Multiple internal reflection poly(dimethylsiloxane ... - Semantic Scholar
PAPER
www.rsc.org/loc | Lab on a Chip
Multiple internal reflection poly(dimethylsiloxane) systems for optical sensing A. Llobera,*ab S. Demming,a R. Wilkea and S. Bu¨ttgenbacha Received 23rd March 2007, Accepted 24th July 2007 First published as an Advance Article on the web 21st August 2007 DOI: 10.1039/b704454b Compact poly(dimethylsiloxane)-based (PDMS) multiple internal reflection systems which comprise self-alignment systems, lenses, microfluidic channels and mirrors have been developed for highly sensitive absorbance measurements. With the proper definition of air mirrors at both sides of the sensing region, the optical path of the light from the LED has been meaningfully lengthened without a dramatic increase of the mean flow cell volume. By recursive positioning of such air mirrors, propagating multiple internal reflection (PMIR) systems have been designed, simulated and characterized. Experimental results confirm the ray-tracing predictions and allow the determining that there are some regions of the mean flow cell volume that do not contribute to the increase of the sensitivity. The tailoring of the sensing region, following the optical path, results in a similar limit of detection (110 nM) for fluorescein diluted in phosphate buffer. Finally, a ring configuration, labelled RMIR, has also been developed. With the addition of a third air mirror, the LOD can be decreased to 41 nM with the additional advantage of a substantial decrease of the length of the sensing region. These results confirm the validity of the proposed systems for high sensitivity measurements.
Introduction Currently, there is a huge demand for lab on chip systems with a low limit of detection (LOD) of (bio)chemical substances in a large variety of fields, such as molecular diagnosis, proteomics or monitoring of toxins. Ideally, such systems should have a high sensitivity, repeatability, reversibility, quick response, small size, low LOD, portability and low cost. To fulfil such strict requirements, several types of transduction mechanisms primarily focused on obtaining increasingly sensitive systems have previously been reported. Among these, optical,1 and mechanical2 transducers are the most commonly reported. Transducers based on the detection of a given optical magnitude (amplitude, phase or wavelength, mainly) have so far presented the highest sensitivity. As an example, optical transducers based on phase modulation with a Mach–Zehnder Interferometer (MZI) configuration have achieved a LOD of DNA hybridization of 300 pM without requiring labelling3 and with a high degree of integration. Nevertheless, the use of nanometre-scale optical structures (such as single mode silicon nitride waveguides) demands the use of expensive equipment that causes a dramatic increase of the cost per system, making it difficult to meet the low cost requirement. The tackling of this drawback is associated with the use of polymers, and especially poly(dimethylsiloxane) (PDMS). Taking advantage of its optical and structural properties, it is possible to monolithically define microfluidics and optical a
Institut fu¨r Mikrotechnik, Technische Universita¨t Braunschweig, Alte Salzdahlumer Straße 203, 38124 Braunschweig, Germany. E-mail: [email protected]; Fax: +49 531/391-9751; Tel: +49 531/ 391-9752 b Centre Nacional de Microelectro`nica. Campus UAB, 08193 Bellaterra, Spain
1560 | Lab Chip, 2007, 7, 1560–1566
structures. Such excellent properties have favoured their useage in laser-induced-fluorescence systems,4 cytometers5 as well as sorters and mixers.6 Moreover, with an accurate design7 or by doping PDMS with the appropriate substance,8 multiplexing or filtering capabilities can be transferred to the system. In addition, its biocompatibility and its low cost are two more advantages that speak for this material. Considering the authors’ previous work using PDMS-based lab on chip systems with multiplexing properties, the hollow prisms have shown their capability to distinguish between excited and emitted wavelength7 in fluorescence measurements, together with their ability to colorimetrically measure changes in the absorption spectra either due to variations of the concentration of a given substance or due to pH variations.9 Optically, the hollow prisms are examples of single internal reflection structures (SIR), since its working principle only requires one reflection. In opposition, multiple internal reflection (MIR) systems have already been presented for infrared spectroscopy for quantitative study of organic contamination of silicon wafers,10 absorption measurement11 and the Brownian motion of fluorescent microspheres.12 Generally, MIR systems are suitable to provide higher sensitivity than SIR systems. Nevertheless, it has several issues that need to be addressed and solved prior to its implementation in lab on chip systems. Firstly, with each additional reflection there is a decrease of intensity, which results in a decrease of the signal-to-noise ratio (SNR). As an example, in the Abbe prism presented in ref. 9, the reflection coefficients of the PDMS-buffer interface were 0.029 for TE-polarization and 0.062 for its TM counterpart. Hence, long integration times were required to obtain a clear signal even in systems with a single reflection, resulting in a low throughput of the hollow prisms and making the implementation of MIR structures with This journal is ß The Royal Society of Chemistry 2007
such low reflection coefficients unaffordable. Secondly, light inherently diverges. Then, as the optical path increases, there is an enlargement of the diameter beam. Therefore, an increase of the optical path beyond a given value generally causes a degradation of the properties of the proposed system, since a fraction of light is not collected by the photodetectors or fiber optics, which have finite dimensions, causing a dramatic decrease of the SNR. In this paper, we have addressed and solved the previously mentioned drawbacks to develop two highly-sensitive configurations of MIR structures, labelled propagating multiple internal reflection (PMIR) and ring multiple internal reflection (RMIR). Both configurations can be obtained in a single technological step and significantly decrease the LOD as compared to SIR structures.9
Structures Design As it was previously mentioned, the improvement of the reflectivity is a primary issue to enhance the performance of any reflection-based photonic lab-on-a-chip system. One possibility for improvement would be to define a mirror that reflects back the outcoming light from the hollow prism and thus achieving its collection by the output fiber optics. Although silicon or metal-coated mirrors can be easily defined, this would lead to a more complex fabrication process and an increase of the cost of the device. The ideal solution would be to obtain a mirror without any additional process steps. This objective can be achieved with the materials used (air, PDMS and buffer solution), if what we have called air mirrors are implemented, as shown in Fig. 1. These structures can be structurally envisaged as air entrapments with a concrete shape and position so as to modify the light path. In the same figure, the ray-tracing simulations of such structures are presented. In the SIR structures presented in ref. 7 and 9, which did not have an air mirror, light was propagating from a media with low refractive index (RI) (n = 1.334) to PDMS (with a higher RI, n = 1.41). Therefore, most of the light was coupled to the PDMS and did not reach the collecting fiber optics. Conversely, as it can be observed in Fig. 1, when the air mirror is included, the light coupled to the PDMS reaches a region with lower RI (air, n = 1.00). If this light has the appropriate incidence angle, total internal reflection (TIR) at the PDMS–air layer is obtained, resulting in a complete reflection of the light towards the PDMS. Considering the refractive indices of the media involved (PDMS and air) and using the Snell’s law, the critical propagation angle is hcPDMS-air = 45.17u. All the propagation angles h . hcPDMS-air (marked in Fig. 1) cause the light to be under TIR conditions and reflected back to the PDMS, reaching the hollow structure. Considering again the RI of the materials involved (PDMS and phosphate buffer (PBS) in this case), in this boundary region the critical angle is hcPDMS-PBS = 70.60u (the angles where the TIR regime happens are also marked in the figure). As can be observed in Fig. 1, the propagation angles match with the TIR regime at the PDMS–air region, but do not at the PDMS–PBS interface. This is the optimal situation, since then light does not remain confined at the PDMS and is This journal is ß The Royal Society of Chemistry 2007
Fig. 1 Ray tracing simulation of the propagating light at the air mirror vicinity, taking into account the refractive index values of air (1.00) poly(dimethylsiloxane) (PDMS, 1.41) and buffer solution (PBS, 1.334). The angles where the condition of total internal reflection (TIR) holds at both PDMS–air and PDMS–PBS interfaces are also shown.
