OD4MI formed from the reaction between TCPO and 4MImH (mole ratio = 1:8) increased with the increase of flow rate up to 100 l/min and then began to decrease. One reason is that excess 4MImH acts as a quencher of OD4MI CL emitted in the detection area of microfluidic device when the flow rate is slower than 100.0 l/min. The other reason is that 200.0 and 390 l/min is so fast that mixture, capable of emitting CL, in the pentagonal chamber flow out to the out-flow channel before attaining the highest CL peak. Table 2 Flow rate effect in FIA system with the microfluidic device and OD4MI CL detection at room temperature. Flow ratea
Hmax (×104 )
W1/2 (s)
A (×105 )
[1-AP] = 4.34 fmol/injection in acetonitrile, [H2 O2 ] = 1.0 M in acetone, [TCPO] = 5.0 mM in ethyl acetate, [4MImH] = 40.0 mM in ethyl acetate. Mean value and standard deviation values for Hmax and W1/2 and A obtained under each experimental condition were determined with values measured with the injection of 1-AP five consecutive times at 30 s intervals. a (l/min).
1002
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
Table 3 H2 O2 effect in FIA system with the microfluidic device and OD4MI CL detection at room temperature. H2 O2 a
Hmax (×104 )
W1/2 (s)
A (×105 )
0.2 0.4 0.6 0.8 1.0
1.74 (±0.034) 2.12 (±0.028) 2.57 (±0.047) 3.19 (±0.052) 3.17 (±0.046)
7.67 (±0.23) 7.56 (±0.31) 7.54 (±0.28) 7.38 (±0.33) 7.43 (±0.29)
1.23 (±0.011) 1.51 (±0.076) 1.81 (±0.091) 2.47 (±0.081) 2.44 (±0.058)
Flow rate: 100.0 l/min. [1-AP] = 3.20 fmol/injection in acetonitrile, [TCPO] = 5.0 mM in ethyl acetate, [4MImH] = 40.0 mM in ethyl acetate. Mean value and standard deviation values for Hmax and W1/2 and A obtained under each experimental condition were determined with values measured with the injection of 1-AP five consecutive times at 30 s intervals. a [M].
However, A measured in OD4MI CL reaction in the presence of OD4MI formed from the reaction between TCPO and 4MImH (mole ratio = 1:8) decreased proportionally with the increase of flow rate. This result indicates that A is more dependent on the flow rate than the chemical reaction in the pentagonal chamber while the opposite is true for Hmax . The effects of flow rate in OD4MI CL reaction under 1:4 mole ratio between TCPO and 4MImH were different from those under 1:8 mole ratio between TCPO and 4MImH. Hmax , W1/2 , and A increased as flow rate decreased because OD4MI molecules inserted into the pentagonal chamber of microfluidic device at 26.0 l/min can stably produce high-energy intermediates in the presence of relatively lower 4MImH, predominantly acting as a catalyst instead of quencher [15–17]. The reason why Hmax decreases with the increase of flow rate is that a portion of OD4MI CL reagents flows out from the pentagonal chamber before producing highenergy intermediate. 3.4. H2 O2 effect in FIA system with microfluidic device and OD4MI CL detection Table 3 shows the effects of H2 O2 concentration in the FIA system with microfluidic device and OD4MI CL detection. With the increase of H2 O2 concentration up to 0.8 M in acetone, Hmax and A increased. And then, Hmax and A measured in the presence of 1.0 M H2 O2 concentration were similar to those in the presence of 0.8 M H2 O2 because, based on the CIEEL mechanism, excess H2 O2 does not act as an activator to form high-energy intermediate capable of transferring energy to 1-AP [18].
a quencher instead of a catalyst, competitively broke down OD4MI formed from the reaction between TCPO and 4MImH within the syringe. For example, relative Hmax measured after 10 min of reaction between TCPO and 4MImH of 1:20 mole ratio was 5.3 times higher than that measured after 30 min of reaction. 3.6. Sensitivity of FIA system with microfluidic device and OD4MI CL detection Based on the results obtained in this research, we determined the appropriate experimental conditions (e.g., 1:4 mole ratio between TCPO and 4MImH, 100 l/min flow rate) to quantify 1-AP using the FIA system we developed. As shown in Fig. 4, we obtained linear calibration curves with good R-square values with Hmax and A measured in the presence of 7 different 1-AP concentrations (0.11–2.3 fmol/injection). The background measured in the absence of 1-AP was 483.5 ± 36.5. The detection limit of 1-AP determined with Hmax using signal to background test (signal/background = 3:1) was as low as 0.05 fmole/injection. It was comparable to that determined with A even though the slope (0.7607 ± 0.00414) of calibration curve for A is smaller than that (0.9941 ± 0.0106) for Hmax . Using signal to background test, the detection limits determined with FIA system with OD4MI CL detection are about 30 times lower than that (1.5 fmol/injection) [21] obtained using HPLC with conventional PO-CL detection and about 500 times lower than that (25 fmol/injection) [21] obtained using HPLC with fluorescence detection. Two apparent peaks (signal/background = 2) obtained in the presence of 0.03 fmole/injection 1-AP (see Fig. 4) had good reproducibility even though the concentration of 1-AP was lower than the detection limit (0.05 fmole/injection) determined with FIA system using signal to background test. This indicates that the FIA system with a microfluidic device and OD4MI CL detection has good reproducibility. Also, the lower limit of detection (LOD = 0.018 fmol/injection) and the limit of quantification (LOQ = 0.043 fmol/injection) were determined using the mean value and standard deviation measured in the absence of 1-AP under the experimental condition and the linear calibration curve for Hmax shown in Fig. 4. Table 2 indicates that 26.0 l/min is the best flow rate to quantify 1-AP using Hmax and A. The detection limits set with Hmax and A at 26.0 l/min were about 2 times lower than those at 100.0 l/min,
