- Email: [email protected]
Chemiluminescence microfluidic system sensor on a chip for determination of glucose in human serum with immobilized reagents
Talanta 59 (2003) 571 /576 www.elsevier.com/locate/talanta
Chemiluminescence microfluidic system sensor on a chip for determination of glucose in human serum with immobilized reagents Yi Lv, Zhujun Zhang *, Funan Chen Department of Chemistry, Institute of Analytical Science, Southwest Normal University, Beibei, Chongqing 400715, People’s Republic of China Received 26 April 2002; received in revised form 17 October 2002; accepted 17 October 2002
Abstract A chemiluminescence (CL) biosensor on a chip coupled to microfluidic system is described in this paper. The CL biosensor measured 25 /45 /5 mm in dimension, was readily produced in analytical laboratory. Glucose oxidase (GOD) was immobilized onto controlled-pore glass (CPG) via glutaraldehyde activation and packed into a reservoir. The analytical reagents, including luminol and ferricyanide, were electrostatically co-immobilized on an anion-exchange resin. The most characteristic of the biosensor was to introduce the air as the carrier flow in stead of the common solution carrier for the first. The glucose was sensed by the CL reaction between hydrogen peroxide produced from the enzymatic reaction and CL reagents, which were released from the anion-exchange resin. The proposed method has been successfully applied to the determination of glucose in human serum. The linear range of the glucose concentration was 1.1 /110 mM and the detection limit was 0.1 mM (3s ). # 2002 Elsevier Science B.V. All rights reserved. Keywords: Microfluidic system; Sensor; Chip; Chemiluminescence; Glucose
1. Introduction The determination of glucose concentration in human blood is an important analysis for the diagnosis and effective treatment of diabetes. Therefore, a large number of efforts have been focused on developing effective diagnostic tools for the benefit of diabetic patients in recent years. As a result, new biosensors have been introduced * Corresponding author. Tel./fax: /86-23-6825-3863. E-mail address: [email protected] (Z. Zhang).
for monitoring glucose [1 /5]. In general, most of these biosensors were based on the electrochemical method, e.g. amperometric biosensor [6 /8]. On the other hand, owing to its inherent high sensitivity and wide linear working range with simple instrumentation, chemiluminescence (CL) analysis has been successfully applied to glucose measurements [9 /11]. Microfluidic analysis systems fabricated on silica and glass microchip, in recent years, have become of major interest, especially to analytical chemists due to its desirable characteristics, such
0039-9140/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 2 ) 0 0 5 6 8 - 4
572
Y. Lv et al. / Talanta 59 (2003) 571 /576
as reduction in reagent consumption, required space, and analysis time [12,13]. CL analysis promises high sensitivity with simple instruments and does not need any light source, so CL have been one of attractive detection methods in m-TAS in recent years [14 /19]. Otherwise, many flow chemical sensors and flow biological sensors [20 / 24], which employed CL reagents in immobilized or solid-state format, have been developed. Those sensors had some advantages of simplicity of apparatus, little reagent consumption, and high sensitivities. In this paper, we developed a novel CL biosensor on a chip coupled to microfluidic system for determination of glucose in serum using air as the carrier flow. The carrier solution was generally necessary for not only flow injection analysis (FIA) but also flow sensor, and the air bubble was forbidden in the carrier flow. However, on this biosensor chip, the carrier flow was the air instead of the ordinary solution. In microfluidic system, solution was usually used as carrier flow, thus small air bubbles taking shape in the rough inner wall of the microchannels were difficult to be eliminated and often effected the stability and the repeatability. Fortunately, the carrier air could avoid that disadvantage [19]. Furthermore, carrier air could reduce the background resulting from the chemical interference to obtain great signal noise ratio (SNR) and high sensitivity. In general, micromechanical pumping devices (such as syringe pumps and silicon or plastic diaphragm pumps) and electro-osmotic flow (EOF) were two important and basic methods of the fluid movement in microfluidic system [25]. However, the fluid movement on this biosensor chip depended upon the negative pressure coming from the peristaltic bump.
dards were prepared daily from the stock solution by phosphate buffer (1.0 mM, pH 7.4). A 0.25 M luminol solution was prepared by dissolving 4.43 g of luminol (Shaanxi Normal University, China, /95%) in 100 ml of 0.5 M NaOH solution. The concentration of potassium ferricyanide was 0.20 g ml 1. Glucose oxidase (GOD) was obtained from Sigma. The reagents used for immobilization of the enzyme were (3-aminopropyl) trimethoxysilone (Wuhan University Chemical Factory), glutaraldehyde (25% w/w; Shanghai Wusi Chemical Regent) and the controlled-pore glass (CPG; Sigma, mesh size 80/120, mean pore diameter 7.4 nm, pore volume 0.47 cm3 g 1, surface area 152.7 m2 g 1). Strongly anion exchange resin D201, purchased from Nankai University, was used for the immobilization of luminol and ferricyanide.