coupled back to the MIR system. In addition, the shape of the air mirror can modify the behaviour of the optical beam; that is, from the initial parallel beams, an adequate curvature of the air mirror make the light converge inside the hollow structure, as is also shown in the ray-tracing diagram of Fig. 1. Hence, with this simple structure, the two most significant issues of the MIR configuration (reflection and beam broadening) can be addressed and solved. Therefore, the optical path can now be enlarged; enhancing the readout signal obtained and at the same time reducing the integration time, as will be presented in the Characterization section. After designing the air mirrors, the configurations of the PMIR and RMIR systems are presented in Fig. 2a and 2b, respectively. From a fluidic point of view, the proposed systems consist of a hollow structure directly connected to fluidic input/output reservoirs. Optically, they form a complex system with a high degree of monolithic integration. They are comprised of self-alignment systems for adequate positioning of fiber optics, lenses and mirrors, which have been designed considering only the refractive indices of PDMS (n = 1.41), buffer solution (PBS, phosphate buffer, pH 7.4, 10 mM, n = 1.334) and air (n = 1.00). In the PMIR system, the inherentlydivergent light emerging from the fiber optics is corrected by the biconvex cylindrical microlens, achieving parallel beams at the output of the microlens. The position of the first air mirror has been selected so as to have the previously mentioned TIR conditions of the incident light and hence assure that most of the light is reflected back. In addition, its shape has been selected to be semicylindrical to focus the reflected light inside Lab Chip, 2007, 7, 1560–1566 | 1561
Fig. 2 Schematic representation of the microfluidic system comprising the hollow multiple internal reflection structures: (a) PMIR, (b) RMIR. The microlenses, the reservoirs, the microchannels and the air mirrors (not to scale) are shown, together with the optical path (dark grey line).
the hollow structure and therefore allow multiple reflections without causing a decrease of the SNR. After this first reflection, light is re-injected into the system and reaches the second air mirror, which again reflects the light while focusing it on the vicinity of the collecting fiber optics. Hence, light propagates in the system following a zig-zag path. Therefore, the regions in light grey in Fig. 2a are not relevant for light 1562 | Lab Chip, 2007, 7, 1560–1566
sensing purposes. Although the reservoirs and the input/output channels are obviously crucial, the sensing region of the system can be tailored in order to reduce the mean flow cell volume while keeping the optical path constant. Therefore, the PMIR geometry can be modified so as to remove the dead volumes. Therefore, two different PMIR configurations will be tested: the PMIR-I will be similar to these presented in Fig. 2a. Conversely, and taking into account the previously mentioned discussion regarding the dead volumes, in the PMIR-II the mean flow cell volume will be tailored, making it match with the optical path. It is noteworthy to mention that with both PMIR configurations proposed, each additional reflection leads to an increase of the overall dimensions of the system. Thus, the number of reflections cannot be made indefinitely large. To tackle this situation, the ring configuration (RMIR) has been designed as presented in Fig. 2b. In RMIR structures, a larger number of reflections can be achieved (enlarging therefore the optical path) for a fixed mean flow cell volume. Finally, from the previous discussion, it becomes clear that the appropriate behaviour of the presented MIR system strongly depends on the relative position between the fiber optics and the microlenses. Therefore, it is required to implement a system not only able to position the fiber optics but also to assure its clamping when the optimal condition is reached. Both issues are obtained by defining a microchannel slightly thinner than the diameter of the fiber optics. Hence, when it is introduced into the microchannel, equal forces of opposite sign are applied, assuring its alignment with the microlenses. A constriction defined in the microchannel in the region where the fiber optics should be positioned causes an increase of the applied forces and its clamping. Then, the optimal situation, when the facet end of the fiber optics is at the focus of the biconvex microlens, can be easily obtained. In such a situation, light emerging from the microlens will have parallel beams, i.e. they will not diverge. Obviously this behaviour will only be obtained with perfect spherical microlenses, which are technologically difficult to obtain13 or require non-standard processing methods.14 For ease of fabrication, cylindrical lenses, which only vary the light direction in one axis (that is, instead of having a focal point, they have a focal plane), are more commonly used. Using such lenses and fixing the fiber optics in its focus results in light emerging from the microlenses with parallel beams in the horizontal direction while it broadens in the vertical axis. Three different configurations have been designed taking into account the previously mentioned considerations and their most significant features are presented in Table 1. It is important to mention that these values correspond to the structures defined on the photolithographic mask and on the expected height. Therefore, accurate magnitudes can be given, although some small discrepancies due to errors in the fabrication process may occur. Concerning the structures, the PMIR-I system is the basic multiple internal reflection with two air mirrors. In the PMIR-II structure, the detection region has been tailored so as to reduce the mean flow cell volume while keeping the number of reflections constant. Finally, in the RMIR structure the optical path has been significantly enlarged by using a third air mirror with a significant reduction of the length of the sensing region. It has to be This journal is ß The Royal Society of Chemistry 2007
Table 1 Features of the different MIR systems used
Label/scheme
PMIR-I
PMIR-II
RMIR
Volume /mL Optical path length/mm Fluidic path length/mm Length of the sensing region/mm
3.31 8336 11439 11340
1.43 6357 9573 8800
5.34 14264 7721 6800
taken into account, however, that if only one output fluidic port was implemented in the RMIR structure, air entrapment may occur. Therefore, two symmetric output ports have been designed and the fluidic path length is identical from the input to each output. Additionally, due to the large optical path of the MIR structures, the absorption of the excitation light will be higher, resulting in a minimization of the noise at the output fiber optics. Fabrication The proposed m-TAS systems have been fabricated by casting of PDMS (Sylgard 184 elastomer kit, Dow Corning, Midland, MI, USA) in an SU-8 master (MicroChem, Corp., Newton, MA, USA). The technology is extremely simple and it has been deeply studied and reported, nevertheless, for completeness it is reproduced here. After cleaning and dehydrating a low-cost 700 mm thick soda-lime glass at 200 uC for 1 h, a double adhesion promoter layer is applied so as to obtain better mechanical stability and robustness, together with an increase of the durability of the fabricated masters. Initially, a thin Crlayer is sputtered. Then, the substrate is dehydrated at 150 uC for 1 h prior to the spinning of an SU-8 layer with a thickness of 4 mm, which will act as a seed layer for the subsequent SU-8 layers. Afterwards, the substrates are baked at 