3.5. Stability of OD4MI and reproducibility The results shown in Table 1 and our previous research results [14,16] indicate that the stability and reproducibility of microfluidic device are dependent on the mole ratio between TCPO and 4MImH to form OD4MI. In other words, it is important to identify the optimum ratio between TCPO and 4MImH in order to obtain good results within the statistical error range. CL peaks of 1-AP (4.3 fmol/injection) measured under the 1:4 mole ratio between TCPO and 4 MIm were constant within statistical error range (coefficient of variation (CV): 3.7%) when 1-AP was injected into the FIA system every 2 min for over 30 min at the flow rate of 26 l/min. We confirmed that Hmax , W1/2 , and A under the same reaction condition decreased with the increase of flow rate from 26 to 390 l/min as shown in Table 2. However, the reproducibility for relative CL peaks was not affected by the flow rate. With the increase of mole ratio between TCPO and 4MImH, the reproducibility for CL peak measured at 2 min intervals for 30 min dropped sharply. This is because excess 4MImH, acting as
Fig. 4. Calibration curves of 1-AP based on relative Hmax () and A (䊉) in OD4MI CL reaction. Reaction condition: [H2 O2 ] = 0.8 M in acetone, [TCPO] = 5.0 mM and [4MImH] = 20.0 mM in ethyl acetate, flow rate: 100.0 l/min. The measurement of Hmax and A at the same concentration of 1-AP were replicated 3 times. The CV for Hmax and A at each concentration of 1-AP were lower than 5.0%.
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
shown in Fig. 4. However, W1/2 measured in the former was about 2 times longer than that in the latter condition. Thus, for rapid quantification of many analytes without band broadening, 26 l/min is not as good as 100.0 l/min as the appropriate flow rate of the FIA system we developed. 4. Conclusions Based on the principle of liquid core waveguide, we fabricated a novel microfluidic device capable of detecting ultrafast OD4MI CL. Peaks obtained in FIA system with the micro-scale pentagonal chamber and OD4MI CL detection showed good resolution as well as excellent reproducibility without band-broadening observed in analytical devices with slow CL detection. As shown in Fig. 1, CL emitted in the micro-scale detection area was so bright that we were able to obtain the integrated CL image within 1/30 s. This indicates that it is possible to monitor and quantify many analytes rapidly and simultaneously using microarray with the OD4MI CL detection. The results shown in Tables 1–3 indicate that amount of 4MImH staying in the pentagonal chamber is in inverse proportion to the flow rate. Also, brightness of CL emitted in the detection area depends on the flow rate of sample and OD4MI CL reagents. In conclusion, based on the results of this research, it will be possible to fabricate advanced microfluidic devices having a smaller detection window, higher efficiency and shorter W1/2 than that shown in Fig. 1. We expect that the FIA system we developed can be applied as a prototype of totally minimized analytical system capable of quantifying environmental toxic chemicals as well as monitoring and diagnosing various diseases. Acknowledgements This work was funded by LST Korea, Inc. (LST-K8 and LST-K420). This company also provided appropriate microfluidic devices for the research.
1003
References [1] D.J. Harrison, K. Fluri, K. Seiler, Z. Fan, C.S. Effenhauser, A. Manz, Science 261 (1993) 895–897. [2] Z.Q. Xu, T. Ando, T. Nishine, A. Arai, T. Hirokawa, Electrophoresis 24 (2003) 3821–3827. [3] K. Faure, M. Loughran, J.D. Glennon, Anal. Chim. Acta 557 (2006) 130– 136. [4] A.R. Stettler, M.A. Schwarz, J. Chromatogr. A 1063 (2005) 217–225. [5] S.-H. Lee, S.I. Cho, C.-S. Lee, B.-G. Kim, Y.-K. Kim, Sens. Actuators B 110 (2005) 164–173. [6] C.F. Duffy, B. MacCraith, D. Diamond, R. O’Kennedy, A. Arriaga, Lab Chip 6 (2006) 1007–1011. [7] J.B. Edel, N.P. Beard, O. Hofmann, J.C. DeMello, D.D.C. Bradley, A.J. deMello, Lab Chip 4 (2004) 136–140. [8] B.-F. Liu, M. Ozaki, H. Hisamoto, Q. Luo, Y. Utsumi, T. Hattori, S. Terabe, Anal. Chem. 77 (2005) 573–578. [9] R.G. Su, J.-M. Lin, K. Uchiyama, M. Yamada, Talanta 64 (2004) 1024– 1029. [10] H. Shen, Q. Fang, Z.-L. Fang, Lab Chip 6 (2006) 1387–1389. [11] M. Amatatongchai, O. Hofmann, D. Nacapricha, O. Chailapakul, A.J. deMello, Anal. Bioanal. Chem. 387 (2007) 277–285. [12] Y.-X. Guan, Z.-R. Xu, J. Dai, Z.-L. Fang, Talanta 68 (2006) 1384–1389. [13] S. Gotz, U. Karst, Anal. Bioanal. Chem. 387 (2007) 183–192. [14] J.H. Lee, J.C. Rock, S.B. Park, M.A. Schlautman, E.R. Carraway, J. Chem. Soc., Perkin Trans. (2002) 802–809. [15] J.H. Lee, J.T. Je, M.A. Schlautman, E.R. Carraway, Chem. Commun. (2003) 270–271. [16] J.H. Lee, J. Je, J. Hur, M.A. Schlautman, E.R. Carraway, Analyst 128 (2003) 1257–1261. [17] J.H. Lee, J. Je, A. Tartaglia, J. Hur, M.A. Schlautman, E.R. Carraway, J. Photochem. Photobiol. A: Chem. 182 (2006) 28–32. [18] P.C. Howard, M.C. Consolo, L.K. Dooley, F.A. Beland, Chem. -Biol. Interact. 95 (1995) 309–325. [19] A. Toriba, H. Kitaoka, R.L. Dills, S. Mizukami, K. Tanabe, N. Takeuchi, M. Ueno, T. Kameda, N. Tang, K. Hayakawa, C.D. Simson, Chem. Res. Toxicol. 20 (2007) 999–1007. [20] K. Hayakawa, R. Kitamura, B.M.N. Imaizumi, M. Miyazaki, Anal. Sci. 7 (1991) 573–577. [21] K. Hayakawa, N. Terai, P.G. Dinning, K. Akutsu, Y. Iwamoto, R. Etoh, T. Murahashi, Biomed. Chromatogr. 10 (1996) 346–350. [22] M.M. Rauhut, L.J. Bollyky, B.G. Roberts, M. Loy, J. Am. Chem. Soc. 89 (1967) 6515.
