Integrated Light Collimating System for Extended Optical-Path-Length

Anal. Chem. 2005, 77, 5160-5166

Integrated Light Collimating System for Extended Optical-Path-Length Absorbance Detection in Microchip-Based Capillary Electrophoresis Kyung Won Ro,† Kwanseop Lim, Bong Chu Shim,‡ and Jong Hoon Hahn*

Department of Chemistry, Division of Molecular and Life Science, Pohang University of Science and Technology, San 31 Hyoja-Dong, Pohang, 790-784, South Korea

We have developed an integrated light collimating system with a microlens and a pair of slits for extended optical path length absorbance detection in a capillary electrophoresis (CE) microchip. The collimating system is made of the same material as the chip, poly(dimethylsiloxane) (PDMS), and it is integrated into the chip during the molding of the CE microchannels. In this microchip, the centers of an extended 500-µm detection cell and two optical fibers are self-aligned, and a planoconvex microlens (r ) 50 µm) for light collimation is placed in front of a light-delivering fiber. To block stray light, two rectangular apertures, realized by a specially designed threedimensional microchannel, are made on each end of the detection cell. In comparison to conventional extended detection cell having no collimator, the percentage of stray radiation readout fraction in the collimator integrated detection cell is significantly reduced from 31.6 to 3.8%. The effective optical path length is increased from 324 to 460 µm in the collimator integrated detection cell. The detection sensitivity is increased by 10 times in the newly developed absorbance detection cell as compared to an unextended, 50-µm-long detection cell. The concentration detection limit (S/N ) 3) for fluorescein in the collimator integrated detection cell is 1.2 µM at the absorbance detection limit of 0.001 AU. UV-visible absorbance detection has been the most popular optical detection mode for microseparations, because it is simple, has wide applicability, and, in general, does not require cumbersome chemical derivatization often involved in fluorescence detection. In absorbance detection, the on-column mode has been most commonly used, but it suffers from the high concentration limit of detection due to the short optical path length available, particularly in capillary electrophoresis (CE). During the past decade, several attempts have been made to enhance the absorbance detection sensitivity in CE by extending the optical path length, which include the use of rectangular capillaries,1 bubble * Corresponding author. E-mail: [email protected]. Phone: 82-54-279-2118. Fax: 82-54-279-5805. † Current address: Medical University of South Carolina, Charleston, SC 29425. ‡ Current address: LG Electronics Institute of Technology, Seoul, South Korea.

5160 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

cells,2,3 multireflection cells,4 and Z-,5-7 U-,8 and upside-down T-shaped flow cells.9,10 Microseparations in capillaries have been successfully adapted for microchip devices,11-14 where UV-visible absorbance detection is also a common detection technique, and the depth of a microchannel, which is only tens of micrometers, is usually used as the optical path length. To extend the path length for light absorption, a Z-cell,15 a multireflection cell,16 and a U-cell17 have been also fabricated into microchip CE devices. In the U-shaped absorbance detection cell, optical fibers were inserted into the microchip to guide the excitation light to the U-cell and the transmitted light to the detector, and thus, light passed through the cell longitudinally.17 However, within the limit of maintaining the linearity and sensitivity in detection, the path length of the cell could be extended to 150 µm at most because of the intrinsic divergence of light from the excitation fiber. Recently, Kutter and co-workers integrated optical waveguides monolithically in a CE microchip with a 750-µm U-shaped absorbance detection cell, but they also observed detrimental effects of stray light due to the light dispersion.18,19 (1) Tsuda, T.; Sweedler, J. V.; Zare, R. N. Anal. Chem. 1990, 62, 2149-2152. (2) Heiger, D. N.; Kaltenbach, P.; Sievert, H.-J. P. Electrophoresis 1994, 15, 1234-1247. (3) Xue, Y.: Yeung, E. S. Anal. Chem. 1994, 66, 3575-3580. (4) Wang, T.; Aiken, J. H.; Huie, C. W.; Hartwick, R. A. Anal. Chem. 1991, 63, 1372-1376. (5) Bruin, G. J. M.; Stegeman, G.; Van Austen, A. C.; Xu, X.; Kraak, J. C.; Poppe, H. J. Chromatogr. 1991, 559, 163-181. (6) Chervet, J. P.; Van Soest, R. E. J.; Ursem, M. J. Chromatogr. 1991, 543, 439-449. (7) Moring, S. E.; Reel, R. T.; Van Soest, R. E. J. Anal. Chem. 1993, 65, 34543459. (8) Mainka, A.; Ba¨chmann, K. J. Chromatogr., A 1997, 767, 241-247. (9) Lim, K.; Kim, S.; Hahn, J. H. Bull. Korean Chem. Soc. 2002, 23, 295-300. (10) Kim, S.; Kim, W.; Hahn, J. H. J. Chromatogr., A 1994, 680, 109-116. (11) Harrison, D. J.; Manz, A.; Fan, Z.; Lu ¨ di, H.; Widmer, H. M. Anal. Chem. 1992, 64, 1926-1932. (12) Effenhauser, C. S.; Manz, A.; Widmer, H. M. Anal. Chem. 1993, 65, 26372642. (13) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897. (14) Jacobson, S. C.; Hergenro¨der, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem. 1994, 66, 1114-1118. (15) Verpoorte, E.; Manz, A.; Lu ¨ di, H.; Bruno, A. E.; Maystre, F.; Krattiger, B.; Widmer, H. M.; van der Schoot, B. H.; de Rooij, N. F. Sens. Actuators, B 1992, 6, 66-70. (16) Salimi-Moosavi, H.; Jiang, Y.; Lester, L.; McKinnon, G.; Harrison, D. J. Electrophoresis 2000, 21, 1291-1299. (17) Liang, Z.; Chiem, N.; Ocvirk, G.; Tang, T.; Fluri, K.; Harrison, D. J. Anal. Chem. 1996, 68, 1040-1046. 10.1021/ac050420c CCC: $30.25