95 uC for 10 minutes and exposed to UV light without using a mask. The post-exposure bake (PEB) at 95 uC for 10 min finishes the definition of the double adhesion promoter layer. Then, with a single spin-on process using SU-8 50, a thickness of 250 mm is obtained, which will allow the hassle-free insertion of the optical fibres (which have a diameter of 230 mm). The PEB was followed by developing the structures in propylene glycol methyl ether acetate (PGMEA, MicroChem, Corp., Newton, MA, USA), finishing the definition of the master. The prepolymer was obtained by mixing the curing agent with the elastomer base in a 1 : 10 ratio (v : v). The resulting mixture was subsequently degassed to remove air bubbles, poured over the master and cured for 20 min at 80 uC. Afterwards, the cured PDMS was peeled off from the master and the fluidic ports were opened. Then, both the PDMS and a second sodalime glass substrate were exposed to an oxygen plasma15 in a barrel etcher (Surface Technology Systems, Newport, UK). Immediately afterwards, both surfaces are brought in contact, causing its irreversible sealing. This covalent bonding between This journal is ß The Royal Society of Chemistry 2007
the PDMS and the soda-lime glass ends the fabrication process. Characterization Light emitted from an SLED with a central wavelength of l = 460 nm is coupled into a multimode fiber optic with a diameter of 230 mm, which is located inside the self-alignment system and it is pushed through until it reaches the constriction. The readout comprises an identical fiber optic, which is also inserted into the output self-alignment system and carries the signal to a spectrometer (P.117, STEAG MicroParts, Dortmund, Germany) with a spectral resolution of 12 nm. Measurements have been done at room temperature in a temperature-controlled lab. The first step in the characterization of the MIR structures has been a ray tracing simulation inside the PMIR-I structure, which is presented in Fig. 3a together with a picture of the same system filled with buffer solution and 10 mM of fluorescein (Fig. 3b). This concentration has been selected to obtain a clear vision of the light path inside the hollow structure, including the reflection at the air mirrors. It has to be noted, however, that fluorescein has a broad absorption band in the blue region of the visible spectra and an emission band around 520 nm. This fluroescence light is isotropically emitted in 4p steradians and therefore not all the light emitted from the fluorescein will have the appropriate solid angle so as to match with the TIR conditions at the air mirrors. Hence, only the fraction of light emitted with the appropriate angle will contribute to the signal collected at the output fiber optics. As can be observed, there is an agreement between the ray tracing evaluation and the experimental results: in the ray tracing simulations, two focuses were predicted inside the hollow structure region. They have been experimentally observed (as seen in Fig. 3b), confirming that such structures allow for the defining of multiple internal analysis systems that tackle the problem of beam broadening during propagation. Also, from Fig. 3b, it can be observed that rays emerging from the input biconvex lens have parallel beams. This point validates the use of constrictions on the self-alignment systems for fast and accurate fiber optics positioning. Finally, as it was predicted in the simulations, it can be observed that there are some regions in the system that do not play any role in terms of sensitivity enhancement, since the light from the LED does not propagate in these regions. Therefore, these regions only Lab Chip, 2007, 7, 1560–1566 | 1563
Fig. 4 (a) Ray-tracing simulation and (b) photograph of a PMIR-II structure filled with buffer solution + 10 mM of fluorescein. As compared to PMIR-I, PMIR-II structures does not have dead volumes. Fig. 3 (a) Ray-tracing simulation and (b) photograph of a PMIR-I structure filled with buffer solution + 10 mM of fluorescein, in which the zig-zag path of the light from the LED inside the sensing region due to the total internal reflection at the air mirrors can be seen, together with the focal planes predicted in the ray tracing simulations.
cause an increase of the mean flow cell volume without increasing the sensitivity. The tackling of this drawback can be done by modifying the geometry, as proposed in the PMIR-II configuration. Ray-tracing simulations (Fig. 4a) and experimental results measured (Fig. 4b) confirm that a large reduction of the hollow structure does not cause variations of the optical path. Numerical and experimental images taken using a concentration of 10 mM of fluorescein also are in good agreement with this PMIR-II configuration. As opposite to the PMIR-I configuration, the PMIR-II does not have dead volumes (noticeable because there are no dark regions on the sensing zone). Finally, the RMIR configuration was also analyzed with ray-tracing (Fig. 5a) and compared with the images of the system filled with the same fluorescein concentration (Fig. 5b). Again, a good agreement is obtained, confirming that it is possible to tailor the light propagation inside the sensing region by using an increasingly large number of air mirrors. To determine the LOD of the proposed system, the following experimental procedure has been carried out: firstly, the system has been filled with buffer solution and the readout obtained was considered as the reference. Then, dilutions with progressively higher concentrations of fluorescein ranging from 20 to 5000 nM (Sigma–Aldrich Chemie GmbH, 1564 | Lab Chip, 2007, 7, 1560–1566
Steinheim, Germany) have been progressively injected. For each concentration the average of 10 consecutive scans was taken. After having measured the highest available fluorescein concentration, the buffer solution is injected again to check possible drifts of the reference signal. In all the measured sets, the reference signal had a constant value within the experimental error. Then, the absorbance as a function of the analyte concentration has been plotted and a linear fit has been done. In accordance to the 3 sigma IUPAC definition, from the values of this linear fit, the LOD can be obtained. It is noteworthy to mention that the LOD is not the lowest detectable analyte concentration, but also depends both on the sensitivity and the accuracy of the linear fit. The first result is a drastic reduction of the required integration time. The experimental results presented in ref. 9 required an integration time of 2.5 s to obtain a SNR between 12 and 14 dB. This excessively long integration time is perhaps the most severe limitation of these SIR systems. Conversely, although the MIR systems have a longer optical path, the air mirror allows integration times of 320 ms while having SNR ranging between 13 and 17 dB (depending on the particular configuration measured). The absorbance as a function of the fluorescein concentration has been studied for the three different MIR systems and their linear fit, their values of R2 and their LOD are presented in Fig. 6 and summarized in Table 2. As it can be observed, there is a good agreement with the Beer–Lambert law,16 which predicts a linear behaviour between the absorbance and the analyte concentration. These results confirm that in the three This journal is ß The Royal Society of Chemistry 2007
Fig. 5 (a) Ray-tracing simulation and (b) photograph of a RMIR structure filled with buffer solution +10 mM of fluorescein.
Fig. 6 Experimental results of the absorbance vs. fluorescein concentration for the PMIR and RMIR systems studied, together with their respective linear fit.