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Microfluidic device capable of sensing ultrafast chemiluminescence Young-Teck Kim a , Seok Oh Ko b,1 , Ji Hoon Lee c,∗ a
Department of Packaging Science, Clemson University, Clemson, SC 29634, USA Department of Civil Engineering, Kyung Hee University, 1, Seochun-Dong, Giheung-Gu, Yongin-Si 409-701, Republic of Korea c Luminescent MD, LLC, 579 Odendhal Avenue, Gaithersburg, MD 20877, USA b
a r t i c l e
i n f o
Article history: Received 28 October 2008 Received in revised form 7 January 2009 Accepted 8 January 2009 Available online 20 January 2009 Keywords: 1,1 -Oxalyldiimidazole (ODI) derivative chemiluminescence Peroxyoxalate Liquid core waveguide (LCW) Microfluidic device Lab on a chip
a b s t r a c t Based on the principle of liquid core waveguide, a novel microfluidic device with micro-scale detection window capable of sensing flashlight emitted from rapid 1,1 -oxalyldi-4-methylimidazole (OD4MI) chemiluminescence (CL) reaction was fabricated. Light emitted from OD4MI CL reaction occurring in the micro-dimensional pentagonal detection window (length of each line segment: 900.0 m, depth: 50.0 m) of the microfluidic device with two inlets and one outlet was so bright that it was possible to take an image every 1/30 s at the optimal focusing distance (60 cm) using a commercial digital camera. Peaks obtained using a flow injection analysis (FIA) system with the micro-scale detection window and OD4MI CL detection show excellent resolution and reproducibility without any band-broadening observed in analytical devices having additional reaction channel(s) to measure light generated from slow CL reaction. Maximum height (Hmax ) and area (A) of peak, reproducibility and sensitivity observed in the FIA system with the microfluidic device and OD4MI CL detection depends on (1) the mole ratio between bis(2,4,6-trichlorophenyl) oxalate and 4-methyl imidazole yielding OD4MI, (2) the flow rate to mix OD4MI, H2 O2 and 1-AP in the detection window of the microfluidic device, and (3) H2 O2 concentration. We obtained linear calibration curves with wide dynamic ranges using Hmax and A. The detection limit of 1-AP determined with Hmax and A was as low as 0.05 fmole/injection (signal/background = 3.0). © 2009 Elsevier B.V. All rights reserved.
1. Introduction Since the first introduction of capillary electrophoresis on a chip with fluorescence detection [1], an enormous number of microfluidic devices with various optical detections (e.g., UV–vis absorbance [2–4], fluorescence [5–7], chemiluminescence [8–12]) have been developed because of the advantages of these miniaturized devices, including rapid separation of complex sample mixtures, portability, reagent/solvent economy, low cost, and broad applications in diverse fields such as biochemistry, environmental toxicology, genetics, and medical diagnostics. The primary objective of fabricating microfluidic devices is to develop totally minimized analytical systems capable of diagnosing and monitoring various diseases or quantifying environmental toxic molecules including biological and chemical warfare agents. Thus, appropriate optical detection for microfluidic devices should have good sensitivity and selectivity as well as be cost effective, small and simple. Application of UV/vis absorbance detection for microfluidic devices with very short optical pathlengths (>30 m)
∗ Corresponding author. Tel.: +1 301 393 9091; fax: +1 301 393 9092. E-mail address: [email protected] (J.H. Lee). 1 Author Seok Oh Ko has equivalently contributed on this paper as a Corresponding Author. 0039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.01.004
is generally limited because of the low sensitivity [13]. Laser induced fluorescence (LIF) detection for microfluidic devices is widely applied because of the high sensitivity. However, LIF may not be suitable light source as a detection of the totally minimized analytical systems because it is expensive, complicate and relatively large [13]. Light-emitting diode (LED) fluorescence detection for microfluidic devices has been developed to solve the problems of LIF detection [7,13]. However, sensitivity of LED detection is not as good as that of LIF detection because of a higher level of background signal generated from the wide half-bandwidth of light emitted from LED [13]. Chemiluminescence (CL) emitted by a chemical reaction is more sensitive and selective than UV/vis and fluorescence generated by a light source because of the low background [13]. Thus, various kinds of microfluidic devices with CL detection have been fabricated [8–12]. However, microfluidic devices with CL detection reported so far are more complicate and larger than those with UV/vis and fluorescence detections because the former needs long reactions channels and wide detection windows to measure optimum emission intensity generated by slow CL reactions (e.g., luminol [12], peroxyoxalate [11]). For example, microfluidic devices with extra flow elements to measure CL do not have better resolution due to the wide band-broadening [13]. Also, it is difficult to measure slow and relatively dim CL using micro-scale detection windows. The smallest dimension of detection window for microfluidic device with CL detection was 2.0 mm × 3.0 mm [13].
