Side-Entry Laser-Beam Zigzag Irradiation of Multiple Channels in a

Aug 21, 2015 - A simple and highly sensitive technique for laser-induced fluorescence detection on multiple channels in a plastic microchip was develo...
0 downloads 0 Views 878KB Size
Download PDF
Technical Note pubs.acs.org/ac

Side-Entry Laser-Beam Zigzag Irradiation of Multiple Channels in a Microchip for Simultaneous and Highly Sensitive Detection of Fluorescent Analytes Takashi Anazawa,* Takahide Yokoi, and Yuichi Uchiho Hitachi, Ltd., Research and Development Group, 1-280 Higashi-koigakubo, Kokubunji, Tokyo 185-8601, Japan S Supporting Information *

ABSTRACT: A simple and highly sensitive technique for laser-induced fluorescence detection on multiple channels in a plastic microchip was developed, and its effectiveness was demonstrated by laser-beam ray-trace simulations and experiments. In the microchip, with refractive index nC, A channels and B channels are arrayed alternately and respectively filled with materials with refractive indexes nA for electrophoresis analysis and nB for laser-beam control. It was shown that a laser beam entering from the side of the channel array traveled straight and irradiated all A channels simultaneously and effectively because the refractive actions by the A and B channels were counterbalanced according to the condition nA < nC < nB. This technique is thus called “side-entry laser-beam zigzag irradiation”. As a demonstration of the technique, when nC = 1.53, nA = 1.41, nB = 1.66, and the cross sections of both eight A channels and seven B channels were the same isosceles trapezoids with 97° base angle, laser-beam irradiation efficiency on the eight A channels by the simulations was 89% on average and coefficient of variation was 4.4%. These results are far superior to those achieved by other conventional methods such as laser-beam expansion and scanning. Furthermore, fluorescence intensity on the eight A channels determined by the experiments agreed well with that determined by the simulations. Therefore, highly sensitive and uniform fluorescence detection on eight A channels was achieved. It is also possible to fabricate the microchips at low cost by plastic-injection molding and to make a simple and compact detection system, thereby promoting actual use of the proposed side-entry laser-beam zigzag irradiation in various fields.

O

density and achieve high sensitivity. Three methods of multicolor fluorescence detection on multiple channels have been reported so far: beam expansion,11−15 beam scanning,16−19 and beam side entry.20 In the case of each method, all channels in the laser-irradiation part are arrayed on the same plane. As for the beam-expansion method, to irradiate all the channels simultaneously, a laser beam is expanded either to form a linear beam across the channel array or to form a circular beam covering the channel array. In the linear case, when the number of channels is N, effective laser-power density for each channel is less than 1/N; however, in the circular case, it is less than 1/N2 as compared with the standard laser-power density (in which case, a focused laser beam with a similar diameter as a channel width is irradiated on the channel). In both methods, sensitivity of fluorescence detection significantly decreases with increasing N. As for the beam-scanning method, a focused laser beam is scanned across the channel array to irradiate all the channels sequentially. Effective laser-power density for each channel is less than 1/N because the duty cycle of the irradiation is less

ne of the most fundamental analytical methods used in the “lab on a chip” (LOC) field is highly sensitive analysis of a sample flowing in the channel in the LOC device (the microchip) by laser-induced multicolor fluorescence detection. This optofluidic analysis has been considerably advanced over the past two decades.1−3 It includes electrophoresis analysis and flow analysis of nucleic acids, proteins, and cells, which are both applied for DNA sequencing,4,5 immunoassay,6 flow cytometry,7,8 and single-cell analysis.9,10 To further improve these analyses and popularize them in fields such as medical practice, food inspection, and environmental monitoring, it is necessary to make it possible to simultaneously analyze multiple samples or multiple items in a single sample by highly sensitive multicolor fluorescence detection using a simple analytical system with a low-cost microchip containing multiple channels. Such low-cost microchips can be manufactured by massproduction technologies such as plastic-injection molding. Moreover, they can be disposable and thus prevent sample contamination and minimize maintenance by users. Multicolor fluorescence detection on multiple channels in microchips is performed by irradiating each channel with a laser beam, exciting various fluorophores flowing in each channel, and spectroscopically detecting the fluorescence emitted from each channel. The key challenge facing this detection is how to efficiently irradiate multiple channels with high laser power © 2015 American Chemical Society