MIR systems a very low LOD is obtained, with values ranging between 41 and 110 nM. The two PMIR configurations have quite similar LOD, although the PMIR-II has a much lower mean flow cell volume. Again, this point confirms that the
increase of the volume on the PMIR-I configuration does not provide additional advantages to the system. The modification of the shape of the detection region, making its shape similar to the optical path, allows a drastic reduction of the mean flow cell volume while keeping the LOD nearly constant. From the slopes of the linear fits, the molar absorptivity of the fluorescein can be determined to be 18 107 ¡ 200 L mol21 cm21 and 21 393 ¡ 500 L mol21 cm21 for the PMIR-I and PMIR-II, while this magnitude decreases to 15 300 ¡ 400 L mol21 cm21 for the RMIR. The significant variation of the molar absorptivity for the RMIR structures can be associated to a variation of the optical path due to a suboptimal positioning of the fiber optics. The RMIR configuration shows the highest sensitivity, which is in agreement with the largest optical path. This configuration shows a LOD nearly 45 times smaller than the previously presented SIR systems14 and with values comparable to these previously reported using avalanche photodiodes (APD),17 but using a much simpler technology that only requires one photolithographic step. Thus, it can be concluded that the combination of soft lithography and air mirrors allows for the defining of extremely sensitive absorbance-based systems with very low LOD.
Table 2 Linear fit, R2 and LOD of the fluorescein measurements performed with the different multiple internal reflection systems. A and C stand for the absorbance and the fluorescein concentration, respectively Linear Fit PMIR-I PMIR-II RMIR
24
25
27
A = (0.0039 ¡ 5 6 10 ) + (1.60 6 10 ¡ 2 6 10 )C A = (0.0371 ¡ 5 6 1024) + (1.36 6 1025 ¡ 3 6 1027)C A = (20.0014 ¡ 3 6 1024) + (2.18 6 1025 ¡ 5 6 1027)C
This journal is ß The Royal Society of Chemistry 2007
R2
LOD/nM
0.998 0.998 0.998
93 ¡ 1 110 ¡ 2 41 ¡ 1
Lab Chip, 2007, 7, 1560–1566 | 1565
Conclusions
Acknowledgements
Multiple internal reflection systems have been presented as candidates for highly sensitive lab-on-a-chip applications. With an appropriate definition of the air mirrors at both sides of the sensing volume, it has been demonstrated that light can match the conditions of total internal reflection and be reflected back to the sensing volume, following a zigzag path. Therefore, an enhancement of the sensitivity and an LOD lowering can be achieved. Moreover, with a proper shape of the air mirrors, light divergence can be corrected, enhancing the SNR and decreasing the losses of the system. Two different families have been studied, namely propagating and ring multiple internal reflection systems (PMIR and RMIR, respectively). When these structures are filled with 10 mM of fluorescein, the light propagating in the structure can be observed, and its behaviour is in agreement with the ray tracing models predictions, which confirms the validity of the proposed lab-on-a-chip systems. The PMIR-I structure uses two air mirrors and has shown an LOD of 93 nM for fluorescein diluted in phosphate buffer. Nevertheless, it has also been shown that there are some regions of this system where the light from the LED (l = 460 nm) does not interact with the mean flow cell volume. A large reduction of the required volume has been achieved with the PMIR-II structure, by fitting its shape to the optical path and keeping constant the number of air mirrors. A similar LOD has been achieved with this latter structure while achieving a reduction of 50% in the mean flow cell volume. Finally, a third air mirror has been used in the ring structures, since its compact shape allows a much larger optical path without a dramatic increase of the size of the system. With a sensing region 30% smaller than the PMIR-I configuration, a LOD of 41 nM has been achieved. These results confirm the validity of the proposed MIR configurations for highly sensitive lab-on-a-chip applications.
The authors thank the German Research Foundation (DFG) for support of this work in the framework of the Collaborative Research Center SFB 516.
1566 | Lab Chip, 2007, 7, 1560–1566
References 1 S. Balslev, A. M. Jorgensen, B. Bilenberg, K. B. Mogensen, D. Snakenborg, O. Geschke, J. P. Kutter and A. Kristensen, Lab Chip, 2006, 6, 213–217. 2 J. Fritz, M. K. Baller, H. P. Lang, H. Rothuizen, P. Vettiger, E. Meyer, H.-J. Gu¨ntherodt, Ch. Gerber and J. K. Gimzewski, Science, 2000, 288, 316–318. 3 B. Sepu´lveda, J. Sa´nchez del Rio, M. Moreno, F. J. Blanco, K. Mayora, C. Domı´nguez and L. M. Lechuga, J. Opt. A: Pure Appl. Opt., 2006, 8, 561–566. 4 S. Camou, H. Fujita and T. Fujii, Lab Chip, 2003, 3, 40–45. 5 C. H. Lin and G. B. Lee, J. Micromech. Microeng., 2003, 13, 447–453. 6 R. W. Applegate, J. Squier, T. Vestad, J. Oakey, D. W. M. Marr, P. Bado, M. A. Dugan and A. A. Said, Lab Chip, 2006, 6(422), 426. 7 A. Llobera, R. Wilke and S. Bu¨ttgenbach, Lab Chip, 2004, 4, 24–27. 8 O. Hofmann, X. Wang, A. Cornwell, S. Beecher, A. Raja, D. C. Bradley, A. J. deMello and J. C. deMello, Lab Chip, 2006, 6, 981–987. 9 A. Llobera, R. Wilke and S. Bu¨ttgenbach, Lab Chip, 2005, 5, 506–511. 10 N. Rochat, M. Olivier, A. Chabli, F. Conne, G. Lefeuvre and C. Boll-Durdet, Appl. Phys. Lett., 2000, 77(14), 2249–2251. 11 H. Salimi-Moosavi, Y. Jiang, L. Lester, G. McKinnon and D. Jed Harrison, Electrophoresis, 2000, 21, 1291–1299. 12 N. Chronis and L. P. Lee, Lab Chip, 2004, 4, 125–130. 13 R. Yang and W. Wang, Opt. Eng., 2004, 43(12), 3096–3103. 14 A. Llobera, R. Wilke, D. W. Johnson and S. Bu¨ttgenbach, IEEE Photon. Technol. Lett., 2005, 17(12), 2628–2630. 15 B. H. Jo, L. M. Van Lerberghe, K. M. Motsegood and D. J. Beebe, J. Microelectromech. Syst., 2000, 9, 76–81. 16 Sensors, a comprehensive survey. Volume 6: Optical Sensors, ed. W. Go¨pel, J. Hesse, J. N. Zemel, E. Wagner, R. Da¨ndliker and K. Spenner, Wiley-VCH, Weinheim, Germany, 1991. 17 M. L. Chabinyc, D. T. Chiu, J. C. McDonald, A. D. Stroock, J. F. Christian, A. M. Karger and G. M. Whitesides, Anal. Chem., 2001, 73, 4491–4498.