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
Recently, we reported that the maximum intensity, Imax , and time required to reach the maximum emission, max , for analytical sample monitored in 1,1 -oxalyldiimidazole (ODI) derivative CL reaction are much higher and faster than their values measured in peroxyoxalate (PO) CL reaction [14,15]. For example, values for Imax and max observed with emission of 1-AP in ODI derivative CL reaction are 61.1 times larger and 15.8 times faster than their respective values obtained from emission of 1-aminopyrene (1-AP) in PO-CL reaction [15]. Ultrafast max (>0.3 s) measured in ODI derivative CL reaction was apparently independent of physical properties (e.g., oxidation potential, dielectric constant, viscosity) of fluorescent molecules and solvents even though max obtained in PO-CL was clearly dependent on the factors [15–17]. Additionally, using the rapid ODI derivative CL reaction, it was possible to quantify analytical molecules dissolved in water without considering critical interferences (e.g., decomposition of CL reagents, side reaction of analytes and CL reagent) observed in relatively slow PO-CL [16,17]. Using the advantages of ODI derivative CL reaction in analytical chemistry, we fabricated a new and simple microfluidic device not having extra flow elements. 1-AP selected as an analyte in this research can emit strong light in ODI derivative CL reaction [15–17]. Also, 1-AP formed under various biological and environmental conditions was quantified and monitored [18–21]. For example, 1-nitropyrene is one of the most mutagenic and carcinogenic nitrated polycyclic hydrocarbons contained in diegel exhaust particular matter [18,19]. 1-AP, as a biomarker, formed in the in vivo metabolism of 1-nitropyrene in living cells was quantified [18,19]. 1-AP formed from the reduction reaction of 1-nitropyrene was quantified to indirectly monitor 1nitropyrene having poor fluorescence and CL quantum efficiency [20,21]. In this paper, we report that the novel microfluidic device with ODI derivative CL detection can quantify trace level of 1-AP with better resolution and reproducibility.
2. Experimental 2.1. Chemicals 1-Aminopyrene, bis(2,4,6-trichlorophenyl)oxalate (TCPO), 4methylimidazole (4MImH), and H2 O2 (50%) were purchased from Aldrich. Spectroscopic grade solvents (acetonitrile, ethyl acetate, and acetone) were purchased from Burdick & Jackson.
999
2.2. Fabrication of lab on a chip As shown in Fig. 1, we designed a new microfluidic device capable of sensing flash light emitted from ultrafst ODI derivative CL reaction. The microfluidic device has two inlets, one outlet and a pentagonal chamber. The micro-dimensional pentagonal chamber was designed to apply as a high-efficiency mixing and reaction channel as well as a detection window. The microfluidic device we designed was fabricated as follows. The lab on a chip shown in Fig. 1 was fabricated at the Institute of Advanced Machinery & Design Korea Bio-IT Foundry Center in Seoul, Republic of Korea. Six microchips like that (26.0 mm × 19.0 mm) shown in Fig. 1(b) were fabricated using two round quartz plates (diameter: 15.2 cm, thickness: 0.5 mm) to observe and quantify OD4ML CL in organic solvent mixture. The diameter and depth of each reservoir on the bottom quartz plate were 2.0 mm and 50.0 m. Holes (diameter: 0.8 mm) for delivering CL reagents or analyte into each reservoir and taking out waste from the chip were drilled into the cover plate (see Fig. 1(c)). Depth, length, and width of in-flow channel between the reservoir and the pentagonal chamber shown in Fig. 1(d) and (e) were 50 m, 1.0 cm, and 450 m, respectively. The depth of the pentagonal chamber, shown in Fig. 1(e), was 50 m, while its line segment, shown in Fig. 1(f), was 900 m. Depth, length, and width of out-flow channel between the waste reservoir and the CL detection area shown in Fig. 1(d) were 50 m, 1.0 cm, and 900 m, respectively. The surface of the bottom plate mediated with 1% HF was bonded with the cover plate under constant high pressure (0.16 MPa) at room temperature for 24 h. Three PEEK tubings (Inner and outer diameters: 250 m and 1/32 in., length: 60 cm) purchased from VICI were connected with two inlets and an outlet of the microfluidic device using commercial epoxy glue (Permatex). The other side of the two PEEK tubings connected to the inlets of the microfluidic device was each connected to a 3.0 ml Luer-Lok syringe (BD Medical). Using a high-resolution microscope (Labophot-2, Nikon) with a color video camera (VK-C370, Hitachi) and a syringe pump (975, Harvard Apparatus) having four syringe holders, we confirmed that aqueous and organic solvents inserted through the two inlets (both sides in Fig. 1(b)) were mixed smoothly and completely on the detection area shown in Fig. 1(e) and (f) and flowed out into the outlet (center in Fig. 1(b)) under various flow rates (26.0–390 l/min).
Fig. 1. Fabrication of lab on a chip capable of sensing ultrafast ODI derivative CL based on the principle of liquid core waveguide. (a) Design to fabricate 6 microchips on two round quartz plates (diameter: 15.2 cm, thickness: 0.5 mm), (b) design to fabricate microchip (26.0 mm × 19.0 mm), (c) reservoir (diameter: 2.0 mm, depth: 50.0 m) and hole (diameter: 0.8 mm), (d) reservoir and in-flow (width: 450.0 m, depth: 50.0 m, length: 1.0 cm) or out-flow (width: 900.0 m, depth: 50.0 m, length: 1.0 cm) channel, (e) two in-flow and out-flow channels and pentagonal CL detection area (length of each line segment: 900.0 m, depth: 50.0 m), (f) pentagonal CL detection area.