Received: April 24, 2014 Accepted: August 18, 2015 Published: August 21, 2015 8623

DOI: 10.1021/acs.analchem.5b02222 Anal. Chem. 2015, 87, 8623−8628

Technical Note

Analytical Chemistry than 1/N. Mathies group cleared the issue by scanning a laser beam with much higher laser-power density than the standard laser-power density.18,19 This was achieved by adopting a higher power laser source and focusing the laser beam to a smaller diameter than each channel width. An epi-illumination confocal system was also employed to reduce laser scattering from the substrate of the microchip for highly sensitive fluorescence detection. The system was originally developed for fluorescence detection of a capillary array (not for a channel array in a microchip)21 and applied to commercial DNA sequencers (MegaBACE DNA Analysis Systems). However, the beamscanning method is more complex than the others because it inherently requires an actuating mechanism. As for the beam-side-entry method, a laser beam enters from the side of the channel-array plane to pass across the channel array and irradiate all the channels simultaneously. Although the method enables highly sensitive fluorescence detection with a simple system configuration, it is difficult to implement due to laser-beam refraction at the boundaries of the channels. To address this difficulty, one reported method20 inputs a laser beam with a diameter larger than the channel width from the side of the plane. As a result, laser-power density and fluorescence-detection sensitivity are lower than the standard values. Highly sensitive fluorescence-detection technologies for slabgel electrophoresis22 and capillary-array electrophoresis23,24 using the beam-side-entry method were previously developed and applied to commercial DNA sequencers. In particular, multiple-laser-beam focusing technique is implemented in a capillary-array DNA sequencer with up to 96 capillaries (Applied Biosystems Genetic Analyzers), where a side-entry laser beam is focused repeatedly (by utilizing the convex-lens action of each capillary) so that it passes across the capillary array and simultaneously irradiates all the capillaries with high power density. In other words, refraction of the laser beam at the boundaries of the capillaries is controlled. As a result, the DNA sequencer is highly sensitive and still widely used in various fields. However, multiple-laser-beam focusing technique cannot be applied to a channel array in a microchip because of differences between the geometries and component refractive indexes of the capillary array and channel array. In this present study, a novel technique, called “side-entry laser-beam zigzag irradiation” of multiple channels in a microchip, is proposed. With this technique, a side-entry laser beam is passed across a channel array in a plastic microchip and simultaneously irradiates all the channels with high power density. It is possible to fabricate such microchips at low cost by plastic-injection molding and to make a fluorescence-detection system simple and compact, thereby expanding actual use of this analysis in fields such as medical practice, food inspection, and environmental monitoring. The validity and usefulness of this technique are demonstrated by both laser-beam ray-trace simulations and experiments.

Figure 1. Configuration diagram of a microchip electrophoresisanalysis system employing side-entry laser-beam irradiation of multiple channels: (A) perspective view of microchip and (B) cross-sectional view of the microchip and multicolor fluorescence detection system. A channels (red) for electrophoresis analysis and B channels (yellow) for laser beam control are alternately arrayed in the microchip (blue). Samples are migrated from A-inlet port to outlet port in A channels and excited at side-entry laser-beam irradiation positions to emit fluorescence. Cross injection or double-T injection parts are omitted for simplicity.