This journal is ß The Royal Society of Chemistry 2007
www.rsc.org/loc | Lab on a Chip
Multiple internal reflection poly(dimethylsiloxane) systems for optical sensing A. Llobera,*ab S. Demming,a R. Wilkea and S. Bu¨ttgenbacha Received 23rd March 2007, Accepted 24th July 2007 First published as an Advance Article on the web 21st August 2007 DOI: 10.1039/b704454b Compact poly(dimethylsiloxane)-based (PDMS) multiple internal reflection systems which comprise self-alignment systems, lenses, microfluidic channels and mirrors have been developed for highly sensitive absorbance measurements. With the proper definition of air mirrors at both sides of the sensing region, the optical path of the light from the LED has been meaningfully lengthened without a dramatic increase of the mean flow cell volume. By recursive positioning of such air mirrors, propagating multiple internal reflection (PMIR) systems have been designed, simulated and characterized. Experimental results confirm the ray-tracing predictions and allow the determining that there are some regions of the mean flow cell volume that do not contribute to the increase of the sensitivity. The tailoring of the sensing region, following the optical path, results in a similar limit of detection (110 nM) for fluorescein diluted in phosphate buffer. Finally, a ring configuration, labelled RMIR, has also been developed. With the addition of a third air mirror, the LOD can be decreased to 41 nM with the additional advantage of a substantial decrease of the length of the sensing region. These results confirm the validity of the proposed systems for high sensitivity measurements.
Introduction Currently, there is a huge demand for lab on chip systems with a low limit of detection (LOD) of (bio)chemical substances in a large variety of fields, such as molecular diagnosis, proteomics or monitoring of toxins. Ideally, such systems should have a high sensitivity, repeatability, reversibility, quick response, small size, low LOD, portability and low cost. To fulfil such strict requirements, several types of transduction mechanisms primarily focused on obtaining increasingly sensitive systems have previously been reported. Among these, optical,1 and mechanical2 transducers are the most commonly reported. Transducers based on the detection of a given optical magnitude (amplitude, phase or wavelength, mainly) have so far presented the highest sensitivity. As an example, optical transducers based on phase modulation with a Mach–Zehnder Interferometer (MZI) configuration have achieved a LOD of DNA hybridization of 300 pM without requiring labelling3 and with a high degree of integration. Nevertheless, the use of nanometre-scale optical structures (such as single mode silicon nitride waveguides) demands the use of expensive equipment that causes a dramatic increase of the cost per system, making it difficult to meet the low cost requirement. The tackling of this drawback is associated with the use of polymers, and especially poly(dimethylsiloxane) (PDMS). Taking advantage of its optical and structural properties, it is possible to monolithically define microfluidics and optical a
Institut fu¨r Mikrotechnik, Technische Universita¨t Braunschweig, Alte Salzdahlumer Straße 203, 38124 Braunschweig, Germany. E-mail: [email protected]; Fax: +49 531/391-9751; Tel: +49 531/ 391-9752 b Centre Nacional de Microelectro`nica. Campus UAB, 08193 Bellaterra, Spain
1560 | Lab Chip, 2007, 7, 1560–1566
structures. Such excellent properties have favoured their useage in laser-induced-fluorescence systems,4 cytometers5 as well as sorters and mixers.6 Moreover, with an accurate design7 or by doping PDMS with the appropriate substance,8 multiplexing or filtering capabilities can be transferred to the system. In addition, its biocompatibility and its low cost are two more advantages that speak for this material. Considering the authors’ previous work using PDMS-based lab on chip systems with multiplexing properties, the hollow prisms have shown their capability to distinguish between excited and emitted wavelength7 in fluorescence measurements, together with their ability to colorimetrically measure changes in the absorption spectra either due to variations of the concentration of a given substance or due to pH variations.9 Optically, the hollow prisms are examples of single internal reflection structures (SIR), since its working principle only requires one reflection. In opposition, multiple internal reflection (MIR) systems have already been presented for infrared spectroscopy for quantitative study of organic contamination of silicon wafers,10 absorption measurement11 and the Brownian motion of fluorescent microspheres.12 Generally, MIR systems are suitable to provide higher sensitivity than SIR systems. Nevertheless, it has several issues that need to be addressed and solved prior to its implementation in lab on chip systems. Firstly, with each additional reflection there is a decrease of intensity, which results in a decrease of the signal-to-noise ratio (SNR). As an example, in the Abbe prism presented in ref. 9, the reflection coefficients of the PDMS-buffer interface were 0.029 for TE-polarization and 0.062 for its TM counterpart. Hence, long integration times were required to obtain a clear signal even in systems with a single reflection, resulting in a low throughput of the hollow prisms and making the implementation of MIR structures with This journal is ß The Royal Society of Chemistry 2007
such low reflection coefficients unaffordable. Secondly, light inherently diverges. Then, as the optical path increases, there is an enlargement of the diameter beam. Therefore, an increase of the optical path beyond a given value generally causes a degradation of the properties of the proposed system, since a fraction of light is not collected by the photodetectors or fiber optics, which have finite dimensions, causing a dramatic decrease of the SNR. In this paper, we have addressed and solved the previously mentioned drawbacks to develop two highly-sensitive configurations of MIR structures, labelled propagating multiple internal reflection (PMIR) and ring multiple internal reflection (RMIR). Both configurations can be obtained in a single technological step and significantly decrease the LOD as compared to SIR structures.9
Structures Design As it was previously mentioned, the improvement of the reflectivity is a primary issue to enhance the performance of any reflection-based photonic lab-on-a-chip system. One possibility for improvement would be to define a mirror that reflects back the outcoming light from the hollow prism and thus achieving its collection by the output fiber optics. Although silicon or metal-coated mirrors can be easily defined, this would lead to a more complex fabrication process and an increase of the cost of the device. The ideal solution would be to obtain a mirror without any additional process steps. This objective can be achieved with the materials used (air, PDMS and buffer solution), if what we have called air mirrors are implemented, as shown in Fig. 1. These structures can be structurally envisaged as air entrapments with a concrete shape and position so as to modify the light path. In the same figure, the ray-tracing simulations of such structures are presented. In the SIR structures presented in ref. 7 and 9, which did not have an air mirror, light was propagating from a media with low refractive index (RI) (n = 1.334) to PDMS (with a higher RI, n = 1.41). Therefore, most of the light was coupled to the PDMS and did not reach the collecting fiber optics. Conversely, as it can be observed in Fig. 1, when the air mirror is included, the light coupled to the PDMS reaches a region with lower RI (air, n = 1.00). If this light has the appropriate incidence angle, total internal reflection (TIR) at the PDMS–air layer is obtained, resulting in a complete reflection of the light towards the PDMS. Considering the refractive indices of the media involved (PDMS and air) and using the Snell’s law, the critical propagation angle is hcPDMS-air = 45.17u. All the propagation angles h . hcPDMS-air (marked in Fig. 1) cause the light to be under TIR conditions and reflected back to the PDMS, reaching the hollow structure. Considering again the RI of the materials involved (PDMS and phosphate buffer (PBS) in this case), in this boundary region the critical angle is hcPDMS-PBS = 70.60u (the angles where the TIR regime happens are also marked in the figure). As can be observed in Fig. 1, the propagation angles match with the TIR regime at the PDMS–air region, but do not at the PDMS–PBS interface. This is the optimal situation, since then light does not remain confined at the PDMS and is This journal is ß The Royal Society of Chemistry 2007
Fig. 1 Ray tracing simulation of the propagating light at the air mirror vicinity, taking into account the refractive index values of air (1.00) poly(dimethylsiloxane) (PDMS, 1.41) and buffer solution (PBS, 1.334). The angles where the condition of total internal reflection (TIR) holds at both PDMS–air and PDMS–PBS interfaces are also shown.