2.2. Instrumentation A schematic diagram of the CL sensor-microfluidic analysis system is shown in Fig. 1. A homemade CL sensor on chip was coupled to peristaltic pump (Shanghai Instrumental Factory, Shanghai, China) by PTFE tubing (0.6 mm i.d.). The signal produced by CL microfluidic system sensor on chip was detected and recorded with a computerized Ultra-weak Luminescence Analyzer (type BPCL, manufactured at the Institute of Biophysics, Academia Sinica, China). The voltage in the CR-105 photomultiplier tube (Hamamatsu Japan) was kept at /800 V. Data acquisition and
2. Experimental 2.1. Reagents and standard solution All chemicals were of analytical reagent grade; doubly distilled water was used for the preparation of solutions. A stock solution of glucose (0.55 M) was stored in refrigerator (4 8C). Working stan-
Fig. 1. Schematic diagram of CL sensor microfluidic analysis system on chip for determination of glucose (A, enzyme reservoir; B, CL immobilized reagents reservoir; R, reactor cell; C, channel).
Y. Lv et al. / Talanta 59 (2003) 571 /576
treatment were performed with running under WINDOWS 95.
BPCL
software
2.3. Fabrication of the microfluidic system sensor The homemade CL sensor on chip shown in Fig. 2 was produced using two transparent glass plates. On the top plate (25 /45 /5 mm), two reservoirs of 5.0 mm length and 2.5 mm i.d. in dimension were made for enzyme reservoir (A) and CL reagent reservoir (B). On the surface of the top plate, a microchannel of 20 mm length, 0.15 mm i.d. was etched to link the enzyme reservoir (A) and CL reagent reservoir (B). Another microchannel (15 mm length, 0.15 mm i.d.) through the reactor cell (R) was produced to link the channel C and the midpoint of the microchannel (AB). The reactor cell (R) was 0.1 mm high and 1.5 mm i.d. in dimension. Finally, the base plate (25 /45 /1 mm) was bonded to the top plate. After the fabrication of the sensor chip was finished, its bottom except for the reaction reservoir was shaded by the black dope. Thus, only the light from the CL reaction in the reaction reservoir could be detected. 2.4. Preparation of immobilized reagent D201 anion exchange resin (0.5 g) was stirred with 25 ml of 0.25 M luminol or 0.1 M potassium
573
ferricyanide for 12 h, then the resin was filtered, washed with doubly-distilled water and kept dry for storing. The most convenient method to determine the amounts of luminol and ferricyanide is to measure the change in their concentrations in immobilization solutions [26]. This was done by UV /Vis absorbance. The concentrations were monitored at 360 nm for luminol and 420 nm for ferricyanide. The amounts of luminol and ferricyanide immobilized were 1.98 and 1.15 mmol g 1 resin, respectively. About 0.02 g of the abovementioned two resins, respectively, were mixed, and packed into the reservoir A, whose bottom already had been plugged with glass wool. 2.5. Preparation of enzyme reactor The immobilization of GOD on CPG was similar to that proposed by I Asfaha et al. [27]. About 0.1 g CPG which had been washed with 5% nitric acid and bidistilled water in order and dried in vacuum at 95 8C for 2 h, was added to 10 ml of 10% solution of (3-aminopropyl) trimethoxysilone in dry toluene, the formed mixture had been stirred for 4 h at 75 8C. The thus-obtained amine-modified CPG was filtered, washed with 96% ethanol and didistilled water, and then dried in vacuum at 100 8C. The above-mentioned amine-modified CPG was added to 10 ml 2.5% glutaraldehyde solution, the mixture was kept for a reaction at room temperature for 3 h, then filtered and washed with 96% ethanol and water. The resulting CPG was added to 10 ml of 0.1 M phosphate buffer solution (pH 7.0) containing 20 mg of GOD, and then kept for a reaction at room temperature for 2 h under constant agitation. The immobilized enzyme CPG was filtered, washed with phosphate buffer solution, and stored in refrigerator at 4 8C. About 0.03 g of the immobilized enzyme CPG was packed in reservoir (B) which had been plugged with glass wool in the bottom. 2.6. Procedures
Fig. 2. Schematic diagram of CL microfluidic system sensor on a chip for determination of glucose (A, enzyme reservoir; B, CL immobilized reagents reservoir; R, reactor cell; C, channel).