of the plane perpendicularly to the channels. The laser-beam axis when the laser beam travels in a straight line within the channel array is defined as the “side-entry axis”, which is indicated by dashed lines in Figure 1. The channels are composed of those for electrophoresis analysis (i.e., A channels) and those for laser-beam control (i.e., B channels), which are alternately arrayed in the microchip. The microchip is made of material with refractive index nC, and the A and B channels are filled with materials with refractive indexes nA and nB, respectively. The materials of the microchip are indicated in blue and those of the A-channel and B-channel fillings are indicated in red and yellow. A- and B-inlet ports are formed at the entrances of A and B channels, respectively, whereas a common outlet port is formed at the exits of A and B channels. Fluorescence from all the channels is spectroscopically and simultaneously detected from the perpendicular direction to the plane. Fluorescence is collimated by a collecting lens, passed through an optical filter for laser-light rejection and a diffraction grating, and focused on a two-dimensional sensor by an imaging lens. Spectroscopic imaging data is processed to obtain time courses of multicolor fluorescence intensity on all the channels and to analyze multiple samples. A microchip was fabricated by plastic-injection molding and composed of ZEONEX (Zeon Corporation), that is, a cycloolefin polymer, with refractive index nC of 1.53. Eight A channels and seven B channels were alternately arrayed at 200 μm intervals. The A channels were numbered from A-ch 1 to Ach 8, and the B channels were numbered from B-ch 1 to B-ch 7, from the nearest channel to the incoming laser beam to the farthest channel from it, respectively. Cross-sectional shapes of all the channels were isosceles trapezoids with lower bases and heights of 40 μm, base angles (B) of 97°, and thereby upper bases of around 49.8 μm. In other words, draft angle (D = B − 90) in plastic-injection molding was 7°. A channels were filled with 70% formamide aqueous solution (70% FA) or 70% FA containing 100 nM dR110 dye-labeled C primer or 100 nM dROX dye-labeled T primer prepared from

■

EXPERIMENTAL SECTION Fluorescence Detection on Multiple Channels. The configuration of a system for microchip electrophoresis analysis employing side-entry laser-beam irradiation of multiple channels is shown in Figure 1. The axes of the channels in the parts of the channels irradiated by the laser beam are densely arrayed in parallel at regular intervals in the same plane. (The same plane is referred to as the “channel-array plane”.) The laser beam in the channel-array plane is input from the side 8624

DOI: 10.1021/acs.analchem.5b02222 Anal. Chem. 2015, 87, 8623−8628

Technical Note

Analytical Chemistry

Figure 2. Results of laser-beam ray-trace simulation for designs (a), (b), (c), and (c′). In designs (a) and (b), eight A channels (red) arrayed in the same plane in the microchip (blue) at 400 μm intervals, where nC = 1.53 and nA = 1.41 (nA < nC). In designs (c) and (c′), eight A channels and seven B channels (yellow) are alternately arrayed at 200 μm intervals, where nC = 1.53, nA = 1.41, and nB = 1.66 (nA < nC < nB) in design (c), whereas nC = 1.53, nA = 1.41, and nB = 1.41 (nA = nB < nC) in design (c′). A laser beam with 40 μm diameter is input from the left side of the channel array.

As the first part of the simulation, a laser beam with wavelength of 505 nm was focused to a 40 μm diameter to enter the left side of the microchip perpendicularly along the channel-array plane, where the central axis of the laser beam coincided with the side-entry axis. All refractive indexes indicated in this paper were adjusted to those at 505 nm. The polarizing axis of the laser beam was parallel to the channel-array plane; i.e., an s-polarized laser beam was irradiated on all channels. The laser beam was composed of 300 beam elements with infinitesimal diameter randomly and uniformly distributed in the 40 μm-diameter cross-sectional area of the focused beam and parallel to the side-entry axis, and in all simulations described in this paper, the distribution of elements was constant. Total laser beam power just before the entrance of the microchip was defined as 1 (100%), so each beam element had equal power of 1/300 (0.33%). As the next part of the simulation, the traveling direction and power of the refracted ray of each beam element were sequentially tracked by Snell’s law and Fresnel’s law on every beam-entrance and beam-exit planes of the microchip and each channel. That is, incident and output angles, reflectance, and transmittance of each element (on each plane) of the spolarized laser beam were calculated. Representative refraction indexes, εA in eq 7 and εB in eq 8 in the Supporting Information, are not used in this tracking. These planes include not only left and right sides but also upper and lower bases of the microchip and each channel. When total reflection occurred, traveling direction and power of the reflected ray were tracked instead of those of the refracted ray. Finally, beam elements that pass through the inside of each channel were extracted and their powers at each channel were summed to obtain irradiation efficiency of the laser beam for each channel. The laser-beam irradiation efficiency on a channel is standardized so that it is 100% when all beam elements irradiate the channel without any power attenuation, and it is proportional to fluorescence intensity from the channel. Although the above-described ray-trace was simulated by the 3D model, the simulation results are shown as projected 2D images for ease of comprehension.