coupled back to the MIR system. In addition, the shape of the air mirror can modify the behaviour of the optical beam; that is, from the initial parallel beams, an adequate curvature of the air mirror make the light converge inside the hollow structure, as is also shown in the ray-tracing diagram of Fig. 1. Hence, with this simple structure, the two most significant issues of the MIR configuration (reflection and beam broadening) can be addressed and solved. Therefore, the optical path can now be enlarged; enhancing the readout signal obtained and at the same time reducing the integration time, as will be presented in the Characterization section. After designing the air mirrors, the configurations of the PMIR and RMIR systems are presented in Fig. 2a and 2b, respectively. From a fluidic point of view, the proposed systems consist of a hollow structure directly connected to fluidic input/output reservoirs. Optically, they form a complex system with a high degree of monolithic integration. They are comprised of self-alignment systems for adequate positioning of fiber optics, lenses and mirrors, which have been designed considering only the refractive indices of PDMS (n = 1.41), buffer solution (PBS, phosphate buffer, pH 7.4, 10 mM, n = 1.334) and air (n = 1.00). In the PMIR system, the inherentlydivergent light emerging from the fiber optics is corrected by the biconvex cylindrical microlens, achieving parallel beams at the output of the microlens. The position of the first air mirror has been selected so as to have the previously mentioned TIR conditions of the incident light and hence assure that most of the light is reflected back. In addition, its shape has been selected to be semicylindrical to focus the reflected light inside Lab Chip, 2007, 7, 1560–1566 | 1561
Fig. 2 Schematic representation of the microfluidic system comprising the hollow multiple internal reflection structures: (a) PMIR, (b) RMIR. The microlenses, the reservoirs, the microchannels and the air mirrors (not to scale) are shown, together with the optical path (dark grey line).
the hollow structure and therefore allow multiple reflections without causing a decrease of the SNR. After this first reflection, light is re-injected into the system and reaches the second air mirror, which again reflects the light while focusing it on the vicinity of the collecting fiber optics. Hence, light propagates in the system following a zig-zag path. Therefore, the regions in light grey in Fig. 2a are not relevant for light 1562 | Lab Chip, 2007, 7, 1560–1566
sensing purposes. Although the reservoirs and the input/output channels are obviously crucial, the sensing region of the system can be tailored in order to reduce the mean flow cell volume while keeping the optical path constant. Therefore, the PMIR geometry can be modified so as to remove the dead volumes. Therefore, two different PMIR configurations will be tested: the PMIR-I will be similar to these presented in Fig. 2a. Conversely, and taking into account the previously mentioned discussion regarding the dead volumes, in the PMIR-II the mean flow cell volume will be tailored, making it match with the optical path. It is noteworthy to mention that with both PMIR configurations proposed, each additional reflection leads to an increase of the overall dimensions of the system. Thus, the number of reflections cannot be made indefinitely large. To tackle this situation, the ring configuration (RMIR) has been designed as presented in Fig. 2b. In RMIR structures, a larger number of reflections can be achieved (enlarging therefore the optical path) for a fixed mean flow cell volume. Finally, from the previous discussion, it becomes clear that the appropriate behaviour of the presented MIR system strongly depends on the relative position between the fiber optics and the microlenses. Therefore, it is required to implement a system not only able to position the fiber optics but also to assure its clamping when the optimal condition is reached. Both issues are obtained by defining a microchannel slightly thinner than the diameter of the fiber optics. Hence, when it is introduced into the microchannel, equal forces of opposite sign are applied, assuring its alignment with the microlenses. A constriction defined in the microchannel in the region where the fiber optics should be positioned causes an increase of the applied forces and its clamping. Then, the optimal situation, when the facet end of the fiber optics is at the focus of the biconvex microlens, can be easily obtained. In such a situation, light emerging from the microlens will have parallel beams, i.e. they will not diverge. Obviously this behaviour will only be obtained with perfect spherical microlenses, which are technologically difficult to obtain13 or require non-standard processing methods.14 For ease of fabrication, cylindrical lenses, which only vary the light direction in one axis (that is, instead of having a focal point, they have a focal plane), are more commonly used. Using such lenses and fixing the fiber optics in its focus results in light emerging from the microlenses with parallel beams in the horizontal direction while it broadens in the vertical axis. Three different configurations have been designed taking into account the previously mentioned considerations and their most significant features are presented in Table 1. It is important to mention that these values correspond to the structures defined on the photolithographic mask and on the expected height. Therefore, accurate magnitudes can be given, although some small discrepancies due to errors in the fabrication process may occur. Concerning the structures, the PMIR-I system is the basic multiple internal reflection with two air mirrors. In the PMIR-II structure, the detection region has been tailored so as to reduce the mean flow cell volume while keeping the number of reflections constant. Finally, in the RMIR structure the optical path has been significantly enlarged by using a third air mirror with a significant reduction of the length of the sensing region. It has to be This journal is ß The Royal Society of Chemistry 2007
Table 1 Features of the different MIR systems used
Label/scheme
PMIR-I
PMIR-II
RMIR
Volume /mL Optical path length/mm Fluidic path length/mm Length of the sensing region/mm
3.31 8336 11439 11340
1.43 6357 9573 8800
5.34 14264 7721 6800
taken into account, however, that if only one output fluidic port was implemented in the RMIR structure, air entrapment may occur. Therefore, two symmetric output ports have been designed and the fluidic path length is identical from the input to each output. Additionally, due to the large optical path of the MIR structures, the absorption of the excitation light will be higher, resulting in a minimization of the noise at the output fiber optics. Fabrication The proposed m-TAS systems have been fabricated by casting of PDMS (Sylgard 184 elastomer kit, Dow Corning, Midland, MI, USA) in an SU-8 master (MicroChem, Corp., Newton, MA, USA). The technology is extremely simple and it has been deeply studied and reported, nevertheless, for completeness it is reproduced here. After cleaning and dehydrating a low-cost 700 mm thick soda-lime glass at 200 uC for 1 h, a double adhesion promoter layer is applied so as to obtain better mechanical stability and robustness, together with an increase of the durability of the fabricated masters. Initially, a thin Crlayer is sputtered. Then, the substrate is dehydrated at 150 uC for 1 h prior to the spinning of an SU-8 layer with a thickness of 4 mm, which will act as a seed layer for the subsequent SU-8 layers. Afterwards, the substrates are baked at 