After 15 ml sample was injected into the enzyme reservoir (A), quiescing for 4 min to allow the enzyme-catalyzed oxidation of glucose, 20 ml
574
Y. Lv et al. / Talanta 59 (2003) 571 /576
eluent (Na3PO4 /NaOH) was injected into the immobilized CL reagent reservoir (B). Then the peristaltic bump was started, the produced H2O2 and the released CL reagents from anion-exchange resin reservoir were encountered and commingled well in reactor cell (R). And a CL signal was produced. The concentrations of glucose were quantified by using the CL intensity.
3. Results and discussion 3.1. Conditions for enzymatic reaction in immobilized reactor The pH of solution is an important parameter for most enzymatic reactions. Fig. 3 shows the effect of the solution pH on the enzymatic activity. The CL intensity increased rapidly from pH 4.5 to 5.0, and the maximum CL intensity was obtained at pH 5.0. Above pH 5.0, the CL intensity is reduced because of the effect of the solution pH on the enzymatic activity. However, pH 7.4 was used in all subsequent studies considering the common pH of the serum and rapid analysis rate of serum sample. The time of that the substrate was allowed to stay in enzyme reservoir (A) controlled the converting rate of glucose to the hydrogen peroxide to be detected. Preliminary experiments showed that
Fig. 3. Effect of the solution pH on the enzymatic activity (glucose: 5.5 mM; flow velocity: 100 cm min 1).
although a higher CL signal could be obtained at a longer quiescing time, the analysis rate might be limited in this condition. To obtain both high sensitivity and a rapid analysis rate, 4 min was finally selected as the quiescing time.
3.2. Influence of eluent Different amounts of luminol and ferricyanide could be released by anions with different eluting abilities being injected through the resin reservoir, which had an effect on the CL intensity. Furthermore the pH condition of eluent had an important effect on the CL reaction too. Thus several different eluents and their mixtures with NaOH, respectively, were examined. The results are shown in Table 1. The mixture eluent of Na3PO4 and NaOH with the highest relative CL intensity was chosen for subsequent work.
3.3. Effect of flow velocity of air stream The advantages of the air carrier flow were to avoid spreading of sample in carrier solution, eliminate the formation of small air bubble and chemical background. Fig. 4 shows a great repeatability of CL intensity and a stable baseline near upon zero were obtained in the assay using air as the carrier flow. The influence of the flow velocity of the air stream on the CL response was investigated in the 40/160 cm min 1 range by changing the speed of the peristaltic pump. Fig. 5 showed that CL intensity increased acutely with the increase of carrier flow velocity up to 100 cm min 1, above which it increased slowly. Because the CL reaction between H2O2 and the luminol, ferricyanide was a fast process, the lower flow velocity leaded to the short and wide peak of the CL intensity, the higher flow velocity resulted in the higher peak of the CL intensity and higher sensitivity, however, a too high velocity brought about bad repeatability. Finally, a flow velocity of 100 cm min 1 was selected as being optimum for a high CL intensity and a stable response.
Y. Lv et al. / Talanta 59 (2003) 571 /576
575
Table 1 Characteristics of eluents for determination of glucose Eluent
NaOH
Na3PO4
Na3PO4 /NaOHa
Na2SO4
Na2SO4 /NaOHa
Na2CO3
Na2CO3 /NaOHa
Relative CL intensity
65
55
100
28
75
47
85
The concentration of each eluent was 3 mM. The concentration ratio of the components in the mixing eluent was 1:1.
a
standard deviation for 5.5 mM glucose was 3.9% (n /7). The measurement throughput of the system is 12 samples per hour.
3.5. The lifetime of the biosensor
Fig. 4. Typical recording of the system’s response to glucose standard (glucose: 5.5 mM, pH 7.4; flow velocity: 100 cm min 1).
The assays showed that the lifetime of luminolreagent beads was generally more than 1 year. No significant changes were observed in response characteristics of the system after 1 week. In this system, the chip with immobilized luminol and ferricyanide could be reused for over 200 times. The chip was stored at 4 8C when it was not being used.
3.6. Application
Fig. 5. Effect of the carrier flow velocity on CL intensity (glucose: 5.5 mM, pH 7.4).