ABI Prism BigDye Primer Cycle Sequencing Ready Reaction Kit (Life Technologies Corporation). In this study, 70% FA was used as a model for POP-7 polymer (Life Technologies Corporation), that is, aqueous solution for electrophoresisbased DNA sequencing, because the refractive indexes of both solutions are 1.41 (i.e., nA = 1.41). On the other hand, B channels were filled with 70% FA (i.e., nB = 1.41) or highrefractive-index liquid (HRI liquid) with refractive index of 1.66 (i.e., nB = 1.66). The HRI liquid, a transmissive and uncured resin adhesive, was custom-ordered from NTT Advanced Technology Corporation. A 505 nm laser beam (Showa Optronics Co., Ltd.) was focused by a lens with f 60 mm, and it entered the side of the microchip at a power of 15 mW. The polarizing axis of the laser beam was parallel to the channel-array plane; i.e., the spolarized laser beam was irradiated on all channels. Both collecting lens and imaging lens were camera lenses with f 50 mm and F/1.4. A CCD 2D sensor (Hamamatsu Photonics K.K.) continuously acquired fluorescence images on eight A channels with exposure time of 200 ms. Laser-Beam Ray-Trace Simulation for Multiple Channels. A 3D-laser-beam ray-trace simulation was carried out using LightTools (Synopsys’ Optical Solutions Group) to demonstrate effectiveness of the proposed concept of side-entry laser-beam irradiation (Supporting Information). Only the region of microchip surrounding the side-entry axis was modeled as explained below. The microchip was a cuboid shape placed in air. Its widths were 3.8 mm in the side-entry axis direction and 1.0 mm in the channel-axis direction, and thickness was 1.0 mm in the vertical direction to the side-entry axis and channel-axis directions. Here, the channel-array plane was in the center of the microchip in the vertical direction. Eight A channels were arrayed at 400 μm intervals, or eight A channels and seven B channels were alternately arrayed at 200 μm intervals. Cross-sectional shapes of all the channels were squares with 40 μm sides or isosceles trapezoids with lower bases and heights of 40 μm and base angles of 97°. The microchip was composed of a plastic with refractive index nC of 1.53. A channels were filled with a medium with refractive index nA of 1.41. B channels were filled with a medium with refractive index nB of 1.41 or 1.66. 8625