© 2005 American Chemical Society Published on Web 07/12/2005

Figure 1. Layout of the absorbance detection microchip. The insets show the three types of detection cells: cell A, unextended, on-column detection cell; cell B, uncollimated Z-shaped detection cell; cell C, collimated Z-shaped detection cell with a microlens (r ) 50 µm) and slits. The optical path length of each cell is 50 (cell A), 500 (cell B), and 500 µm (cell C), respectively.

The stray light can be minimized by collimating the divergent light from an excitation fiber and by restricting the collimated light within the cross section of the microchannel for detection. In this work, we have realized such a collimating system by integrating a microconvex lens in front of the excitation fiber and a rectangular aperture on each end of the detection microchannel. A CE microchip including these microoptical components has been fabricated by a simple, conventional photolithography replica molding method and by using a single-substrate material, poly(dimethylsiloxane) (PDMS).20-26 Since our collimating system is easy to integrate into a PDMS chip and significantly improves the concentration range of linearity, effective optical path length, and sensitivity, it has great potential to be a common part of extended optical path length absorbance detection cells for separation microchips. EXPERIMENTAL SECTION Microchip Design. Figure 1 shows a schematic of the channel structure of the microchip with three types of absorbance detection cells, A-C. Cell A is an unextended, on-column detection cell, cell B is an extended optical path length detection cell of conventional type, and cell C has the same path length as cell B, but it is equipped with an integrated light collimator consisting a collimating microlens and a pair of slit channels. Each cell has a (18) Mogensen, K. B.; Petersen, N. J.; Hu ¨ bner, J.; Kutter, J. P. Electrophoresis 2001, 22, 3930-3938. (19) Petersen, N. J.; Mogensen, K. B.; Kutter, J. P. Electrophoresis 2002, 23, 3528-3536. (20) Ro, K. W.; Lim, K.; Kim, H.; Hahn, J. H. Electrophoresis 2002, 23, 11291137. (21) Ro, K. W.; Chang, W. J.; Kim, H.; Koo, Y. M.; Hahn, J. H. Electrophoresis 2003, 24, 3253-3259. (22) Makamba, H.; Kim, J. H., Lim, K., Park, N., Hahn, J. H. Electrophoresis 2003, 24, 3607-3619. (23) Sia, S. K.; Whitesides, G. M. Electrophoresis 2003, 24, 3563-3576. (24) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M. Anal. Chem. 1998, 70, 4974-4984. (25) Ocvirk, G.; Munroe, M.; Tang, T.; Oleschuk, R.; Westra, K.; Harrison, D. J. Electrophoresis 2000, 21, 107-115. (26) Xiao, D.; Le, T. V.; Wirth, M. J. Anal. Chem. 2004, 76, 2055-2061.