3.4. Performance of the biosensor for glucose measurements Under the selected conditions, the response to the glucose concentration was linear over range 1.1 /110 mM with a regression equation of I / 5.09C (mM)/12.1 (r2 /0.9991, n /7) and a detection limit of 0.1 mM (3s). The relative
The CL biosensor on chip was applied to the determination of glucose in human serum. After 15.0 ml of the serum was transferred into the enzymatic reactor for 4 min, the eluent (Na3PO4 / NaOH) was injected into the immobilized reagent reactor B, and then used for glucose analysis. In order to evaluate the present method, the official method [28] also be employed at the same time. The results are given in Table 2. As can be seen, the results with the present method agreed with those obtained by the official method. Table 2 Determination of glucose in human serum sample Sample
Proposed method (mM)a Official method (mM)a
Number 1 5.5 (9/3.6%) Number 2 4.2 (9/2.9%) Number 3 4.4 (9/3.9%) a
5.3 (9/5.2%) 4.1 (9/5.0%) 4.5 (9/4.5%)
Average of three replicates (9/R.S.D.).
576
Y. Lv et al. / Talanta 59 (2003) 571 /576
Acknowledgements This work was supported by the National Natural Science Foundation of China (No. 20175039).
References [1] D.S. Bindra, Y. Zhang, G.S. Wilson, R. Sternberg, D.R. Thevenot, D. Moatti, G. Reach, Anal. Chem. 63 (1991) 1692. [2] F. Moussy, D.J. Harrison, D.W. O’Brien, R.V. Rajotte, Anal. Chem. 65 (1993) 2072. [3] V. Poitout, D. Moatti-Sirat, G. Reach, Y. Zhang, G.S. Wilson, F. Lemonnier, J.C. Klein, Diabetologia 36 (1993) 658. [4] Y. Zhang, G.S. Wilson, Anal. Chim. Acta 281 (1993) 513. [5] J. Wang, L. Chen, M. Jiang, F. Lu, Anal. Chem. 71 (1999) 5009. [6] P.G. Osborne, O. Niwa, K. Yamamoto, Anal. Chem. 70 (1998) 1701. [7] Z. Rosenzweig, R. Kopelman, Anal. Chem. 68 (1996) 1408. [8] J. Wang, M.P. Chatrathi, B. Tian, Anal. Chem. 73 (2001) 1296. [9] K. Matsumoto, K. Waki, Anal. Chim. Acta 380 (1999) 1. [10] B. Li, Z. Zhang, Y. Jin, Anal. Chim. Acta 432 (2001) 95.
[11] Q. Fang, X.T. Shi, Y.Q. Sun, Z.L. Fang, Anal. Chem. 69 (1997) 3570. [12] D.J. Harrison, K. Fluri, K. Seiler, Z. Fan, C.S. Effenhauser, Science 261 (1993) 895. [13] Y. Tanaka, M.N. Slyadnev, K. Sato, Anal. Sci. 17 (2001) 809. [14] G.M. Greenway, L.J. Nelstrop, S.N. Port, Anal. Chim. Acta 405 (2000) 43. [15] Y. Xu, F.G. Bessoth, J.C.T. Eijkel, A. Manz, Analyst 125 (2000) 677. [16] K. Tsukagoshi, M. Hashimoto, R. Nakajima, A. Arai, Anal. Sci. 16 (2000) 1111. [17] M. Hashimoto, K. Tsukagoshi, R. Nakajima, K. Kondo, A. Arai, J. Chromatogr. A 867 (2000) 271. [18] P.A. Greenwood, C. Merrin, T. McCreedy, G.M. Greenway, Talanta 56 (2002) 539. [19] Y. Lv, Z. Zhang, F. Chen, Analyst 127 (2002) 1176. [20] X.Z. Zhou, M.A. Arnold, Anal. Chim. Acta 304 (1995) 147. [21] A. Daridon, M. Sequeira, G. Pennarun-Thomasb, H. Dirace, Sensors Actuators B 76 (2001) 235. [22] B. Li, Z. Zhang, Sens. Actuat. B 69 (2000) 70. [23] W. Qin, Z. Zhang, Y. Peng, Anal. Chim. Acta 407 (2000) 81. [24] B. Li, Z. Zhang, Y. Jin, Anal. Chim. Acta 432 (2001) 95. [25] T. Mccreedy, Trends Anal. Chem. 19 (2000) 396. [26] Z. Zhang, W. Qin, Talanta 43 (1996) 119. [27] I. Asfaha, H.A. Mottola, Anal. Chem. 52 (1980) 2332. [28] Clinical Test Handbook, People’s Heath Press, Beijing, 1990, p. 175.
