Fluorescence excitation and observation is a common technique of investigation in the field of biochemistry and cell biology. In our approach, we demonstrate the excitation of fluorescence on an integrated optofluidic system.1 Microfluidic systems can be used as analysing platforms requiring only a minimum amount of biological or chemical analyte. By the use of microoptical waveguides, specific spots of analyte or cells can be addressed optically,2, 3, 4 which makes the combination of waveguides and fluorescence microscopy a useful tool for cell biology.5, 6, 7 The implementation of fluorescence detection in such a microfluidic and microoptical system may open up the possibility for, e.g., cell based sensors, where it is often advantageous to investigate a very limited amount of a certain cell culture during the study of its behavior and characteristics.
The integration of optical waveguides and channels in one system has been shown for a combination of different materials, for instance, glasses and poly(dimethylsiloxane) (PDMS),8 glasses and silicon,9 and polymers and silicon.10 In our system, we combined microfluidic channels and microoptical waveguides on a single polymer substrate made from poly(methyl methacrylate) (PMMA). This approach is favorable due to biocompatibility, optical properties, fabrication flexibility, and the low cost of the raw material.11 After fabrication of the chip, stained living animal cells or fluorophore-labeled phospholipids were deposited onto the planar chip or into a microfluidic channel in the chip. The fluorescent labels were excited by light coupled into the integrated waveguide and guided to the cells or phospholipids. Emitted fluorescence was observed with a microscope.
For the implementation of a fluorescence measurement on an integrated polymer chip, two different types of fluorescent medium and two different layouts were set up (Fig. 1 ). The two media are fluorescently labeled phospholipids [Figs. 1 and 1] on the one hand and fluorescently stained cells [Figs. 1 and 1] on the other hand. Those were either introduced into a channel crossing an optical waveguide [Figs. 1 and 1] or directly applied on top of an optical waveguide [Figs. 1 and 1].
The two layouts were realized on a polymer substrate. Here, photochemical degradation of PMMA by deep ultraviolet light (DUV) is used to process the channels and optical waveguides (Fig. 2 ).
The microchannels in setups A and B were generated by DUV lithography and subsequent development. In a first step, the channel structures are patterned by conventional shadow-printing lithography using a quartz-chromium mask. Bulk PMMA (Hesa®Glas by NotzPlastics) was used as semifinished product and exposed to DUV radiation [Fig. 2] with a commercial UV-lithography system (maskaligner EVG 620) operated at a wavelength of . The DUV-exposed PMMA is photochemically degraded (i.e., molecular chains of the polymer are cracked such that the average molecular weight is reduced). This reduction leads to alteration of chemical and optical properties of the exposed PMMA. Because of the altered solubility of the irradiated PMMA, the exposed areas can be dissolved using a developer. In our case, an exposure dose of was applied and the development of the irradiated PMMA was carried out using methyl isobutyl ketone (MIBK) and isopropanole (IPA),12 resulting in a trench of depth [Fig. 2].
An exposure dose of applied with the same UV-lithography system was chosen to create the optical waveguides. Because of the DUV exposure, the irradiated PMMA shows a refractive index increase of refractive index units in comparison to the refractive index of 1.48 for the substrate. The irradiated PMMA can be used as core of an optical waveguide13 with a graded index profile and an exponential decay in vertical direction.
In our case, the optical waveguides were structured into the PMMA substrate as lines of 5 or width by DUV lithography. For setups C and D, the waveguides were patterned into the bulk PMMA substrate while for setups A and B, they were patterned perpendicularly to the microfluidic channels.
After the patterning of microchannels and optical waveguides, the chip was mounted on a transparent polymer ground plate. This groundplate was used as support for fixing a glass fiber with glue in a way such that light could be coupled into the waveguide. Fixed on the ground plate and coupled to the glass fiber, the optofluidic chip could be used for the culturing of cells and the following experiments under a conventional fluorescence microscope.
Labeling and Observation of Fluorescence
After fabrication of microfludic channels (setups A and B) and optical waveguides in PMMA the chips were tested with fluorescently labeled phospholipids (cases A and C) or cells (cases B and D).
The labeled phospholipids were deposited via Dip-Pen nanolithography (DPN), which can be regarded as a modified and parallelized atomic force microscopy (AFM) setup.14, 15 Here, the AFM tips work similar to a pen, taking up a solution of phospholipids. By directing the AFM-tips onto the polymer surface, the phospholipid can be deposited with a lateral precision of several hundred nanometers. Thus, this method is particularly suited for deposition of phospholipid into, e.g., microfluidic channels as in our case.16
For the fluorescence measurement, the PMMA chip was mounted on a polymer base plate. A glass fiber was fixed on the base plate in order to allow stable end fire coupling to the optical waveguide on the chip. For fluorescence excitation light was coupled into the waveguide and fluorescence detected with a microscope objective (Fig. 3 ).