1000
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
2.3. Measurement CL emission with microfluidic device Fig. 2 shows the flow injection analysis (FIA) system with the microfluidic device and 1,1 -oxalyldi-4-methylimidazole (OD4MI), one of ODI derivatives, CL detection. It consists of a syringe pump (975, Harvard Apparatus) capable of pumping 1–4 syringes at the same flow rate, an injection valve (D) having 250 nl loop (CN2-4340, Valco Instruments Co, Inc. (VICI)), a microfluidic device (F) we fabricated, detection system (G) having a photomultiplier tube (PTI, Inc.), and a personal computer (H) for data collection and analysis. In order to protect light from the exterior, F and G were placed in a chamber (E, PTI, Inc). One (A) of the two syringes contains acetonitrile and the other syringe (B) has OD4MI formed from the reaction between TCPO and 4MImH in ethyl acetate at room temperature (21–23 ◦ C). 1-AP solution containing H2 O2 was injected into D through D-1 with a microsyringe (10.0 l). Constant loading time for the sample solution was 30 s. The sample solution filled in the 250 nl loop connected between D-3 and D-6 was inserted into the detection area in F through D-5 and one of the two inlets of F. OD4MI was inserted into the detection area in F through the other inlet of F. Flow rate range of CL reagents and sample inserted into the microfluidic device was 26.0–390 l/min. CL emitted with the fast OD4MI CL reaction in the detection area was detected with G at 427 nm, which is the fluorescence emission wavelength of 1-AP in the solvent of acetonitrile and ethyl acetate mixture. CL measured with G was transferred to H for data collection and analysis. In order to quantify 1-AP under the optimum condition of the FIA system with the microfluidic device and OD4MI CL detection, we observed three parameters (maximum height (Hmax ), half-width (W1/2 ), and area (A)) of each peak under chemical and physical conditions. Experimental results under each reaction condition were analyzed with Origin 7.5 (OriginLab Corporation). 3. Results and discussion 3.1. Imaging of CL emitted from OD4MI CL reaction in the detection area of the microfluidic device Using the movie mode (resolution: 640 pixels × 480 pixels, 30 frame/s) of a digital camera (Powershot A640, Canon), instead of the PMT (G) and the computer (H) shown in Fig. 2, we obtained CL images emitted with the ultrafast OD4MI CL reaction in the pentagonal chamber of the microfluidic device. Distance between the
Fig. 3. Images of CL emitted in the detection area of microfluidic device in FIA system. Integration time for each image: 1/30 s, [1-AP] = 20 fmol/injection, [TCPO] = 5.0 mM, [4MImH] = 60.0 mM, [H2 O2 ] = 1.0 M, Flow rate for 1-AP and OD4MI CL reagents: 100 l/min.
microfluidic device and the camera to maintain optimum focus was 60.0 cm. The peak shown in Fig. 3 was obtained with FIA system shown in Fig. 2, and the experimental condition for obtaining the peak was the same as that for obtaining the images. As shown in Fig. 3, brightness of CL emitted in the pentagonal chamber was dependent on the concentration of 1-AP. In other words, (1) CL emission was not observed in the absence of 1-AP, (2) brightness of CL was enhanced with the increase of 1-AP concentration, and (3) brightness of CL began to dim with the decrease of 1-AP. The CL emission in the presence of relatively high concentration of 1-AP was observed in the pentagonal chamber as well as outside the pentagonal chamber (out-flow channel and outlet) because the half-life of bright CL emitted in OD4MI CL reaction was as long as 1.7 (±0.2) s [15,16]. The outlet diameter (2.0 mm) of the microfluidic device is wider than the width (900 m) of the flow channel. As a result, CL was observed in the outlet while CL in the flow channel next to the outlet was not apparent. Based on the results observed in the section, we studied in detail the effects of OD4MI CL reagents and flow rate in FIA system with PMT instead of a commercial digital camera. 3.2. 4MImH effect in FIA system with microfluidic device and OD4MI CL detection
Fig. 2. Diagram of the flow injection analysis (FIA) system with a microfluidic device and OD4MI CL detection. (A) Solvent (acetonitrile), (B) OD4MI, (C) syringe pump, (D) injection valve ((1) sample, (2) vent/waste, (3 and 6) sample loop, (4) carrier/solvent, (5) microfluidic device), (E) chamber, (F) microfluidic device, (G) photomultiplier tube (PMT) and (H) data collection and analysis.