Chemiluminescence microfluidic system sensor on a chip for determination of glucose in human serum with immobilized reagents Yi Lv, Zhujun Zhang *, Funan Chen Department of Chemistry, Institute of Analytical Science, Southwest Normal University, Beibei, Chongqing 400715, People’s Republic of China Received 26 April 2002; received in revised form 17 October 2002; accepted 17 October 2002
Abstract A chemiluminescence (CL) biosensor on a chip coupled to microfluidic system is described in this paper. The CL biosensor measured 25 /45 /5 mm in dimension, was readily produced in analytical laboratory. Glucose oxidase (GOD) was immobilized onto controlled-pore glass (CPG) via glutaraldehyde activation and packed into a reservoir. The analytical reagents, including luminol and ferricyanide, were electrostatically co-immobilized on an anion-exchange resin. The most characteristic of the biosensor was to introduce the air as the carrier flow in stead of the common solution carrier for the first. The glucose was sensed by the CL reaction between hydrogen peroxide produced from the enzymatic reaction and CL reagents, which were released from the anion-exchange resin. The proposed method has been successfully applied to the determination of glucose in human serum. The linear range of the glucose concentration was 1.1 /110 mM and the detection limit was 0.1 mM (3s ). # 2002 Elsevier Science B.V. All rights reserved. Keywords: Microfluidic system; Sensor; Chip; Chemiluminescence; Glucose
1. Introduction The determination of glucose concentration in human blood is an important analysis for the diagnosis and effective treatment of diabetes. Therefore, a large number of efforts have been focused on developing effective diagnostic tools for the benefit of diabetic patients in recent years. As a result, new biosensors have been introduced * Corresponding author. Tel./fax: /86-23-6825-3863. E-mail address: [email protected] (Z. Zhang).
for monitoring glucose [1 /5]. In general, most of these biosensors were based on the electrochemical method, e.g. amperometric biosensor [6 /8]. On the other hand, owing to its inherent high sensitivity and wide linear working range with simple instrumentation, chemiluminescence (CL) analysis has been successfully applied to glucose measurements [9 /11]. Microfluidic analysis systems fabricated on silica and glass microchip, in recent years, have become of major interest, especially to analytical chemists due to its desirable characteristics, such
0039-9140/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 2 ) 0 0 5 6 8 - 4
572
Y. Lv et al. / Talanta 59 (2003) 571 /576
as reduction in reagent consumption, required space, and analysis time [12,13]. CL analysis promises high sensitivity with simple instruments and does not need any light source, so CL have been one of attractive detection methods in m-TAS in recent years [14 /19]. Otherwise, many flow chemical sensors and flow biological sensors [20 / 24], which employed CL reagents in immobilized or solid-state format, have been developed. Those sensors had some advantages of simplicity of apparatus, little reagent consumption, and high sensitivities. In this paper, we developed a novel CL biosensor on a chip coupled to microfluidic system for determination of glucose in serum using air as the carrier flow. The carrier solution was generally necessary for not only flow injection analysis (FIA) but also flow sensor, and the air bubble was forbidden in the carrier flow. However, on this biosensor chip, the carrier flow was the air instead of the ordinary solution. In microfluidic system, solution was usually used as carrier flow, thus small air bubbles taking shape in the rough inner wall of the microchannels were difficult to be eliminated and often effected the stability and the repeatability. Fortunately, the carrier air could avoid that disadvantage [19]. Furthermore, carrier air could reduce the background resulting from the chemical interference to obtain great signal noise ratio (SNR) and high sensitivity. In general, micromechanical pumping devices (such as syringe pumps and silicon or plastic diaphragm pumps) and electro-osmotic flow (EOF) were two important and basic methods of the fluid movement in microfluidic system [25]. However, the fluid movement on this biosensor chip depended upon the negative pressure coming from the peristaltic bump.
dards were prepared daily from the stock solution by phosphate buffer (1.0 mM, pH 7.4). A 0.25 M luminol solution was prepared by dissolving 4.43 g of luminol (Shaanxi Normal University, China, /95%) in 100 ml of 0.5 M NaOH solution. The concentration of potassium ferricyanide was 0.20 g ml 1. Glucose oxidase (GOD) was obtained from Sigma. The reagents used for immobilization of the enzyme were (3-aminopropyl) trimethoxysilone (Wuhan University Chemical Factory), glutaraldehyde (25% w/w; Shanghai Wusi Chemical Regent) and the controlled-pore glass (CPG; Sigma, mesh size 80/120, mean pore diameter 7.4 nm, pore volume 0.47 cm3 g 1, surface area 152.7 m2 g 1). Strongly anion exchange resin D201, purchased from Nankai University, was used for the immobilization of luminol and ferricyanide.