For setup A, biotinylated phospholipids (18:1 Lissamine Rhodamine PE, Avanti No. 810150, Avanti Polar Lipids Inc.) were deposited inside the microfluidic channel in the intersection of channel and waveguide, forming a spot of lipids on the bottom of the channel [Fig. 1]. Subsequently, streptavidin (Cy-3, Product No. S6402-1ML, Sigma-Aldrich Co.) labeled with the fluorescent dye Cy3 (absorption maximum , emission maximum ) was bound to the biotinylated lipids. After depositing Cy3-labeled lipids into the channel-waveguide intersection by DPN, fluorescence was excited by green light and observed with a fluorescence microscope. In a first step, a reference image was taken with a Nikon TE 2000 microscope using a mercury source filtered for excitation at . Emission was detected at [Fig. 4 , second column]. Then, the fluorescent labels were excited by green laser light (wavelength ), which was coupled to and guided through the polymer waveguide directly to the channel-waveguide intersection. Herewith, the part of the phospholipid spot directly located in the intersection area could be excited selectively by the light coupled out of the end-facet of the polymer waveguide inside the channel [Fig. 4, third column].
For setup B, the same experiment was carried out using living L929 mouse fibroblasts. The cells were cultured following a conventional culturing.17 After trypsination, the cells were stained in suspension with DiD Vybrant™ Cell-Labeling solution (Catalogue No. V-22887, Molecular Probes), according to the manufacturers recommendations. This lipophilic carbocyanine dye (absorption maximum , emission maximum ) selectively stains the cell membrane. After washing the cell suspension to remove unbound dye, the cell suspension was introduced into the channel and microscopy images were recorded while the cells remained in suspension. Results from this experiment are similar to the one with labeled phospholipids. Under flood exposure in the wavelength range of with a Zeiss Apotome, all cells floating through the channel were excited. The emitted fluorescent light was detected at [Fig. 4, second column]. Coupling red laser light (wavelength ) into the waveguide allowed to excite solely the cells crossing the laser light beam at the moment of capture [Fig. 4, third column].
For setup C, DPN was carried out depositing phospholipids labeled with Rhodamine B (absorption maximum , emission maximum ) directly onto a polymer waveguide [Fig. 1]. In this case (as well as in setup D), our goal was to excite the fluorescent labels via the evanescent field of the optical mode propagating through the waveguide buried in the polymer. The evanescent field of the waveguide mode extends a few hundred nanometers into the medium above the polymer (Fig. 5 ). Fluorescence was excited by flood exposure with the same microscope setup as in case A as a reference measurement. The labeled spot of phospholipids is clearly visible [Fig. 4, second column]. Afterward, green laser light (wavelength ) was coupled into the waveguide, in order to excite the phospholipids via the evanescent field. Solely, the lipids directly covering the waveguide were visible in the fluorescence microscope [Fig. 4, third column]. Hence, the evanescent field portion is sufficient to excite the fluorescent labels in the lipid layer.
In setup D, we wanted to visualize fluorophore stained cells via the evanescent field of a waveguide mode. Thus, the mouse fibroplasts stained with the DiD were allowed to adhere to the isopropanole washed polymer surface for three days. Although the cell number approximately doubled during this time due to cell proliferation, the staining was found to be still sufficient for imaging. However, the morphology of the fibroblasts was rather nonphysiological, indicating weak adhesion to the substrate, which might have to be improved by surface modifications. After a measurement under flood exposure with the same microscope setup as in case B [Fig. 4, second column], the cells were excited by red laser light (wavelength ) guided along the underlying waveguide. As in case C, fluorescence could be excited selectively for the cells adhered directly above the waveguide [Fig. 4, third column].
We realized the visualization of fluorescence excited via waveguides in an integrated optofluidic microsystem for fluorescently labeled phospholipids and stained cells. The measurement is a proof of concept for the layout of a cell-based optofluidic microsensor built as a monolithic polymer device. The advantage of the presented layout compared to other waveguide excitation systems5, 6 is the lateral definition of the waveguide, which makes it possible to address groups of cells selectively. Applications of such a system could be the selective investigation of morphology changes and different cellular responses2, 18 as well as the detection of cells in separation systems or kinetics of binding events via monitoring fluorescence reactions.
The presented results could also be interesting for lab-on-a-chip systems. For the presented optofluidic system to be used as handheld device the microfluidic channels need to be covered. This can be accomplished by thermal bonding or laser welding of another polymer sheet on top of the optofluidic chip. Further steps on the road toward a handheld lab-on-a-chip system are the integration of light sources and fluorescence detectors. Regarding the integration of light sources meeting the requirements for low bandwidth fluorescence excitation, this could be realized by organic semiconductor lasers. Those lasers can be directly integrated into a PMMA substrate and optically pumped from the outside. The detection could be realized by application of photodiodes, which are combined with a filter just as the fluorescence filters used in the fluorescence microscopes. The filters could be realized integrating Bragg gratings that are patterned into a polymer waveguide, which is guiding the emitted fluorescent light from the analyte to the photodiode. The integration of those elements is subject of ongoing research in our group.
Timo Mappes’ Young Investigator Group ‘YIG 08’ received financial support by the “Concept for the Future” of Karlsruhe Institute of Technology within the framework of the German Excellence Initiative. Mauno Schelb and Christoph Vannahme acknowledge financial support from the “Karlsruhe School of Optics and Photonics.” We acknowledge the Karlsruhe Nano Micro Facility (KNMF, www.kit.edu/knmf) of the Karlsruhe Institute of Technology for provision of access to instruments at their laboratories and we would like to thank Steven Lenhert for assistance in using the Dip-Pen Nanolithography Laboratory.