Figure 2. Anatomic view of the collimated absorbance detection cell with an extended optical path length flow cell, a microlens, and a pair of slits (cell C). The unit of each length is micrometers.

pair of channels (125 µm wide × 127 µm deep) for inserting two optical fibers: one for delivering the excitation light and the other for collecting the transmitted light. The channel lengths from reservoirs 1-3 to the injection cross are 6, 9, and 9 mm, respectively, and the separation channel lengths from the injection cross to cells A-C are 15, 20, and 25 mm, respectively. These channels have a cross section of 50 µm wide × 27 µm deep. The optical path lengths of cells A-C are 50, 500, and 500 µm, respectively. The gaps between the fiber channels and the separation channels of cells A-C are 70, 70, and 270 µm, respectively. The radius of the collimating microlens in cell C is 50 µm. Microchip Fabrication. The microchip is fabricated by stacking three PDMS plates, top, membrane, and bottom (Figure 2). This fabrication has been achieved by modifying the “membrane sandwich” technique.27,28 Figure 2A anatomically shows how the optical axes of fibers can be aligned with the longitudinal axis of the flow cell and also how the collimating lens and the slits can be realized in cell C. As shown in Figure 2B, the membrane of 27-µm thickness defines the depths of separation channels and apertures of slits. The fiber channels and slit channels have a depth of 127 µm that is the sum of 27 and 100 µm (50 µm each from top and bottom plates). The apertures of the slits are formed by filling black ink into the stacked slit channels (Figure 2A), and they are (27) Anderson, J. R.; Chiu, D. T.; Jackman, R. J.; Cherniavskaya, O.; McDonald, J. C.; Wu, H.; Whitesides, S. H.; Whitesides, G. M. Anal. Chem. 2000, 72, 3158-3164. (28) Wu, H.; Odom, T. W.; Chiu, D. T., Whitesides, G. M. J. Am. Chem. Soc. 2003, 125, 554-559.

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5161

30 and 50 µm wide in inlet and outlet slits, respectively. The microlens has a cylindrical shape with 50-µm radius. To make the top and bottom plates, ∼50-µm-thick film of a negative photoresist (SU-8 50; MicroChem Corp., Newton, MA) is spin-coated on silicone wafers of 100-mm diameter (Siltron Inc.) and soft-baked to drive off solvent. The photoresist is exposed to UV light through a photomask, postexposure-baked, and developed in propylene glycol methyl acetate (Aldrich, Milwaukee, WI) to reveal a master with a positive relief pattern of a channel manifold. The master is silanized by placing it in a vacuum desiccator for 2 h along with a vial containing a few drops of a silanizing agent, trichloro(3,3,3-trifluoropropyl)silane (Aldrich). The silanization facilitates the removal of the PDMS replica from the master. A 10:1 mixture of PDMS oligomer and cross-linking agent (Sylgard 184; Dow Corning, Midland, MI) is poured onto the master and then degassed under vacuum. After 3 h of curing at 75 °C, a PDMS replica is peeled from the master to yield the negative relief of a channel network. The reservoirs for sample, buffer, black ink, etc., are defined on the top PDMS plate by punching holes at the ends of the flow and slit channels. The master for the membrane, a 27-µm-high positive relief pattern of a channel manifold, is made by the same fabrication method used for the masters of top and bottom plates and silanized. A flat slab of PDMS is oxidized by corona discharge generated from a Tesla coil (BD-10A; Electro Technic Products Inc., Chicago, IL) for 2 min, silanized, and then placed face down on top of the membrane master with a drop of PDMS prepolymer between. Pressure greater than 2 g/mm2 (20 kPa) is applied by placing a glass plate on top of the flat PDMS plate until prepolymer is excluded from between the PDMS plate and the membrane master, which are in contact. After 3 h of curing at 75 °C, the assembly of the membrane and the flat PDMS plate is peeled off from the membrane master. To transfer the membrane to the bottom plate, the bottom surface of the membrane and the bottom plate are oxidized with a Tesla coil, aligned under a CCD camera, and brought into a conformal contact. After complete irreversible bonding is accomplished between the membrane and the bottom plate, the bonded plate is peeled off from the flat PDMS plate. To complete the chip fabrication, the top plate and the top surface of the membrane are oxidized, aligned, and bonded irreversibly. Two rectangular apertures serving as slits are realized by filling black ink into the slit channels through reservoir 5. After filling the slit channels, the reservoirs are sealed with Parafilm to avoid evaporation of the ink. Optical fibers with core diameters of 3 (Spec Tran Specialty Optics Co., Avon, CT) and 50 µm (Fiberguide Industries Inc., Stirling, NJ) are employed for the source and the collection fibers, respectively. Their jacket and cladding diameters are 250 and 125 µm, respectively. Before being inserted into the fiber channels, the optical fibers are cleaved with a fiber cleaver (F-BK2, Newport, Irvine, CA), stripped of their jackets, and then etched slightly in an HF/HNO3 mixture (49% HF:70% HNO3 ) 10:1) to allow easy insertion. The fibers are inserted into the fiber channels using a homemade fiber positioner and a CCD camera. Reagents. All of the chemicals used in this study are of analytical grade. Fluorescein, orange II (acid orange 7), and new coccine (acid red 18) were purchased from Aldrich. Sodium borate decahydrate and sodium dodecyl sulfate (SDS) were obtained 5162