2.2. Instrumentation A schematic diagram of the CL sensor-microfluidic analysis system is shown in Fig. 1. A homemade CL sensor on chip was coupled to peristaltic pump (Shanghai Instrumental Factory, Shanghai, China) by PTFE tubing (0.6 mm i.d.). The signal produced by CL microfluidic system sensor on chip was detected and recorded with a computerized Ultra-weak Luminescence Analyzer (type BPCL, manufactured at the Institute of Biophysics, Academia Sinica, China). The voltage in the CR-105 photomultiplier tube (Hamamatsu Japan) was kept at /800 V. Data acquisition and
2. Experimental 2.1. Reagents and standard solution All chemicals were of analytical reagent grade; doubly distilled water was used for the preparation of solutions. A stock solution of glucose (0.55 M) was stored in refrigerator (4 8C). Working stan-
Fig. 1. Schematic diagram of CL sensor microfluidic analysis system on chip for determination of glucose (A, enzyme reservoir; B, CL immobilized reagents reservoir; R, reactor cell; C, channel).
Y. Lv et al. / Talanta 59 (2003) 571 /576
treatment were performed with running under WINDOWS 95.
BPCL
software
2.3. Fabrication of the microfluidic system sensor The homemade CL sensor on chip shown in Fig. 2 was produced using two transparent glass plates. On the top plate (25 /45 /5 mm), two reservoirs of 5.0 mm length and 2.5 mm i.d. in dimension were made for enzyme reservoir (A) and CL reagent reservoir (B). On the surface of the top plate, a microchannel of 20 mm length, 0.15 mm i.d. was etched to link the enzyme reservoir (A) and CL reagent reservoir (B). Another microchannel (15 mm length, 0.15 mm i.d.) through the reactor cell (R) was produced to link the channel C and the midpoint of the microchannel (AB). The reactor cell (R) was 0.1 mm high and 1.5 mm i.d. in dimension. Finally, the base plate (25 /45 /1 mm) was bonded to the top plate. After the fabrication of the sensor chip was finished, its bottom except for the reaction reservoir was shaded by the black dope. Thus, only the light from the CL reaction in the reaction reservoir could be detected. 2.4. Preparation of immobilized reagent D201 anion exchange resin (0.5 g) was stirred with 25 ml of 0.25 M luminol or 0.1 M potassium
573
ferricyanide for 12 h, then the resin was filtered, washed with doubly-distilled water and kept dry for storing. The most convenient method to determine the amounts of luminol and ferricyanide is to measure the change in their concentrations in immobilization solutions [26]. This was done by UV /Vis absorbance. The concentrations were monitored at 360 nm for luminol and 420 nm for ferricyanide. The amounts of luminol and ferricyanide immobilized were 1.98 and 1.15 mmol g 1 resin, respectively. About 0.02 g of the abovementioned two resins, respectively, were mixed, and packed into the reservoir A, whose bottom already had been plugged with glass wool. 2.5. Preparation of enzyme reactor The immobilization of GOD on CPG was similar to that proposed by I Asfaha et al. [27]. About 0.1 g CPG which had been washed with 5% nitric acid and bidistilled water in order and dried in vacuum at 95 8C for 2 h, was added to 10 ml of 10% solution of (3-aminopropyl) trimethoxysilone in dry toluene, the formed mixture had been stirred for 4 h at 75 8C. The thus-obtained amine-modified CPG was filtered, washed with 96% ethanol and didistilled water, and then dried in vacuum at 100 8C. The above-mentioned amine-modified CPG was added to 10 ml 2.5% glutaraldehyde solution, the mixture was kept for a reaction at room temperature for 3 h, then filtered and washed with 96% ethanol and water. The resulting CPG was added to 10 ml of 0.1 M phosphate buffer solution (pH 7.0) containing 20 mg of GOD, and then kept for a reaction at room temperature for 2 h under constant agitation. The immobilized enzyme CPG was filtered, washed with phosphate buffer solution, and stored in refrigerator at 4 8C. About 0.03 g of the immobilized enzyme CPG was packed in reservoir (B) which had been plugged with glass wool in the bottom. 2.6. Procedures
Fig. 2. Schematic diagram of CL microfluidic system sensor on a chip for determination of glucose (A, enzyme reservoir; B, CL immobilized reagents reservoir; R, reactor cell; C, channel).