DOI: 10.1021/acs.analchem.5b02222 Anal. Chem. 2015, 87, 8623−8628

Technical Note

Analytical Chemistry

■

RESULTS AND DISCUSSION

Side-entry laser-beam irradiation of multiple channels with circular cross sections was studied before that with quadrangular cross sections (see the Supporting Information). It was found that the circular cross section was appropriate to control the lens actions of the channels and effectively irradiate multiple channels. However, it is difficult to fabricate such a circular cross section by mass-production technologies such as plasticinjection molding, at least, at this moment, although it is possible to fabricate it by small-quantity-production technologies such as 3D printing. In the case of usual mass production of microchips, two plastic plates, where one has multiple grooves on its surface and the other is flat, are molded and then bonded together to make the grooves into channels. In this case, the cross section of the channels is a tapered form (such as a trapezoid or a semicircle) toward the bottoms of grooves, because a draft angle of more than zero is necessary to easily extract the mold from the grooved plate. Positional error between the two plates in the bonding process does not affect their cross-sectional shapes and the intervals between channels. On the other hand, two plastic plates both with multiple grooves (whose cross sections are semicircles with the same diameter) are bound together to make grooves into channels with circular cross sections. However, it is often the case that the semicircle axis on one plate does not match the corresponding semicircle axis on the other plate; that is, a complete circle is not obtained, because the two plates cannot be bound together with micrometer-level positional accuracy. Such an incompletely circular cross section is not preferable for either side-entry laser-beam irradiation or electrophoresis. The results of the laser-beam ray-trace simulations for designs (a) to (c) (Figure S-1) are shown in Figure 2. Each design shows the central part of the cross-section of the microchip (indicated in blue), including eight A channels (red) for designs (a) and (b) or eight A channels (red) and seven B channels (yellow) for designs (c) and (c′). The laser beam (black arrow) was input from the left side of the microchip and composed of 300 beam elements (black lines). Although the diameter of the laser beam on the left side of A-ch 1 (i.e., before it enters the microchip and A-ch 1) was kept at 40 μm, the patterns of the laser beam on the right side of A-ch 1 (i.e., after it leaves A-ch 1) in all the simulations were different. Figure 3 plots laser-beam irradiation efficiency against A-ch number for each design. In design (a) in Figure 2, cross-sectional shapes of eight A channels (A-ch 1 to 8) are squares with 40 μm sides. However, it is difficult to fabricate such channels by mass-production technologies such as plastic-injection molding, because the draft angle for square channels is zero. Incidence angles of all beam elements were always zero, and no laser beam refraction occurred at boundaries of A channels. Consequently, the elements traveled in straight lines parallel to the side-entry axis and contributed to irradiation of all A channels. Ideal laserbeam irradiation efficiency against A-ch number was obtained as shown by closed-square plots in Figure 3, where slight attenuation of the efficiency, i.e., approximately 0.33% decline per A channel, was accounted for only by laser-beam reflection loss at the boundaries of A channels. The attenuation rate of 0.33% is also calculated as follows. Transmittances for normal incident light at left and right sides of an A channel are 99.833% (4nAnC/(nA + nC)2 = 4 × 1.41 × 1.53/(1.41 + 1.53)2); therefore, the attenuation rate by the A channel is 100% −

Figure 3. Laser-beam-irradiation efficiency against A-channel number for designs (a), (b), (c), and (c′). Conditions are the same as those for Figure 2.

99.833%2 = 0.334%. The efficiency of 95% at A-ch 1 is accounted for by laser-beam reflection loss at the entrance of the microchip placed in air. (Similarly, 4n0nC/(n0 + nC)2 = 4 × 1.00 × 1.53/(1.00 + 1.53)2 = 95.612%, where n0 is refractive index of air.) As a result, irradiation efficiency was 93% (95% × (100% − 0.33%)7) at A-ch 8 and more than 93% at all other A channels. Average and coefficient of variation of laser-beam irradiation efficiency for eight A channels were 94% and 0.8%, respectively. Design (a) is therefore ideal for side-entry laserbeam irradiation of eight A channels. In design (b) in Figure 2, cross-sectional shapes of eight A channels were isosceles trapezoids with lower bases and heights of 40 μm, and base angles (B) of 97°, i.e., D = 7°. As shown in the Supporting Information, net refraction angle by a single A channel (εA) was calculated as −1.1° from eq 7, where D = 7°, nC = 1.53, and nA = 1.41. Because εA < 0, the laser beam refracted downward. Furthermore, refraction angle was increased by increasing the number of A channels the laser beam passed through. Therefore, the laser beam deflected at an increasing rate from the side-entry axis, or the channel array, and deviated completely at A-ch 4 and beyond. As a result, laser-beam irradiation efficiency decreased drastically from A-ch 2 to 4 and was zero at A-ch 4 and beyond, as shown by closedtriangle plots in Figure 3. In design (b), it is concluded that no more than two A channels can be effectively irradiated by the side-entry laser beam. The proposed technique, namely, side-entry laser-beam zigzag irradiation, is validated in design (c) in Figure 2. B channels (B-ch 1 to 7) with the same cross section and refractive index nB of 1.66 were alternately placed in between A channels (A-ch 1 to 8) in design (b). The relation nA (1.41) < nC (1.53) < nB (1.66) was satisfied to counterbalance refractions by the A and B channels. Net refraction angle by a single B channel (εB) was calculated as 1.2° from eq 8, where D = 7°, nC = 1.53, and nB= 1.66, whereas net refraction angle by a single A channel (εA) was −1.1°. Consequently, net refraction angle by a pair of A and B channels was εA + εB = 0.1°, whose absolute value |εA + εB| = 0.1° was much smaller than that of net refraction angle by a single A channel, |εA| = 1.1°. For this reason, unlike in the case of design (b), the largest part of the side-entry laser beam traveled along the side-entry axis and passed through and effectively irradiated all A and B channels. 8626