from Sigma (St. Louis, MO). All aqueous solutions were prepared by using water from a Milli-Q purification system (Millipore, Milford, MA). The running buffer for electrophoresis is 15 mM sodium borate (pH 10.0) with 25 mM SDS. Instrumentation. Microfluidics for injection and separation is controlled by using a computer-controlled high-voltage supply system consisting of a high-voltage power supply (MP5; Spellman High Voltage Electronics, Plainview, NY), a high-voltage relay (K45C332; Kilovac, Santa Barbara, CA), and a homemade voltage dividing system. A LabVIEW program (National Instruments, Austin, TX) written in-house and a multifunction I/O board (LabPC-1200; National Instruments) are used for instrument control and data acquisition. A 488-nm argon ion laser (Lexel 95, Lexel Laser Inc., Fremont, CA) is employed as the light source for absorbance detection, and the laser beam is coupled into an optical fiber using a single-mode fiber coupler (F-1015, Newport). The transmitted light is collected with an optical fiber and directed to a photomultiplier tube module (HC120-01; Hamamatsu, Bridgewater, NJ). The light power incident to the detector is ∼5 nW. CE Procedures. The microchannels in a microchip are preconditioned sequentially with 0.1 N NaOH, deionized water, and the running buffer for 5 min each, and then a sample solution is introduced at reservoir 1. Gated injection method20 is employed to inject a sample into the separation channel. During conditioning and separation, 1.0, 1.5, 0.4, and 0 kV are applied to reservoirs 1 (sample), 2 (buffer), 3 (sample waste), and 4 (buffer waste), respectively. For sample injection, the potential at the buffer reservoir is floated for 0.5 s, and then reapplied. The light beam path in a detection flow channel is monitored with a CCD camera by floating the potential at the buffer reservoir until the entire flow path is filled with a steady stream of a fluorescein solution. RESULTS AND DISCUSSION Comparison of Beam Path. Optical fibers are used in this work in order to obtain a more flexible and convenient coupling of light into the absorbance detection cell fabricated in the microchip. However, one of the disadvantages of using optical fiber for transmitting the light is that the light radiated from the excitation fiber is diverged. The divergent light passing through the absorbance detection cell causes stray light, poor sensitivity, and nonlinear response of a detector. This deleterious effect becomes worse in the extended long optical path length detection cell. To overcome the above limitation and improve the detection efficiency in the microchip, we fabricate a collimator integrated absorbance detection cell that can collimate the divergent light by integrating a microlens at the end of the excitation fiber insertion channel with two rectangular apertures at both ends of the detection flow cell as shown in Figure 2. In Figures 3 and 4, the beam paths passing through the extended flow cells are compared in the uncollimated Z-shaped detection cell (cell B) (Figure 3) and the collimator integrated Z-shaped detection cell (cell C) (Figure 4). Figure 3A shows a top view of detection cell B and the optical fibers before being illuminated by a laser. The optical axes of fibers are well aligned with the longitudinal axis of the flow cell. Since the ends of the fiber insertion channels made in the PDMS have relatively flat and soft walls as seen in the Figure 3, the optical fibers can be inserted to the ends of the channels without any gap. One of the advantages of the flat wall formed at the end of the PDMS