After 15 ml sample was injected into the enzyme reservoir (A), quiescing for 4 min to allow the enzyme-catalyzed oxidation of glucose, 20 ml
574
Y. Lv et al. / Talanta 59 (2003) 571 /576
eluent (Na3PO4 /NaOH) was injected into the immobilized CL reagent reservoir (B). Then the peristaltic bump was started, the produced H2O2 and the released CL reagents from anion-exchange resin reservoir were encountered and commingled well in reactor cell (R). And a CL signal was produced. The concentrations of glucose were quantified by using the CL intensity.
3. Results and discussion 3.1. Conditions for enzymatic reaction in immobilized reactor The pH of solution is an important parameter for most enzymatic reactions. Fig. 3 shows the effect of the solution pH on the enzymatic activity. The CL intensity increased rapidly from pH 4.5 to 5.0, and the maximum CL intensity was obtained at pH 5.0. Above pH 5.0, the CL intensity is reduced because of the effect of the solution pH on the enzymatic activity. However, pH 7.4 was used in all subsequent studies considering the common pH of the serum and rapid analysis rate of serum sample. The time of that the substrate was allowed to stay in enzyme reservoir (A) controlled the converting rate of glucose to the hydrogen peroxide to be detected. Preliminary experiments showed that
Fig. 3. Effect of the solution pH on the enzymatic activity (glucose: 5.5 mM; flow velocity: 100 cm min 1).
although a higher CL signal could be obtained at a longer quiescing time, the analysis rate might be limited in this condition. To obtain both high sensitivity and a rapid analysis rate, 4 min was finally selected as the quiescing time.
3.2. Influence of eluent Different amounts of luminol and ferricyanide could be released by anions with different eluting abilities being injected through the resin reservoir, which had an effect on the CL intensity. Furthermore the pH condition of eluent had an important effect on the CL reaction too. Thus several different eluents and their mixtures with NaOH, respectively, were examined. The results are shown in Table 1. The mixture eluent of Na3PO4 and NaOH with the highest relative CL intensity was chosen for subsequent work.
3.3. Effect of flow velocity of air stream The advantages of the air carrier flow were to avoid spreading of sample in carrier solution, eliminate the formation of small air bubble and chemical background. Fig. 4 shows a great repeatability of CL intensity and a stable baseline near upon zero were obtained in the assay using air as the carrier flow. The influence of the flow velocity of the air stream on the CL response was investigated in the 40/160 cm min 1 range by changing the speed of the peristaltic pump. Fig. 5 showed that CL intensity increased acutely with the increase of carrier flow velocity up to 100 cm min 1, above which it increased slowly. Because the CL reaction between H2O2 and the luminol, ferricyanide was a fast process, the lower flow velocity leaded to the short and wide peak of the CL intensity, the higher flow velocity resulted in the higher peak of the CL intensity and higher sensitivity, however, a too high velocity brought about bad repeatability. Finally, a flow velocity of 100 cm min 1 was selected as being optimum for a high CL intensity and a stable response.
Y. Lv et al. / Talanta 59 (2003) 571 /576
575
Table 1 Characteristics of eluents for determination of glucose Eluent
NaOH
Na3PO4
Na3PO4 /NaOHa
Na2SO4
Na2SO4 /NaOHa
Na2CO3
Na2CO3 /NaOHa
Relative CL intensity
65
55
100
28
75
47
85
The concentration of each eluent was 3 mM. The concentration ratio of the components in the mixing eluent was 1:1.
a
standard deviation for 5.5 mM glucose was 3.9% (n /7). The measurement throughput of the system is 12 samples per hour.
3.5. The lifetime of the biosensor
Fig. 4. Typical recording of the system’s response to glucose standard (glucose: 5.5 mM, pH 7.4; flow velocity: 100 cm min 1).
The assays showed that the lifetime of luminolreagent beads was generally more than 1 year. No significant changes were observed in response characteristics of the system after 1 week. In this system, the chip with immobilized luminol and ferricyanide could be reused for over 200 times. The chip was stored at 4 8C when it was not being used.
3.6. Application
Fig. 5. Effect of the carrier flow velocity on CL intensity (glucose: 5.5 mM, pH 7.4).