DOI: 10.1021/acs.analchem.5b02222 Anal. Chem. 2015, 87, 8623−8628

Technical Note

Analytical Chemistry As a result, laser-beam irradiation efficiency at all A channels was more than 83%, as shown by closed circle plots in Figure 3. Average and coefficient of variation of laser-beam irradiation efficiency for eight A channels were 89% and 4.4%, respectively. Therefore, laser-beam irradiation efficiency for design (c) is comparable to that for design (a), shown by closed-square plots. Design (c) can achieve highly sensitive and simultaneous fluorescence detection on eight A channels. On the other hand, it was found by a closer look at the laser-beam trajectory in the channel array that the laser beam was oscillated slightly up and slightly down by B and A channels, respectively. Side-entry laser-beam zigzag irradiation is therefore an appropriate name for this beam propagation. As shown in Figure 2, compared to refractive index nB of 1.66 in design (c), that in design (c′) was decreased to 1.41, i.e., nB = 1.41. Because εA = εB = −1.1°, net refraction angle by a pair of A and B channels was εA + εB = −2.2°. The laser beam was therefore deflected from the channel array at a higher rate than that for design (b). As a result, as shown by the open-circle plots in Figure 3, laser-beam-irradiation efficiency decreased drastically from A-ch 1 to 3 and was zero at A-ch 4 and beyond. With design (c′), along with design (b), it is possible to irradiate no more than two A channels effectively by the sideentry laser beam. The above results also mean εA in eq 7 and εB in eq 8 are adequate indexes of laser-beam refraction in the channel array. Irradiation efficiency of the s-polarized laser beam was simulated. Although, in general, laser-beam reflection loss for a p-polarized laser beam is smaller than that for an s-polarized laser beam, the difference is negligible for all designs in Figure 2, because incident angles of the laser beam at the boundaries of all channels are very small. Figure 4 shows a series of experimental results of fluorescence detection on eight A channels for designs (c) and (c′). Each line indicates a fluorescence-intensity profile at 546 nm (maximum fluorescence-emission wavelength of dR110) along the side-entry axis. The eight peak positions of the green and red solid lines correspond to eight A-channel positions; i.e., the leftmost to rightmost peaks correspond to Ach 1 to 8, respectively. As for the first result, the background-intensity profile for design (c) (indicated by a black solid line) was obtained when A- and B-channels were filled with 70% FA with 0-nM dR110 (nA = 1.41) and HRI liquid (nB = 1.66), respectively. As for the second result, the fluorescence-intensity profile for design (c) (indicated by a green solid line) was obtained when the Achannel fillings were replaced with 70% FA with 100 nM dR110 (nA = 1.41). Fluorescence intensity on each A channel was nearly equal (slightly decreased against A-channel number) and comparable to that obtained by using a capillary under the same condition. As for the third result, the fluorescence-intensity profile for design (c′) (indicated by a blue solid line) was obtained when the B-channel fillings were replaced with 70% FA with 0 nM dR110 (nB = 1.41). Fluorescence intensity drastically decreased against increasing number of A-channels, whereas fluorescence intensity on A-ch 1 was nearly the same as that for design (c) (green solid line). As for the fourth result, fluorescence intensity profile for design (c) (indicated by a red solid line) was obtained when the B-channel fillings were returned to HRI liquid (nB = 1.66). It was restored to a similar profile to that shown by the green solid line. Small fluctuation in fluorescence intensity (e.g., the red peak is smaller than the

Figure 4. A series of experimental results of fluorescence detection on eight A channels in the same microchip for designs (c) and (c′). Black and green solid lines indicate background and dR110 fluorescence intensity across the side-entry axis for design (c). Eight peaks of the green solid line correspond to eight A channels. Blue solid line indicates fluorescence intensity for design (c′). Green- and blue-cross plots and dashed lines indicate simulated fluorescence intensity for designs (c) and (c′), respectively. Red solid line indicates dR110 fluorescence intensity for the second set of results for design (c). Experimental conditions are the same as the simulation conditions in Figure 2.