Figure 3. CCD images of an uncollimated Z-shaped detection cell (cell B). (A) Images of optical fibers and a flow cell before illumination. The core diameters of excitation and collection fibers are 3 and 50 µm, respectively. The Z-shaped flow cell has a 50- × 27-µm cross section and an optical path length of 500 µm. (B) The image of divergent beam path shown in an extended optical path length flow cell, which is filled with 50 µM fluorescein solution and illuminated by an argon ion laser (488 nm). (C) Optical simulation result by using an optical ray tracing method, which shows the divergent beam path in the extended flow cell.

channels is that much less optical distortion and scattering are generated in comparison to the curved glass walls formed by a chemical, isotropic etching method.17 Figure 3B shows a beam path illuminated by an argon ion laser (488 nm) through an optical fiber in an extended flow cell. The flow cell is filled with 50 µM fluorescein solution. Figure 3C shows the optical simulation result by using an optical ray tracing method with the help of an OSLO LT program (Lambda Research Co, Littleton, MA). Although a single-mode optical fiber with 3-µm core diameter and numerical aperture of 0.1 is used as the excitation fiber to reduce the divergence of light, light divergence is clearly observed in the extended flow cell as shown in Figure 3B and C. However, in a collimator integrated detection cell (cell C), this shortcoming has been significantly overcome, as shown in Figure 4. Figure 4A shows well-aligned centers of an extended 500-µmlong flow cell, two optical fibers, a pair of apertures, and a microlens (radius 50 µm). Moreover, it clearly indicates that the slit channels located in the membrane plate are disconnected for the transmission of light, but they are connected with other slit channels fabricated on the top and bottom plates. Threedimensional slit structures with apertures are formed by filling these slit channels with black ink, as shown in Figure 4B and C. Figure 4B shows the collimated beam path accomplished in an extended flow cell by means of a planoconvex microlens. The light beam illuminated by an argon ion laser passes perfectly through a 500-µm-long flow cell filled with a 50 µM fluorescein solution. Distortion of collimated light is not observed at the interface of PDMS wall and water in the flow cell due to the flat wall of the flow detection cell. To acquire the collimated beam as shown in Figure 4B, the exact position of the excitation fiber in the fiber channel is determined by simultaneously observing the beam shape in the flow cell and measuring the absorbance signal. When

the excitation fiber is positioned to the exact focal length of the microlens, the collimated light is formed in the flow cell and the absorbance signal reaches a maximum value, and the fiber is fixed at that position. The focal length (f) of a planoconvex lens for obtaining the collimated light is given by29

f ) n1R/(n2 - n1)

(1)

where n1 and n2 are the refractive indices of air and PDMS, respectively, and R is the radius of the lens. For n1 ) 1.0, n2 ) 1.41,30 and R ) 50 µm, the focal length (f) of the planoconvex lens for light collimation is 122 µm. Measurement of distance from the end of fiber to the lens in Figure 4B gives 120 ( 8 µm, which agrees well with the theoretical focal length of the planoconvex microlens. Figure 4D shows the result of ray tracing simulation that the collimated beam made from a PDMS planoconvex microlens passing through the entire extended flow cell. The focal length of the planoconvex lens for light collimation, 122 µm, is used in this optical simulation. However, since the microlens in cell C has the cylindrical lens structure due to fabrication constraints, the light collimation is obtained only in the horizontal direction, so the light in the vertical direction is still diverged. To block this diverged light effectively, specially designed three-dimensional slits are integrated in the detection cell C (Figure 2). The slits are realized by filling the black ink into the slit channels as shown in Figure 4. Only the light entirely passing through the flow cell can enter the aperture and be transmitted to the collection fiber for detection. Figure (29) Hecht, E. Optics, 3rd ed.; Addison-Wesley: Boston, 1998; Chapter 5. (30) Llobera, A.; Wilke, R.; Buttgenbach, S. Lab Chip 2004, 4, 24-27.


5163

Figure 4. CCD images of a collimated Z-shaped detection cell (cell C). (A) Images of a flow cell and slits before filling with black ink. The core diameters of excitation and collection fibers are 3 and 50 µm, respectively. The Z-shaped flow cell has a 50- × 27-µm cross section and an optical path length of 500 µm. (B) The images of the slits are after filling with ink and the collimated beam path shown in an extended optical path length flow cell, which is filled with 50 µM fluorescein solution and illuminated by an argon ion laser (488 nm). (C) The cross-sectional image of an outlet slit with an aperture (50 µm × 27 µm). (D) Optical simulation result by using an optical ray tracing method, which shows the collimated beam path in the extended flow cell.