95 uC for 10 minutes and exposed to UV light without using a mask. The post-exposure bake (PEB) at 95 uC for 10 min finishes the definition of the double adhesion promoter layer. Then, with a single spin-on process using SU-8 50, a thickness of 250 mm is obtained, which will allow the hassle-free insertion of the optical fibres (which have a diameter of 230 mm). The PEB was followed by developing the structures in propylene glycol methyl ether acetate (PGMEA, MicroChem, Corp., Newton, MA, USA), finishing the definition of the master. The prepolymer was obtained by mixing the curing agent with the elastomer base in a 1 : 10 ratio (v : v). The resulting mixture was subsequently degassed to remove air bubbles, poured over the master and cured for 20 min at 80 uC. Afterwards, the cured PDMS was peeled off from the master and the fluidic ports were opened. Then, both the PDMS and a second sodalime glass substrate were exposed to an oxygen plasma15 in a barrel etcher (Surface Technology Systems, Newport, UK). Immediately afterwards, both surfaces are brought in contact, causing its irreversible sealing. This covalent bonding between This journal is ß The Royal Society of Chemistry 2007
the PDMS and the soda-lime glass ends the fabrication process. Characterization Light emitted from an SLED with a central wavelength of l = 460 nm is coupled into a multimode fiber optic with a diameter of 230 mm, which is located inside the self-alignment system and it is pushed through until it reaches the constriction. The readout comprises an identical fiber optic, which is also inserted into the output self-alignment system and carries the signal to a spectrometer (P.117, STEAG MicroParts, Dortmund, Germany) with a spectral resolution of 12 nm. Measurements have been done at room temperature in a temperature-controlled lab. The first step in the characterization of the MIR structures has been a ray tracing simulation inside the PMIR-I structure, which is presented in Fig. 3a together with a picture of the same system filled with buffer solution and 10 mM of fluorescein (Fig. 3b). This concentration has been selected to obtain a clear vision of the light path inside the hollow structure, including the reflection at the air mirrors. It has to be noted, however, that fluorescein has a broad absorption band in the blue region of the visible spectra and an emission band around 520 nm. This fluroescence light is isotropically emitted in 4p steradians and therefore not all the light emitted from the fluorescein will have the appropriate solid angle so as to match with the TIR conditions at the air mirrors. Hence, only the fraction of light emitted with the appropriate angle will contribute to the signal collected at the output fiber optics. As can be observed, there is an agreement between the ray tracing evaluation and the experimental results: in the ray tracing simulations, two focuses were predicted inside the hollow structure region. They have been experimentally observed (as seen in Fig. 3b), confirming that such structures allow for the defining of multiple internal analysis systems that tackle the problem of beam broadening during propagation. Also, from Fig. 3b, it can be observed that rays emerging from the input biconvex lens have parallel beams. This point validates the use of constrictions on the self-alignment systems for fast and accurate fiber optics positioning. Finally, as it was predicted in the simulations, it can be observed that there are some regions in the system that do not play any role in terms of sensitivity enhancement, since the light from the LED does not propagate in these regions. Therefore, these regions only Lab Chip, 2007, 7, 1560–1566 | 1563
Fig. 4 (a) Ray-tracing simulation and (b) photograph of a PMIR-II structure filled with buffer solution + 10 mM of fluorescein. As compared to PMIR-I, PMIR-II structures does not have dead volumes. Fig. 3 (a) Ray-tracing simulation and (b) photograph of a PMIR-I structure filled with buffer solution + 10 mM of fluorescein, in which the zig-zag path of the light from the LED inside the sensing region due to the total internal reflection at the air mirrors can be seen, together with the focal planes predicted in the ray tracing simulations.
cause an increase of the mean flow cell volume without increasing the sensitivity. The tackling of this drawback can be done by modifying the geometry, as proposed in the PMIR-II configuration. Ray-tracing simulations (Fig. 4a) and experimental results measured (Fig. 4b) confirm that a large reduction of the hollow structure does not cause variations of the optical path. Numerical and experimental images taken using a concentration of 10 mM of fluorescein also are in good agreement with this PMIR-II configuration. As opposite to the PMIR-I configuration, the PMIR-II does not have dead volumes (noticeable because there are no dark regions on the sensing zone). Finally, the RMIR configuration was also analyzed with ray-tracing (Fig. 5a) and compared with the images of the system filled with the same fluorescein concentration (Fig. 5b). Again, a good agreement is obtained, confirming that it is possible to tailor the light propagation inside the sensing region by using an increasingly large number of air mirrors. To determine the LOD of the proposed system, the following experimental procedure has been carried out: firstly, the system has been filled with buffer solution and the readout obtained was considered as the reference. Then, dilutions with progressively higher concentrations of fluorescein ranging from 20 to 5000 nM (Sigma–Aldrich Chemie GmbH, 1564 | Lab Chip, 2007, 7, 1560–1566
Steinheim, Germany) have been progressively injected. For each concentration the average of 10 consecutive scans was taken. After having measured the highest available fluorescein concentration, the buffer solution is injected again to check possible drifts of the reference signal. In all the measured sets, the reference signal had a constant value within the experimental error. Then, the absorbance as a function of the analyte concentration has been plotted and a linear fit has been done. In accordance to the 3 sigma IUPAC definition, from the values of this linear fit, the LOD can be obtained. It is noteworthy to mention that the LOD is not the lowest detectable analyte concentration, but also depends both on the sensitivity and the accuracy of the linear fit. The first result is a drastic reduction of the required integration time. The experimental results presented in ref. 9 required an integration time of 2.5 s to obtain a SNR between 12 and 14 dB. This excessively long integration time is perhaps the most severe limitation of these SIR systems. Conversely, although the MIR systems have a longer optical path, the air mirror allows integration times of 320 ms while having SNR ranging between 13 and 17 dB (depending on the particular configuration measured). The absorbance as a function of the fluorescein concentration has been studied for the three different MIR systems and their linear fit, their values of R2 and their LOD are presented in Fig. 6 and summarized in Table 2. As it can be observed, there is a good agreement with the Beer–Lambert law,16 which predicts a linear behaviour between the absorbance and the analyte concentration. These results confirm that in the three This journal is ß The Royal Society of Chemistry 2007
Fig. 5 (a) Ray-tracing simulation and (b) photograph of a RMIR structure filled with buffer solution +10 mM of fluorescein.
Fig. 6 Experimental results of the absorbance vs. fluorescein concentration for the PMIR and RMIR systems studied, together with their respective linear fit.