green peak on A-ch 2) was due to inhomogeneous fluorophore concentration. As for the fifth result, a similar fluorescence intensity profile at 616 nm (maximum fluorescence-emission wavelength of dROX) as that indicated by the green solid line in Figure 4 was obtained when the A-channel fillings were replaced with 70% FA with 100 nM dROX (Figure S-6). Signal-to-noise ratio (S/ N) on A channels was ∼9000, where S was averaged fluorescence intensity in the range of 600−630 nm, and N was standard deviation of averaged background intensity in the same range when A-channel fillings were 70% FA with 0 nM dROX. Therefore, limit of detection was estimated to be ∼22 pM dROX with a S/N of 2. The fluorescence detection sensitivity is high and comparable to that by the scanning confocal detection system.18,19 On the other hand, in Figure 4, for design (c), green-cross plots and the dashed line indicate the simulated fluorescence intensity on eight A channels (SF(n), where n is A-channel number). The intensity was calculated as SF(n) = E(n)/E(1) × (F(1) − B(1)) + B(n), where E(n) is given as laser-beam irradiation efficiency on eight A channels for design (c) (closedcircle plots) in Figure 3, F(1) is given as fluorescence intensity on A-ch 1 (green-solid line), and B(n) is given as background intensity on eight A channels (black-solid line) in Figure 4. Similarly, blue-cross plots and the dashed line indicate the simulated fluorescence intensity on eight A channels for design (c′). The intensity was calculated from laser-beam irradiation efficiency on eight A channels for design (c′) (open-circle plots) in Figure 3 and from fluorescence intensity on A-ch 1 (green-solid line) and background intensity on eight A channels (black-solid line) in Figure 4. 8627

DOI: 10.1021/acs.analchem.5b02222 Anal. Chem. 2015, 87, 8623−8628

Analytical Chemistry

■

Figure 4 shows that the simulated fluorescence intensities agree well with experimentally measured ones. This agreement confirms the side-entry laser beam actually passes across the channel array and irradiates all A channels simultaneously, effectively, and equally, as shown in design (c) in Figure 2, where refractions by A and B channels are counterbalanced according to the relation nA (1.41) < nC (1.53) < nB (1.66). On the other hand, the side-entry laser beam actually deviates from the channel array and does not irradiate all A channels, as shown in design (c′) in Figure 2, where refractions by A and B channels work synergistically according to the relation nA (1.41) = nB (1.41) < nC (1.53).

CONCLUSIONS A simple and highly sensitive method of fluorescence detection on multiple channels in a plastic microchip was proposed, and its effectiveness was demonstrated by laser-beam ray-trace simulations and experiments. It is possible to fabricate the microchips at low cost by plastic-injection molding. The metal mold for the grooved plate, the lower part of the microchip, should have the appropriate draft angle, be highly polished, and have multiple ejector pins. These are important not only to improve mold releasability but also to reduce surface roughness of channels without deteriorating the fine structure of the channels. Bonding between the upper and lower parts of the microchip should be carried out by thermocompression without any adhesive material other than the microchip material, because the laser beam partly passes through the boundary. The proposed technique is applicable to not only electrophoresis analysis but also any laser-irradiation-based analysis in microchips, which includes PCR, immunoassay, flow cytometer, and single-cell analysis. It is also possible to make a fluorescence-detection system simple and compact, thereby expanding actual use of this analysis in fields such as medical practice, food inspection, and environmental monitoring. ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b02222. Concept of side-entry laser-beam zigzag irradiation; other factors influencing fluorescence detection; study on sideentry laser-beam irradiation of multiple channels with circular cross sections; experimental results of dROX fluorescence detection on eight channels with trapezoidal cross sections (PDF)