4C shows a cross-sectional image of the outlet slit with an aperture having a 50 µm × 27 µm cross section. Evaluation of Cell Efficiencies. The absorption efficiency of a collimated detection cell (cell C) is compared with that of an uncollimated detection cell (cell B). Both flow cells are filled with a new coccine solution with concentrations ranging from 6 µM to 0.6 mM, and the resulting absorbances are monitored. New coccine is chosen for the test because it has a broad absorption band in the vicinity of the wavelength of an argon ion laser (488 nm). The calibration curve of new coccine obtained using an Agilent 8453 UV-visible spectrophotometer is shown by dotted line (Figure 5). The other two solid lines represent calibration curves for cell B and cell C fabricated in the microchip. The calibration curve for cell C shows relatively good linearity in the entire concentration range; however, cell B shows strong nonlinearity in the high-concentration range because of the higher amount of stray light. The presence of stray light causes the measured transmittance to be larger than normal value, because the majority of the stray light is less strongly absorbed by the analyte. As the analyte concentration increases, this effect progressively worsens and causes a negative deviation in the calibration curve (Figure 5), because the transmitted stray light becomes a larger fraction of the total transmitted light. Therefore, 5164


Figure 5. Dependence of the absorbance on the concentration of new coccine in a collimated (cell C) and an uncollimated (cell B) detection cells. The dotted line is obtained by measuring the absorbance using an Agilent 8453 UV-visible spectrophotometer.

a higher amount of stray light in cell B results in a stronger negative deviation character than in cell C.

The amount of stray light is expressed as the stray radiation fraction (fSR), defined as the fraction of light that does not entirely participate in the absorption process to the total incident light beam,31

fSR ) PSR/(PO + PSR)

(3)

If the readout signal due to the stray radiation is identical for the reference and analyte solution, the measured transmittance (T*) is given by

T* ) (Es + ESR)/(Er + ESR)

(4)

where Es is the readout signal from sample solution. Using eq 3, the eq 4 can be rewritten as

T* ) (T + R)/(1 + R)

detection cell

path length (µm)

effective path length (µm)

stray radiation readout raction (R), %

linear range (mAU)

collimated cell (cell C) uncollimated cell (cell B)

500 500

460 324

3.8 31.6

410 150

(2)

where PSR is stray radiant power and PO is the radiant power output. In the absorbance detection system, the stray radiation readout fraction (R) is defined as the relative contribution of stray radiation (ESR) to the ideal reference signal (Er),

R ) ESR/Er

Table 1. Comparison of Absorbance Cell Performances in an Uncollimated and a Collimated Z-Shaped Detection Cells

(5)

where T is the ideal transmittance when R ) 0. Thus, R can be expressed as

The linear range is calculated by fitting a second-order polynomial to the calibration curve at each detection cell and defined as the absorbance where the polynomial deviated 5% from the extrapolated line, which is obtained by fitting a straight line to the calibration curve at lower concentration.32 The effective path length of each cell was determined to be 324 µm for cell B and 460 µm for cell C. Table 1 compares effective path length, stray radiation readout fraction, and linear range in each flow detection cell. It can be noticed that despite having same optical path length (500 µm), cell B shows 35% reduction in sensitivity over the linear range, whereas, only 8% reduction is shown in cell C. Sensitivity Enhancement. The sensitivity of the collimated Z-shaped detection cell (cell C) and an unextended, on-column detection cell (cell A) is compared for the CE separation. Figure 6 shows the separation results of fluorescein (10 µM), orange II

R ) (T* - T)/(1 - T*) ) (10-A* - 10-A)/(1 - 10-A*) (6) where A* and A are the measured and ideal absorbances, respectively. From eq 6 and Figure 5, the stray radiation readout fraction (R) of each detection cell in the microchip can be determined by measuring the maximum absorbance (A*) at a high concentration of new coccine. In this work, measured absorbances (A*) in both detection cells are obtained at 0.6 mM concentration of new coccine solution. The ideal absorbance, A, is obtained by the Beer’s law,