Scheme 1 shows the mechanism of OD4MI CL reaction. High-energy intermediates (X) capable of transferring energy to fluorescent compounds (F) are produced when OD4MI molecules formed from the 1:2 chemical reactions between TCPO and 4MImH react with H2 O2 [14,15]. Finally, ultrafast and strong CL emitted from the interaction between X and fluorescent compound (F), based on the chemically initiated electron exchange luminescence (CIEEL) mechanism [18], is observed. Even though OD4MI molecules are formed from the 1:2 reactions between TCPO and 4MImH, as shown in Scheme 1, we reported that appropriate mole ratio in the chemical reaction between TCPO and 4MImH for the fast formation of OD4MI should be larger than 1:2 [14,22]. This is necessary in order to enhance quantum efficiency and sensitivity of OD4MI CL with the elimination of interference effects observed from other POCL reactions (e.g., TCPO-CL reaction) occurring competitively with OD4MI CL reaction [14,22]. This is because 4MImH molecules act
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
1001
Scheme 1. OD4MI CL reaction.
as a reagent to form OD4MI as well as a catalyst. However, maximum CL intensity (Imax ) observed in the presence of over a certain concentration of 4MImH decreased as excess 4MImH molecules decompose OD4MI quickly before it reacts with H2 O2 to form X [14]. Thus, it is important to determine appropriate mole ratio between TCPO and 4MImH for quantifying trace level of analytes using the FIA system we developed. We prepared five different 4MImH solutions (20.0–100.0 mM in ethyl acetate) based on the previous research results (observing steady-state OD4MI CL using a stopped-flow injection) [15–17]. Table 1 shows that sensitivity of the FIA system shown in Fig. 2 is dependent on 4MImH concentration used to produce OD4MI. The mixture of 1-AP and H2 O2 was injected into the FIA system after OD4MI was formed from the reaction between TCPO and 4MImH for 1 min in the FIA system. Based on our previous batch experimental results obtained with a stopped-flow system [15–17], we expected that relative Hmax and A in the presence of 80 or 100 mM 4MImH would be higher and wider than those in the presence of lower 4MImH concentrations, respectively. However, the concentration of 4MImH to obtain the best Hmax and A observed in the FIA system was 20 mM (TCPO:4MImH = 1:4) and then, Hmax and A decreased proportionally with the increase of 4MImH concentration. This indicates that environmental condition emitting CL in the detection area of the microfluidic device of FIA system is different from that in the stopped-flow injection system. The W1/2 of peak as shown in Table 1 was constant in the presence of various concentrations of 4MImH because it is dependent on the fixed flow rate (100.0 l/min) of FIA system. However, the
results shown in Table 1 imply that the appropriate mole ratio between TCPO and 4MImH to enhance the sensitivity of OD4MI CL depends on the flow rate of 1-AP and OD4MI CL reagents in FIA system.
Table 1 4MImH effect in FIA system with the microfluidic device and OD4MI CL detection at room temperature.
TCPO:4MImH = 1:8 26.0 2.42 (±0.088) 52.0 3.10 (±0.073) 100.0 3.38 (±0.065) 200.0 1.63 (±0.069) 390.0 0.83 (±0.015)
16.78 (±0.45) 11.15 (±0.24) 7.42 (±0.21) 4.83 (±0.23) 3.23 (±0.14)
4.67 (±0.100) 3.58 (±0.074) 2.63 (±0.039) 0.83 (±0.010) 0.29 (±0.007)
TCPO:4MImH = 1:4 26.0 8.29 (±0.141) 52.0 6.49 (±0.105) 100.0 4.43 (±0.078)
16.48 (±0.49) 11.16 (±0.24) 7.56 (±0.14)
14.19 (±0.265) 7.00 (±0.134) 3.41 (±0.064)
4MImHa
Hmax (×104 )
W1/2 (s)
A (×105 )
20.0 40.0 60.0 80.0 100.0
4.43 (±0.078) 3.38 (±0.065) 2.22 (±0.059) 1.39 (±0.062) 0.96 (±0.069)
7.56 (±0.14) 7.42 (±0.21) 7.54 (±0.26) 7.52 (±0.39) 7.66 (±0.37)
3.41 (±0.064) 2.63 (±0.039) 1.97 (±0.041) 1.22 (±0.052) 0.76 (±0.047)
Flow rate: 100.0 l/min [1-AP] = 4.34 fmol/injection in acetonitrile, [H2 O2 ] = 1.0 M in acetone, [TCPO] = 5.0 mM in ethyl acetate. Mean value and standard deviation values for Hmax and W1/2 and A obtained under each experimental condition were determined with values measured with the injection of 1-AP five consecutive times at 30 s intervals. a [mM].
3.3. Flow rate effect in FIA system with the microfluidic device and OD4MI CL detection Table 2 shows the results observed under different flow rates of 1-AP and OD4MI CL reagents in the presence of ODI molecules formed under two different mole ratios (1:4 and 1:8) of TCPO and 4MImH. The mixture of 1-AP and H2 O2 was injected into the FIA system after OD4MI was formed from the reaction between TCPO and 4MImH for 1 min in the system. Hmax measured in OD4MI CL reaction in the presence of OD4MI formed from the reaction between TCPO and 4MImH (mole ratio = 1:8) increased with the increase of flow rate up to 100 l/min and then began to decrease. One reason is that excess 4MImH acts as a quencher of OD4MI CL emitted in the detection area of microfluidic device when the flow rate is slower than 100.0 l/min. The other reason is that 200.0 and 390 l/min is so fast that mixture, capable of emitting CL, in the pentagonal chamber flow out to the out-flow channel before attaining the highest CL peak. Table 2 Flow rate effect in FIA system with the microfluidic device and OD4MI CL detection at room temperature. Flow ratea
Hmax (×104 )
W1/2 (s)
A (×105 )
[1-AP] = 4.34 fmol/injection in acetonitrile, [H2 O2 ] = 1.0 M in acetone, [TCPO] = 5.0 mM in ethyl acetate, [4MImH] = 40.0 mM in ethyl acetate. Mean value and standard deviation values for Hmax and W1/2 and A obtained under each experimental condition were determined with values measured with the injection of 1-AP five consecutive times at 30 s intervals. a (l/min).