3.4. Performance of the biosensor for glucose measurements Under the selected conditions, the response to the glucose concentration was linear over range 1.1 /110 mM with a regression equation of I / 5.09C (mM)/12.1 (r2 /0.9991, n /7) and a detection limit of 0.1 mM (3s). The relative
The CL biosensor on chip was applied to the determination of glucose in human serum. After 15.0 ml of the serum was transferred into the enzymatic reactor for 4 min, the eluent (Na3PO4 / NaOH) was injected into the immobilized reagent reactor B, and then used for glucose analysis. In order to evaluate the present method, the official method [28] also be employed at the same time. The results are given in Table 2. As can be seen, the results with the present method agreed with those obtained by the official method. Table 2 Determination of glucose in human serum sample Sample
Proposed method (mM)a Official method (mM)a
Number 1 5.5 (9/3.6%) Number 2 4.2 (9/2.9%) Number 3 4.4 (9/3.9%) a
5.3 (9/5.2%) 4.1 (9/5.0%) 4.5 (9/4.5%)
Average of three replicates (9/R.S.D.).
576
Y. Lv et al. / Talanta 59 (2003) 571 /576
Acknowledgements This work was supported by the National Natural Science Foundation of China (No. 20175039).
References [1] D.S. Bindra, Y. Zhang, G.S. Wilson, R. Sternberg, D.R. Thevenot, D. Moatti, G. Reach, Anal. Chem. 63 (1991) 1692. [2] F. Moussy, D.J. Harrison, D.W. O’Brien, R.V. Rajotte, Anal. Chem. 65 (1993) 2072. [3] V. Poitout, D. Moatti-Sirat, G. Reach, Y. Zhang, G.S. Wilson, F. Lemonnier, J.C. Klein, Diabetologia 36 (1993) 658. [4] Y. Zhang, G.S. Wilson, Anal. Chim. Acta 281 (1993) 513. [5] J. Wang, L. Chen, M. Jiang, F. Lu, Anal. Chem. 71 (1999) 5009. [6] P.G. Osborne, O. Niwa, K. Yamamoto, Anal. Chem. 70 (1998) 1701. [7] Z. Rosenzweig, R. Kopelman, Anal. Chem. 68 (1996) 1408. [8] J. Wang, M.P. Chatrathi, B. Tian, Anal. Chem. 73 (2001) 1296. [9] K. Matsumoto, K. Waki, Anal. Chim. Acta 380 (1999) 1. [10] B. Li, Z. Zhang, Y. Jin, Anal. Chim. Acta 432 (2001) 95.
[11] Q. Fang, X.T. Shi, Y.Q. Sun, Z.L. Fang, Anal. Chem. 69 (1997) 3570. [12] D.J. Harrison, K. Fluri, K. Seiler, Z. Fan, C.S. Effenhauser, Science 261 (1993) 895. [13] Y. Tanaka, M.N. Slyadnev, K. Sato, Anal. Sci. 17 (2001) 809. [14] G.M. Greenway, L.J. Nelstrop, S.N. Port, Anal. Chim. Acta 405 (2000) 43. [15] Y. Xu, F.G. Bessoth, J.C.T. Eijkel, A. Manz, Analyst 125 (2000) 677. [16] K. Tsukagoshi, M. Hashimoto, R. Nakajima, A. Arai, Anal. Sci. 16 (2000) 1111. [17] M. Hashimoto, K. Tsukagoshi, R. Nakajima, K. Kondo, A. Arai, J. Chromatogr. A 867 (2000) 271. [18] P.A. Greenwood, C. Merrin, T. McCreedy, G.M. Greenway, Talanta 56 (2002) 539. [19] Y. Lv, Z. Zhang, F. Chen, Analyst 127 (2002) 1176. [20] X.Z. Zhou, M.A. Arnold, Anal. Chim. Acta 304 (1995) 147. [21] A. Daridon, M. Sequeira, G. Pennarun-Thomasb, H. Dirace, Sensors Actuators B 76 (2001) 235. [22] B. Li, Z. Zhang, Sens. Actuat. B 69 (2000) 70. [23] W. Qin, Z. Zhang, Y. Peng, Anal. Chim. Acta 407 (2000) 81. [24] B. Li, Z. Zhang, Y. Jin, Anal. Chim. Acta 432 (2001) 95. [25] T. Mccreedy, Trends Anal. Chem. 19 (2000) 396. [26] Z. Zhang, W. Qin, Talanta 43 (1996) 119. [27] I. Asfaha, H.A. Mottola, Anal. Chem. 52 (1980) 2332. [28] Clinical Test Handbook, People’s Heath Press, Beijing, 1990, p. 175.