A ) cL

(7)

where is the extinction coefficient, c is the concentration of analyte, and L is the optical path length. The extinction coefficient of new coccine is calculated to be 2.76 × 104 M-1 cm-1 by measuring the slope of calibration curve of the UV-visible spectrophotometer (Figure 5). The ideal absorbance at 0.6 mM concentration of the new coccine solution is calculated to be 0.83. The percentage of stray radiation readout fractions (R) calculated using eq 5 in cell B and cell C are 31.6 and 3.8%, respectively. This result demonstrates that the microlens and slits of cell C play a key role for significantly reducing the diverged and scattered lights reaching the detector. The effective path length that shows sensitivity in the linear range is calculated by dividing the slope in the linear range with the extinction coefficient of new coccine (2.76 × 104 M-1 cm-1). (31) Ingle, J. D., Jr.; Crouch, S. R. Spectrochemical Analysis, 1st ed.; Prentice Hall International, Inc: Englewood Cliffs, NJ, 1988; Chapter 13.

Figure 6. Comparison of detector sensitivity in an unextended, on-column detection cell (cell A) and a collimated Z-shaped detection cell (cell C). Running buffer: 15 mM Borax (pH 10), 25 mM SDS. Analytes: (1) fluorescein (10 µM), (2) orange II (120 µM), and (3) new coccine (60 µM). The effective separation channel lengths to cell A and cell C are 15 and 25 mm, respectively. The separation voltage between buffer reservoir and buffer waste reservoir is 1.5 kV.

(120 µM), and new coccine (60 µM) in cell C and cell A. Cell C shows an ∼10 times increase in sensitivity due to the 10 times longer optical path length than cell A. This result shows that the collimating system integrated in cell C certainly contributes to the increase in the effective path length of the flow cell, and the increase in detection sensitivity of the collimated detection cell can be accomplished in proportion to the optical path length of the flow cell. There is ∼5% decrease in the separation efficiency (32) Lindberg, P.; Hanning, A.; Lindberg, T.; Roeraade, J. J. Chromatogr., A 1998, 809, 181-189.


5165

of cell C as compared to that of cell A (N ) 8700 in cell C and 9200 in cell A for the peak 2 in Figure 6). The height equivalent to a theoretical plate contributed by the detector volume, Hdet, can be calculated from the following equation using the standard formula,33

Hdet ) l2/12did

(8)

where l is the optical path length of the detector cell and did is the distance from injector to detector. An optical path length of 500 µm in cell C will contribute 0.83 µm of plate height. From the plate height, 2.87 µm, given by peak 2 in cell C, the band broadening caused by the longer detector volume would contribute to the plate height of 29%. It is expected that this contribution can be decreased by using a tapering detection cell structure18,19 and a longer separation channel. The separation efficiency in cell C could be further improved by optimizing the injection volume and the detection cell geometry along with decreasing the plate height contributed by the detector volume. The calibration curve is determined by varying concentrations of three different dyes in the range of 1-500 µM in cell C. The plot of absorbance versus concentration of each dye is found to be linear, and the calculated R2 factors are 0.9989 for fluorescein, 0.9983 for orange II, and 0.9996 for new coccine. The concentration detection limits (S/N ) 3) for fluorescein, orange II, and new (33) Sternberg, J. C. Adv. Chromatogr. 1966, 2, 206-270.

5166


coccine are 1.2, 2.9, and 3.5 µM, respectively, at the absorbance detection limits of 0.001 AU. The concentration and absorbance detection limits in the collimator integrated detection cell (cell C) are better than those of the microfabricated U-cell in the glass substrate.17 The detection limit can be improved by incorporation of a bifurcated optical fiber and a second detector for a reference monitoring. CONCLUSIONS We have fabricated an integrated light collimating system for a highly efficient and sensitive absorbance detection of microchip CE in PDMS. This collimating system integrated with a microlens and a pair of slits provides an indispensable tool for extended optical path length absorbance detection by greatly reducing the stray light. We have obtained enhanced detection sensitivity in the collimated Z-shaped detection cell without reduction in the effective path length of a flow cell. The ability to use an absorbance detector for microchips is expected to greatly extend the applicability of these devices to various chemical, biological, and environmental samples. ACKNOWLEDGMENT This work has been financially supported by the R&D Program of Fusion Strategies for Advanced Technologies, the Ministry of Commerce, Industry and Energy, South Korea. Received for review March 10, 2005. Accepted June 4, 2005. AC050420C

Integrated Light Collimating System for Extended Optical-Path-Length

Recommend Documents