1002
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
Table 3 H2 O2 effect in FIA system with the microfluidic device and OD4MI CL detection at room temperature. H2 O2 a
Hmax (×104 )
W1/2 (s)
A (×105 )
0.2 0.4 0.6 0.8 1.0
1.74 (±0.034) 2.12 (±0.028) 2.57 (±0.047) 3.19 (±0.052) 3.17 (±0.046)
7.67 (±0.23) 7.56 (±0.31) 7.54 (±0.28) 7.38 (±0.33) 7.43 (±0.29)
1.23 (±0.011) 1.51 (±0.076) 1.81 (±0.091) 2.47 (±0.081) 2.44 (±0.058)
Flow rate: 100.0 l/min. [1-AP] = 3.20 fmol/injection in acetonitrile, [TCPO] = 5.0 mM in ethyl acetate, [4MImH] = 40.0 mM in ethyl acetate. Mean value and standard deviation values for Hmax and W1/2 and A obtained under each experimental condition were determined with values measured with the injection of 1-AP five consecutive times at 30 s intervals. a [M].
However, A measured in OD4MI CL reaction in the presence of OD4MI formed from the reaction between TCPO and 4MImH (mole ratio = 1:8) decreased proportionally with the increase of flow rate. This result indicates that A is more dependent on the flow rate than the chemical reaction in the pentagonal chamber while the opposite is true for Hmax . The effects of flow rate in OD4MI CL reaction under 1:4 mole ratio between TCPO and 4MImH were different from those under 1:8 mole ratio between TCPO and 4MImH. Hmax , W1/2 , and A increased as flow rate decreased because OD4MI molecules inserted into the pentagonal chamber of microfluidic device at 26.0 l/min can stably produce high-energy intermediates in the presence of relatively lower 4MImH, predominantly acting as a catalyst instead of quencher [15–17]. The reason why Hmax decreases with the increase of flow rate is that a portion of OD4MI CL reagents flows out from the pentagonal chamber before producing highenergy intermediate. 3.4. H2 O2 effect in FIA system with microfluidic device and OD4MI CL detection Table 3 shows the effects of H2 O2 concentration in the FIA system with microfluidic device and OD4MI CL detection. With the increase of H2 O2 concentration up to 0.8 M in acetone, Hmax and A increased. And then, Hmax and A measured in the presence of 1.0 M H2 O2 concentration were similar to those in the presence of 0.8 M H2 O2 because, based on the CIEEL mechanism, excess H2 O2 does not act as an activator to form high-energy intermediate capable of transferring energy to 1-AP [18].
a quencher instead of a catalyst, competitively broke down OD4MI formed from the reaction between TCPO and 4MImH within the syringe. For example, relative Hmax measured after 10 min of reaction between TCPO and 4MImH of 1:20 mole ratio was 5.3 times higher than that measured after 30 min of reaction. 3.6. Sensitivity of FIA system with microfluidic device and OD4MI CL detection Based on the results obtained in this research, we determined the appropriate experimental conditions (e.g., 1:4 mole ratio between TCPO and 4MImH, 100 l/min flow rate) to quantify 1-AP using the FIA system we developed. As shown in Fig. 4, we obtained linear calibration curves with good R-square values with Hmax and A measured in the presence of 7 different 1-AP concentrations (0.11–2.3 fmol/injection). The background measured in the absence of 1-AP was 483.5 ± 36.5. The detection limit of 1-AP determined with Hmax using signal to background test (signal/background = 3:1) was as low as 0.05 fmole/injection. It was comparable to that determined with A even though the slope (0.7607 ± 0.00414) of calibration curve for A is smaller than that (0.9941 ± 0.0106) for Hmax . Using signal to background test, the detection limits determined with FIA system with OD4MI CL detection are about 30 times lower than that (1.5 fmol/injection) [21] obtained using HPLC with conventional PO-CL detection and about 500 times lower than that (25 fmol/injection) [21] obtained using HPLC with fluorescence detection. Two apparent peaks (signal/background = 2) obtained in the presence of 0.03 fmole/injection 1-AP (see Fig. 4) had good reproducibility even though the concentration of 1-AP was lower than the detection limit (0.05 fmole/injection) determined with FIA system using signal to background test. This indicates that the FIA system with a microfluidic device and OD4MI CL detection has good reproducibility. Also, the lower limit of detection (LOD = 0.018 fmol/injection) and the limit of quantification (LOQ = 0.043 fmol/injection) were determined using the mean value and standard deviation measured in the absence of 1-AP under the experimental condition and the linear calibration curve for Hmax shown in Fig. 4. Table 2 indicates that 26.0 l/min is the best flow rate to quantify 1-AP using Hmax and A. The detection limits set with Hmax and A at 26.0 l/min were about 2 times lower than those at 100.0 l/min,
3.5. Stability of OD4MI and reproducibility The results shown in Table 1 and our previous research results [14,16] indicate that the stability and reproducibility of microfluidic device are dependent on the mole ratio between TCPO and 4MImH to form OD4MI. In other words, it is important to identify the optimum ratio between TCPO and 4MImH in order to obtain good results within the statistical error range. CL peaks of 1-AP (4.3 fmol/injection) measured under the 1:4 mole ratio between TCPO and 4 MIm were constant within statistical error range (coefficient of variation (CV): 3.7%) when 1-AP was injected into the FIA system every 2 min for over 30 min at the flow rate of 26 l/min. We confirmed that Hmax , W1/2 , and A under the same reaction condition decreased with the increase of flow rate from 26 to 390 l/min as shown in Table 2. However, the reproducibility for relative CL peaks was not affected by the flow rate. With the increase of mole ratio between TCPO and 4MImH, the reproducibility for CL peak measured at 2 min intervals for 30 min dropped sharply. This is because excess 4MImH, acting as
Fig. 4. Calibration curves of 1-AP based on relative Hmax () and A (䊉) in OD4MI CL reaction. Reaction condition: [H2 O2 ] = 0.8 M in acetone, [TCPO] = 5.0 mM and [4MImH] = 20.0 mM in ethyl acetate, flow rate: 100.0 l/min. The measurement of Hmax and A at the same concentration of 1-AP were replicated 3 times. The CV for Hmax and A at each concentration of 1-AP were lower than 5.0%.