MIR systems a very low LOD is obtained, with values ranging between 41 and 110 nM. The two PMIR configurations have quite similar LOD, although the PMIR-II has a much lower mean flow cell volume. Again, this point confirms that the
increase of the volume on the PMIR-I configuration does not provide additional advantages to the system. The modification of the shape of the detection region, making its shape similar to the optical path, allows a drastic reduction of the mean flow cell volume while keeping the LOD nearly constant. From the slopes of the linear fits, the molar absorptivity of the fluorescein can be determined to be 18 107 ¡ 200 L mol21 cm21 and 21 393 ¡ 500 L mol21 cm21 for the PMIR-I and PMIR-II, while this magnitude decreases to 15 300 ¡ 400 L mol21 cm21 for the RMIR. The significant variation of the molar absorptivity for the RMIR structures can be associated to a variation of the optical path due to a suboptimal positioning of the fiber optics. The RMIR configuration shows the highest sensitivity, which is in agreement with the largest optical path. This configuration shows a LOD nearly 45 times smaller than the previously presented SIR systems14 and with values comparable to these previously reported using avalanche photodiodes (APD),17 but using a much simpler technology that only requires one photolithographic step. Thus, it can be concluded that the combination of soft lithography and air mirrors allows for the defining of extremely sensitive absorbance-based systems with very low LOD.
Table 2 Linear fit, R2 and LOD of the fluorescein measurements performed with the different multiple internal reflection systems. A and C stand for the absorbance and the fluorescein concentration, respectively Linear Fit PMIR-I PMIR-II RMIR
24
25
27
A = (0.0039 ¡ 5 6 10 ) + (1.60 6 10 ¡ 2 6 10 )C A = (0.0371 ¡ 5 6 1024) + (1.36 6 1025 ¡ 3 6 1027)C A = (20.0014 ¡ 3 6 1024) + (2.18 6 1025 ¡ 5 6 1027)C
This journal is ß The Royal Society of Chemistry 2007
R2
LOD/nM
0.998 0.998 0.998
93 ¡ 1 110 ¡ 2 41 ¡ 1
Lab Chip, 2007, 7, 1560–1566 | 1565
Conclusions
Acknowledgements
Multiple internal reflection systems have been presented as candidates for highly sensitive lab-on-a-chip applications. With an appropriate definition of the air mirrors at both sides of the sensing volume, it has been demonstrated that light can match the conditions of total internal reflection and be reflected back to the sensing volume, following a zigzag path. Therefore, an enhancement of the sensitivity and an LOD lowering can be achieved. Moreover, with a proper shape of the air mirrors, light divergence can be corrected, enhancing the SNR and decreasing the losses of the system. Two different families have been studied, namely propagating and ring multiple internal reflection systems (PMIR and RMIR, respectively). When these structures are filled with 10 mM of fluorescein, the light propagating in the structure can be observed, and its behaviour is in agreement with the ray tracing models predictions, which confirms the validity of the proposed lab-on-a-chip systems. The PMIR-I structure uses two air mirrors and has shown an LOD of 93 nM for fluorescein diluted in phosphate buffer. Nevertheless, it has also been shown that there are some regions of this system where the light from the LED (l = 460 nm) does not interact with the mean flow cell volume. A large reduction of the required volume has been achieved with the PMIR-II structure, by fitting its shape to the optical path and keeping constant the number of air mirrors. A similar LOD has been achieved with this latter structure while achieving a reduction of 50% in the mean flow cell volume. Finally, a third air mirror has been used in the ring structures, since its compact shape allows a much larger optical path without a dramatic increase of the size of the system. With a sensing region 30% smaller than the PMIR-I configuration, a LOD of 41 nM has been achieved. These results confirm the validity of the proposed MIR configurations for highly sensitive lab-on-a-chip applications.
The authors thank the German Research Foundation (DFG) for support of this work in the framework of the Collaborative Research Center SFB 516.
1566 | Lab Chip, 2007, 7, 1560–1566
References 1 S. Balslev, A. M. Jorgensen, B. Bilenberg, K. B. Mogensen, D. Snakenborg, O. Geschke, J. P. Kutter and A. Kristensen, Lab Chip, 2006, 6, 213–217. 2 J. Fritz, M. K. Baller, H. P. Lang, H. Rothuizen, P. Vettiger, E. Meyer, H.-J. Gu¨ntherodt, Ch. Gerber and J. K. Gimzewski, Science, 2000, 288, 316–318. 3 B. Sepu´lveda, J. Sa´nchez del Rio, M. Moreno, F. J. Blanco, K. Mayora, C. Domı´nguez and L. M. Lechuga, J. Opt. A: Pure Appl. Opt., 2006, 8, 561–566. 4 S. Camou, H. Fujita and T. Fujii, Lab Chip, 2003, 3, 40–45. 5 C. H. Lin and G. B. Lee, J. Micromech. Microeng., 2003, 13, 447–453. 6 R. W. Applegate, J. Squier, T. Vestad, J. Oakey, D. W. M. Marr, P. Bado, M. A. Dugan and A. A. Said, Lab Chip, 2006, 6(422), 426. 7 A. Llobera, R. Wilke and S. Bu¨ttgenbach, Lab Chip, 2004, 4, 24–27. 8 O. Hofmann, X. Wang, A. Cornwell, S. Beecher, A. Raja, D. C. Bradley, A. J. deMello and J. C. deMello, Lab Chip, 2006, 6, 981–987. 9 A. Llobera, R. Wilke and S. Bu¨ttgenbach, Lab Chip, 2005, 5, 506–511. 10 N. Rochat, M. Olivier, A. Chabli, F. Conne, G. Lefeuvre and C. Boll-Durdet, Appl. Phys. Lett., 2000, 77(14), 2249–2251. 11 H. Salimi-Moosavi, Y. Jiang, L. Lester, G. McKinnon and D. Jed Harrison, Electrophoresis, 2000, 21, 1291–1299. 12 N. Chronis and L. P. Lee, Lab Chip, 2004, 4, 125–130. 13 R. Yang and W. Wang, Opt. Eng., 2004, 43(12), 3096–3103. 14 A. Llobera, R. Wilke, D. W. Johnson and S. Bu¨ttgenbach, IEEE Photon. Technol. Lett., 2005, 17(12), 2628–2630. 15 B. H. Jo, L. M. Van Lerberghe, K. M. Motsegood and D. J. Beebe, J. Microelectromech. Syst., 2000, 9, 76–81. 16 Sensors, a comprehensive survey. Volume 6: Optical Sensors, ed. W. Go¨pel, J. Hesse, J. N. Zemel, E. Wagner, R. Da¨ndliker and K. Spenner, Wiley-VCH, Weinheim, Germany, 1991. 17 M. L. Chabinyc, D. T. Chiu, J. C. McDonald, A. D. Stroock, J. F. Christian, A. M. Karger and G. M. Whitesides, Anal. Chem., 2001, 73, 4491–4498.
This journal is ß The Royal Society of Chemistry 2007