■

REFERENCES

(1) Psaltis, D.; Quake, S. R.; Yang, C. Nature 2006, 442, 381−386. (2) Gai, H.; Li, Y.; Yeung, E. S. Top. Curr. Chem. 2011, 304, 171− 201. (3) Ban, E.; Song, E. J. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2013, 929, 180−186. (4) Fredlake, C. P.; Hert, D. G.; Kan, C. W.; Chiesl, T. N.; Root, B. E.; Forster, R. E.; Barron, A. E. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 476−481. (5) Shi, Y. Electrophoresis 2006, 27, 3703−3711. (6) Huang, Y.; Zhao, S.; Shi, M.; Liu, J.; Liang, H. Anal. Chim. Acta 2011, 694, 162−166. (7) Skommer, J.; Akagi, J.; Takeda, K.; Fujimura, Y.; Khoshmanesh, K.; Wlodkowic, D. Biosens. Bioelectron. 2013, 42, 586−591. (8) Piyasena, M. E.; Graves, S. W. Lab Chip 2014, 14, 1044−1059. (9) Huang, N. T.; Zhang, H. L.; Chung, M. T.; Seo, J. H.; Kurabayashi, K. Lab Chip 2014, 14, 1230−1245. (10) Tung, Y. C.; Huang, N. T.; Oh, B. R.; Patra, B.; Pan, C. C.; Qiu, T.; Chu, P. K.; Zhang, W.; Kurabayashi, K. Lab Chip 2012, 12, 3552− 3565. (11) Dang, F.; Shinohara, S.; Tabata, O.; Yamaoka, Y.; Kurokawa, M.; Shinohara, Y.; Ishikawa, M.; Baba, Y. Lab Chip 2005, 5, 472−478. (12) Dang, F.; Tabata, O.; Kurokawa, M.; Ewis, A. A.; Zhang, L.; Yamaoka, Y.; Shinohara, S.; Shinohara, Y.; Ishikawa, M.; Baba, Y. Anal. Chem. 2005, 77, 2140−2146. (13) Okagbare, P. I.; Soper, S. A. Analyst 2009, 134, 97−106. (14) Nikcevic, I.; Piruska, A.; Wehmeyer, K. R.; Seliskar, C. J.; Limbach, P. A.; Heineman, W. R. Electrophoresis 2010, 31, 2796−2803. (15) Liu, Z.; Shi, H.; Liu, L.; Deng, S.; Ji, Y.; Ma, S.; Ma, H.; He, Y. Sensors 2011, 11, 10038−10047. (16) Huang, Z.; Munro, N.; Huhmer, A. F.; Landers, J. P. Anal. Chem. 1999, 71, 5309−5314. (17) Cheng, S. B.; Skinner, C. D.; Taylor, J.; Attiya, S.; Lee, W. E.; Picelli, G.; Harrison, D. J. Anal. Chem. 2001, 73, 1472−1479. (18) Shi, Y.; Simpson, P. C.; Scherer, J. R.; Wexler, D.; Skibola, C.; Smith, M. T.; Mathies, R. A. Anal. Chem. 1999, 71, 5354−5361. (19) Scherer, J. R.; Liu, P.; Mathies, R. A. Rev. Sci. Instrum. 2010, 81, 113105. (20) Emory, J. M.; Soper, S. A. Anal. Chem. 2008, 80, 3897−3903. (21) Huang, X. C.; Quesada, M. A.; Mathies, R. A. Anal. Chem. 1992, 64, 2149−2154. (22) Kambara, H.; Nishikawa, T.; Katayama, Y.; Yamaguchi, T. Bio/ Technology 1988, 6, 816−821. (23) Anazawa, T.; Takahashi, S.; Kambara, H. Anal. Chem. 1996, 68, 2699−2704. (24) Anazawa, T.; Takahashi, S.; Kambara, H. Electrophoresis 1999, 20, 539−546.

■

■

Technical Note

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Tomoyuki Sakai, Kunio Harada, and Toru Habu of Hitachi, Ltd. for their kind support in the simulations and experiments. We also thank Dr. Takanobu Haga and Dr. Kenichi Takeda of Hitachi, Ltd. for their valuable discussions on the manuscript. We are grateful to Motohiro Yamazaki and Yoshitaka Kodama of Hitachi High-Technologies Corporation for the facilitation of this research. 8628

DOI: 10.1021/acs.analchem.5b02222 Anal. Chem. 2015, 87, 8623−8628