Y.-T. Kim et al. / Talanta 78 (2009) 998–1003
shown in Fig. 4. However, W1/2 measured in the former was about 2 times longer than that in the latter condition. Thus, for rapid quantification of many analytes without band broadening, 26 l/min is not as good as 100.0 l/min as the appropriate flow rate of the FIA system we developed. 4. Conclusions Based on the principle of liquid core waveguide, we fabricated a novel microfluidic device capable of detecting ultrafast OD4MI CL. Peaks obtained in FIA system with the micro-scale pentagonal chamber and OD4MI CL detection showed good resolution as well as excellent reproducibility without band-broadening observed in analytical devices with slow CL detection. As shown in Fig. 1, CL emitted in the micro-scale detection area was so bright that we were able to obtain the integrated CL image within 1/30 s. This indicates that it is possible to monitor and quantify many analytes rapidly and simultaneously using microarray with the OD4MI CL detection. The results shown in Tables 1–3 indicate that amount of 4MImH staying in the pentagonal chamber is in inverse proportion to the flow rate. Also, brightness of CL emitted in the detection area depends on the flow rate of sample and OD4MI CL reagents. In conclusion, based on the results of this research, it will be possible to fabricate advanced microfluidic devices having a smaller detection window, higher efficiency and shorter W1/2 than that shown in Fig. 1. We expect that the FIA system we developed can be applied as a prototype of totally minimized analytical system capable of quantifying environmental toxic chemicals as well as monitoring and diagnosing various diseases. Acknowledgements This work was funded by LST Korea, Inc. (LST-K8 and LST-K420). This company also provided appropriate microfluidic devices for the research.
1003
References [1] D.J. Harrison, K. Fluri, K. Seiler, Z. Fan, C.S. Effenhauser, A. Manz, Science 261 (1993) 895–897. [2] Z.Q. Xu, T. Ando, T. Nishine, A. Arai, T. Hirokawa, Electrophoresis 24 (2003) 3821–3827. [3] K. Faure, M. Loughran, J.D. Glennon, Anal. Chim. Acta 557 (2006) 130– 136. [4] A.R. Stettler, M.A. Schwarz, J. Chromatogr. A 1063 (2005) 217–225. [5] S.-H. Lee, S.I. Cho, C.-S. Lee, B.-G. Kim, Y.-K. Kim, Sens. Actuators B 110 (2005) 164–173. [6] C.F. Duffy, B. MacCraith, D. Diamond, R. O’Kennedy, A. Arriaga, Lab Chip 6 (2006) 1007–1011. [7] J.B. Edel, N.P. Beard, O. Hofmann, J.C. DeMello, D.D.C. Bradley, A.J. deMello, Lab Chip 4 (2004) 136–140. [8] B.-F. Liu, M. Ozaki, H. Hisamoto, Q. Luo, Y. Utsumi, T. Hattori, S. Terabe, Anal. Chem. 77 (2005) 573–578. [9] R.G. Su, J.-M. Lin, K. Uchiyama, M. Yamada, Talanta 64 (2004) 1024– 1029. [10] H. Shen, Q. Fang, Z.-L. Fang, Lab Chip 6 (2006) 1387–1389. [11] M. Amatatongchai, O. Hofmann, D. Nacapricha, O. Chailapakul, A.J. deMello, Anal. Bioanal. Chem. 387 (2007) 277–285. [12] Y.-X. Guan, Z.-R. Xu, J. Dai, Z.-L. Fang, Talanta 68 (2006) 1384–1389. [13] S. Gotz, U. Karst, Anal. Bioanal. Chem. 387 (2007) 183–192. [14] J.H. Lee, J.C. Rock, S.B. Park, M.A. Schlautman, E.R. Carraway, J. Chem. Soc., Perkin Trans. (2002) 802–809. [15] J.H. Lee, J.T. Je, M.A. Schlautman, E.R. Carraway, Chem. Commun. (2003) 270–271. [16] J.H. Lee, J. Je, J. Hur, M.A. Schlautman, E.R. Carraway, Analyst 128 (2003) 1257–1261. [17] J.H. Lee, J. Je, A. Tartaglia, J. Hur, M.A. Schlautman, E.R. Carraway, J. Photochem. Photobiol. A: Chem. 182 (2006) 28–32. [18] P.C. Howard, M.C. Consolo, L.K. Dooley, F.A. Beland, Chem. -Biol. Interact. 95 (1995) 309–325. [19] A. Toriba, H. Kitaoka, R.L. Dills, S. Mizukami, K. Tanabe, N. Takeuchi, M. Ueno, T. Kameda, N. Tang, K. Hayakawa, C.D. Simson, Chem. Res. Toxicol. 20 (2007) 999–1007. [20] K. Hayakawa, R. Kitamura, B.M.N. Imaizumi, M. Miyazaki, Anal. Sci. 7 (1991) 573–577. [21] K. Hayakawa, N. Terai, P.G. Dinning, K. Akutsu, Y. Iwamoto, R. Etoh, T. Murahashi, Biomed. Chromatogr. 10 (1996) 346–350. [22] M.M. Rauhut, L.J. Bollyky, B.G. Roberts, M. Loy, J. Am. Chem. Soc. 89 (1967) 6515.